CINXE.COM

<!DOCTYPE html> <html lang="en"> <head> <meta charset="utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1"> <meta name="description" content="Keras documentation"> <meta name="author" content="Keras Team"> <link rel="shortcut icon" href="https://keras.io/img/favicon.ico"> <link rel="canonical" href="https://keras.io/examples/vision/zero_dce/" />  <meta property="og:title" content="Keras documentation: Zero-DCE for low-light image enhancement"> <meta property="og:image" content="https://keras.io/img/logo-k-keras-wb.png"> <meta name="twitter:title" content="Keras documentation: Zero-DCE for low-light image enhancement"> <meta name="twitter:image" content="https://keras.io/img/k-keras-social.png"> <meta name="twitter:card" content="summary"> <title>Zero-DCE for low-light image enhancement</title>  <link href="/css/bootstrap.min.css" rel="stylesheet">  <link href="https://fonts.googleapis.com/css2?family=Open+Sans:wght@400;600;700;800&display=swap" rel="stylesheet">  <link href="/css/docs.css" rel="stylesheet"> <link href="/css/monokai.css" rel="stylesheet">  <script>(function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start': new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0], j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src= 'https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f); })(window,document,'script','dataLayer','GTM-5DNGF4N'); </script> <script> (function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','https://www.google-analytics.com/analytics.js','ga'); ga('create', 'UA-175165319-128', 'auto'); ga('send', 'pageview'); </script>  <script async defer src="https://buttons.github.io/buttons.js"></script> </head> <body>  <noscript><iframe src="https://www.googletagmanager.com/ns.html?id=GTM-5DNGF4N" height="0" width="0" style="display:none;visibility:hidden"></iframe></noscript>  <div class='k-page'> <div class="k-nav" id="nav-menu"> <a href='/'><img src='/img/logo-small.png' class='logo-small' /></a> <div class="nav flex-column nav-pills" role="tablist" aria-orientation="vertical"> <a class="nav-link" href="/about/" role="tab" aria-selected="">About Keras</a> <a class="nav-link" href="/getting_started/" role="tab" aria-selected="">Getting started</a> <a class="nav-link" href="/guides/" role="tab" aria-selected="">Developer guides</a> <a class="nav-link" href="/api/" role="tab" aria-selected="">Keras 3 API documentation</a> <a class="nav-link" href="/2.18/api/" role="tab" aria-selected="">Keras 2 API documentation</a> <a class="nav-link active" href="/examples/" role="tab" aria-selected="">Code examples</a> <a class="nav-sublink active" href="/examples/vision/">Computer Vision</a> <a class="nav-sublink2" href="/examples/vision/image_classification_from_scratch/">Image classification from scratch</a> <a class="nav-sublink2" href="/examples/vision/mnist_convnet/">Simple MNIST convnet</a> <a class="nav-sublink2" href="/examples/vision/image_classification_efficientnet_fine_tuning/">Image classification via fine-tuning with EfficientNet</a> <a class="nav-sublink2" href="/examples/vision/image_classification_with_vision_transformer/">Image classification with Vision Transformer</a> <a class="nav-sublink2" href="/examples/vision/attention_mil_classification/">Classification using Attention-based Deep Multiple Instance Learning</a> <a class="nav-sublink2" href="/examples/vision/mlp_image_classification/">Image classification with modern MLP models</a> <a class="nav-sublink2" href="/examples/vision/mobilevit/">A mobile-friendly Transformer-based model for image classification</a> <a class="nav-sublink2" href="/examples/vision/xray_classification_with_tpus/">Pneumonia Classification on TPU</a> <a class="nav-sublink2" href="/examples/vision/cct/">Compact Convolutional Transformers</a> <a class="nav-sublink2" href="/examples/vision/convmixer/">Image classification with ConvMixer</a> <a class="nav-sublink2" href="/examples/vision/eanet/">Image classification with EANet (External Attention Transformer)</a> <a class="nav-sublink2" href="/examples/vision/involution/">Involutional neural networks</a> <a class="nav-sublink2" href="/examples/vision/perceiver_image_classification/">Image classification with Perceiver</a> <a class="nav-sublink2" href="/examples/vision/reptile/">Few-Shot learning with Reptile</a> <a class="nav-sublink2" href="/examples/vision/semisupervised_simclr/">Semi-supervised image classification using contrastive pretraining with SimCLR</a> <a class="nav-sublink2" href="/examples/vision/swin_transformers/">Image classification with Swin Transformers</a> <a class="nav-sublink2" href="/examples/vision/vit_small_ds/">Train a Vision Transformer on small datasets</a> <a class="nav-sublink2" href="/examples/vision/shiftvit/">A Vision Transformer without Attention</a> <a class="nav-sublink2" href="/examples/vision/image_classification_using_global_context_vision_transformer/">Image Classification using Global Context Vision Transformer</a> <a class="nav-sublink2" href="/examples/vision/oxford_pets_image_segmentation/">Image segmentation with a U-Net-like architecture</a> <a class="nav-sublink2" href="/examples/vision/deeplabv3_plus/">Multiclass semantic segmentation using DeepLabV3+</a> <a class="nav-sublink2" href="/examples/vision/basnet_segmentation/">Highly accurate boundaries segmentation using BASNet</a> <a class="nav-sublink2" href="/examples/vision/fully_convolutional_network/">Image Segmentation using Composable Fully-Convolutional Networks</a> <a class="nav-sublink2" href="/examples/vision/retinanet/">Object Detection with RetinaNet</a> <a class="nav-sublink2" href="/examples/vision/keypoint_detection/">Keypoint Detection with Transfer Learning</a> <a class="nav-sublink2" href="/examples/vision/object_detection_using_vision_transformer/">Object detection with Vision Transformers</a> <a class="nav-sublink2" href="/examples/vision/3D_image_classification/">3D image classification from CT scans</a> <a class="nav-sublink2" href="/examples/vision/depth_estimation/">Monocular depth estimation</a> <a class="nav-sublink2" href="/examples/vision/nerf/">3D volumetric rendering with NeRF</a> <a class="nav-sublink2" href="/examples/vision/pointnet_segmentation/">Point cloud segmentation with PointNet</a> <a class="nav-sublink2" href="/examples/vision/pointnet/">Point cloud classification</a> <a class="nav-sublink2" href="/examples/vision/captcha_ocr/">OCR model for reading Captchas</a> <a class="nav-sublink2" href="/examples/vision/handwriting_recognition/">Handwriting recognition</a> <a class="nav-sublink2" href="/examples/vision/autoencoder/">Convolutional autoencoder for image denoising</a> <a class="nav-sublink2" href="/examples/vision/mirnet/">Low-light image enhancement using MIRNet</a> <a class="nav-sublink2" href="/examples/vision/super_resolution_sub_pixel/">Image Super-Resolution using an Efficient Sub-Pixel CNN</a> <a class="nav-sublink2" href="/examples/vision/edsr/">Enhanced Deep Residual Networks for single-image super-resolution</a> <a class="nav-sublink2 active" href="/examples/vision/zero_dce/">Zero-DCE for low-light image enhancement</a> <a class="nav-sublink2" href="/examples/vision/cutmix/">CutMix data augmentation for image classification</a> <a class="nav-sublink2" href="/examples/vision/mixup/">MixUp augmentation for image classification</a> <a class="nav-sublink2" href="/examples/vision/randaugment/">RandAugment for Image Classification for Improved Robustness</a> <a class="nav-sublink2" href="/examples/vision/image_captioning/">Image captioning</a> <a class="nav-sublink2" href="/examples/vision/nl_image_search/">Natural language image search with a Dual Encoder</a> <a class="nav-sublink2" href="/examples/vision/visualizing_what_convnets_learn/">Visualizing what convnets learn</a> <a class="nav-sublink2" href="/examples/vision/integrated_gradients/">Model interpretability with Integrated Gradients</a> <a class="nav-sublink2" href="/examples/vision/probing_vits/">Investigating Vision Transformer representations</a> <a class="nav-sublink2" href="/examples/vision/grad_cam/">Grad-CAM class activation visualization</a> <a class="nav-sublink2" href="/examples/vision/near_dup_search/">Near-duplicate image search</a> <a class="nav-sublink2" href="/examples/vision/semantic_image_clustering/">Semantic Image Clustering</a> <a class="nav-sublink2" href="/examples/vision/siamese_contrastive/">Image similarity estimation using a Siamese Network with a contrastive loss</a> <a class="nav-sublink2" href="/examples/vision/siamese_network/">Image similarity estimation using a Siamese Network with a triplet loss</a> <a class="nav-sublink2" href="/examples/vision/metric_learning/">Metric learning for image similarity search</a> <a class="nav-sublink2" href="/examples/vision/metric_learning_tf_similarity/">Metric learning for image similarity search using TensorFlow Similarity</a> <a class="nav-sublink2" href="/examples/vision/nnclr/">Self-supervised contrastive learning with NNCLR</a> <a class="nav-sublink2" href="/examples/vision/video_classification/">Video Classification with a CNN-RNN Architecture</a> <a class="nav-sublink2" href="/examples/vision/conv_lstm/">Next-Frame Video Prediction with Convolutional LSTMs</a> <a class="nav-sublink2" href="/examples/vision/video_transformers/">Video Classification with Transformers</a> <a class="nav-sublink2" href="/examples/vision/vivit/">Video Vision Transformer</a> <a class="nav-sublink2" href="/examples/vision/bit/">Image Classification using BigTransfer (BiT)</a> <a class="nav-sublink2" href="/examples/vision/gradient_centralization/">Gradient Centralization for Better Training Performance</a> <a class="nav-sublink2" href="/examples/vision/token_learner/">Learning to tokenize in Vision Transformers</a> <a class="nav-sublink2" href="/examples/vision/knowledge_distillation/">Knowledge Distillation</a> <a class="nav-sublink2" href="/examples/vision/fixres/">FixRes: Fixing train-test resolution discrepancy</a> <a class="nav-sublink2" href="/examples/vision/cait/">Class Attention Image Transformers with LayerScale</a> <a class="nav-sublink2" href="/examples/vision/patch_convnet/">Augmenting convnets with aggregated attention</a> <a class="nav-sublink2" href="/examples/vision/learnable_resizer/">Learning to Resize</a> <a class="nav-sublink2" href="/examples/vision/adamatch/">Semi-supervision and domain adaptation with AdaMatch</a> <a class="nav-sublink2" href="/examples/vision/barlow_twins/">Barlow Twins for Contrastive SSL</a> <a class="nav-sublink2" href="/examples/vision/consistency_training/">Consistency training with supervision</a> <a class="nav-sublink2" href="/examples/vision/deit/">Distilling Vision Transformers</a> <a class="nav-sublink2" href="/examples/vision/focal_modulation_network/">Focal Modulation: A replacement for Self-Attention</a> <a class="nav-sublink2" href="/examples/vision/forwardforward/">Using the Forward-Forward Algorithm for Image Classification</a> <a class="nav-sublink2" href="/examples/vision/masked_image_modeling/">Masked image modeling with Autoencoders</a> <a class="nav-sublink2" href="/examples/vision/sam/">Segment Anything Model with 🤗Transformers</a> <a class="nav-sublink2" href="/examples/vision/segformer/">Semantic segmentation with SegFormer and Hugging Face Transformers</a> <a class="nav-sublink2" href="/examples/vision/simsiam/">Self-supervised contrastive learning with SimSiam</a> <a class="nav-sublink2" href="/examples/vision/supervised-contrastive-learning/">Supervised Contrastive Learning</a> <a class="nav-sublink2" href="/examples/vision/temporal_latent_bottleneck/">When Recurrence meets Transformers</a> <a class="nav-sublink2" href="/examples/vision/yolov8/">Efficient Object Detection with YOLOV8 and KerasCV</a> <a class="nav-sublink" href="/examples/nlp/">Natural Language Processing</a> <a class="nav-sublink" href="/examples/structured_data/">Structured Data</a> <a class="nav-sublink" href="/examples/timeseries/">Timeseries</a> <a class="nav-sublink" href="/examples/generative/">Generative Deep Learning</a> <a class="nav-sublink" href="/examples/audio/">Audio Data</a> <a class="nav-sublink" href="/examples/rl/">Reinforcement Learning</a> <a class="nav-sublink" href="/examples/graph/">Graph Data</a> <a class="nav-sublink" href="/examples/keras_recipes/">Quick Keras Recipes</a> <a class="nav-link" href="/keras_tuner/" role="tab" aria-selected="">KerasTuner: Hyperparameter Tuning</a> <a class="nav-link" href="/keras_hub/" role="tab" aria-selected="">KerasHub: Pretrained Models</a> <a class="nav-link" href="/keras_cv/" role="tab" aria-selected="">KerasCV: Computer Vision Workflows</a> <a class="nav-link" href="/keras_nlp/" role="tab" aria-selected="">KerasNLP: Natural Language Workflows</a> </div> </div> <div class='k-main'> <div class='k-main-top'> <script> function displayDropdownMenu() { e = document.getElementById("nav-menu"); if (e.style.display == "block") { e.style.display = "none"; } else { e.style.display = "block"; document.getElementById("dropdown-nav").style.display = "block"; } } function resetMobileUI() { if (window.innerWidth <= 840) { document.getElementById("nav-menu").style.display = "none"; document.getElementById("dropdown-nav").style.display = "block"; } else { document.getElementById("nav-menu").style.display = "block"; document.getElementById("dropdown-nav").style.display = "none"; } var navmenu = document.getElementById("nav-menu"); var menuheight = navmenu.clientHeight; var kmain = document.getElementById("k-main-id"); kmain.style.minHeight = (menuheight + 100) + 'px'; } window.onresize = resetMobileUI; window.addEventListener("load", (event) => { resetMobileUI() }); </script> <div id='dropdown-nav' onclick="displayDropdownMenu();"> <svg viewBox="-20 -20 120 120" width="60" height="60"> <rect width="100" height="20"></rect> <rect y="30" width="100" height="20"></rect> <rect y="60" width="100" height="20"></rect> </svg> </div> <form class="bd-search d-flex align-items-center k-search-form" id="search-form"> <input type="search" class="k-search-input" id="search-input" placeholder="Search Keras documentation..." aria-label="Search Keras documentation..." autocomplete="off"> <button class="k-search-btn"> <svg width="13" height="13" viewBox="0 0 13 13"><title>search</title><path d="m4.8495 7.8226c0.82666 0 1.5262-0.29146 2.0985-0.87438 0.57232-0.58292 0.86378-1.2877 0.87438-2.1144 0.010599-0.82666-0.28086-1.5262-0.87438-2.0985-0.59352-0.57232-1.293-0.86378-2.0985-0.87438-0.8055-0.010599-1.5103 0.28086-2.1144 0.87438-0.60414 0.59352-0.8956 1.293-0.87438 2.0985 0.021197 0.8055 0.31266 1.5103 0.87438 2.1144 0.56172 0.60414 1.2665 0.8956 2.1144 0.87438zm4.4695 0.2115 3.681 3.6819-1.259 1.284-3.6817-3.7 0.0019784-0.69479-0.090043-0.098846c-0.87973 0.76087-1.92 1.1413-3.1207 1.1413-1.3553 0-2.5025-0.46363-3.4417-1.3909s-1.4088-2.0686-1.4088-3.4239c0-1.3553 0.4696-2.4966 1.4088-3.4239 0.9392-0.92727 2.0864-1.3969 3.4417-1.4088 1.3553-0.011889 2.4906 0.45771 3.406 1.4088 0.9154 0.95107 1.379 2.0924 1.3909 3.4239 0 1.2126-0.38043 2.2588-1.1413 3.1385l0.098834 0.090049z"></path></svg> </button> </form> <script> var form = document.getElementById('search-form'); form.onsubmit = function(e) { e.preventDefault(); var query = document.getElementById('search-input').value; window.location.href = '/search.html?query=' + query; return False } </script> </div> <div class='k-main-inner' id='k-main-id'> <div class='k-location-slug'> ► <a href='/examples/'>Code examples</a> / <a href='/examples/vision/'>Computer Vision</a> / Zero-DCE for low-light image enhancement </div> <div class='k-content'> <h1 id="zerodce-for-lowlight-image-enhancement">Zero-DCE for low-light image enhancement</h1> Author: <a href="http://github.com/soumik12345">Soumik Rakshit</a> Date created: 2021/09/18 Last modified: 2023/07/15 Description: Implementing Zero-Reference Deep Curve Estimation for low-light image enhancement. <div class='example_version_banner keras_3'>ⓘ This example uses Keras 3</div> <img class="k-inline-icon" src="https://colab.research.google.com/img/colab_favicon.ico"/> <a href="https://colab.research.google.com/github/keras-team/keras-io/blob/master/examples/vision/ipynb/zero_dce.ipynb">View in Colab</a> •<img class="k-inline-icon" src="https://github.com/favicon.ico"/> <a href="https://github.com/keras-team/keras-io/blob/master/examples/vision/zero_dce.py">GitHub source</a> <hr /> <h2 id="introduction">Introduction</h2> Zero-Reference Deep Curve Estimation or Zero-DCE formulates low-light image enhancement as the task of estimating an image-specific <a href="https://en.wikipedia.org/wiki/Curve_(tonality)">tonal curve</a> with a deep neural network. In this example, we train a lightweight deep network, DCE-Net, to estimate pixel-wise and high-order tonal curves for dynamic range adjustment of a given image. Zero-DCE takes a low-light image as input and produces high-order tonal curves as its output. These curves are then used for pixel-wise adjustment on the dynamic range of the input to obtain an enhanced image. The curve estimation process is done in such a way that it maintains the range of the enhanced image and preserves the contrast of neighboring pixels. This curve estimation is inspired by curves adjustment used in photo editing software such as Adobe Photoshop where users can adjust points throughout an image’s tonal range. Zero-DCE is appealing because of its relaxed assumptions with regard to reference images: it does not require any input/output image pairs during training. This is achieved through a set of carefully formulated non-reference loss functions, which implicitly measure the enhancement quality and guide the training of the network. <h3 id="references">References</h3> <ul> <li><a href="https://arxiv.org/abs/2001.06826">Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement</a></li> <li><a href="https://helpx.adobe.com/photoshop/using/curves-adjustment.html">Curves adjustment in Adobe Photoshop</a></li> </ul> <hr /> <h2 id="downloading-loldataset">Downloading LOLDataset</h2> The LoL Dataset has been created for low-light image enhancement. It provides 485 images for training and 15 for testing. Each image pair in the dataset consists of a low-light input image and its corresponding well-exposed reference image. <div class="codehilite"><pre><code>import os os.environ["KERAS_BACKEND"] = "tensorflow" import random import numpy as np from glob import glob from PIL import Image, ImageOps import matplotlib.pyplot as plt import keras from keras import layers import tensorflow as tf </code></pre></div> <div class="codehilite"><pre><code>!wget https://huggingface.co/datasets/geekyrakshit/LoL-Dataset/resolve/main/lol_dataset.zip !unzip -q lol_dataset.zip && rm lol_dataset.zip </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>--2023-11-20 20:01:50-- https://huggingface.co/datasets/geekyrakshit/LoL-Dataset/resolve/main/lol_dataset.zip Resolving huggingface.co (huggingface.co)... 3.163.189.74, 3.163.189.90, 3.163.189.114, ... Connecting to huggingface.co (huggingface.co)|3.163.189.74|:443... connected. HTTP request sent, awaiting response... 302 Found Location: https://cdn-lfs.huggingface.co/repos/d9/09/d909ef7668bb417b7065a311bd55a3084cc83a1f918e13cb41c5503328432db2/419fddc48958cd0f5599939ee0248852a37ceb8bb738c9b9525e95b25a89de9a?response-content-disposition=attachment%3B+filename*%3DUTF-8%27%27lol_dataset.zip%3B+filename%3D%22lol_dataset.zip%22%3B&response-content-type=application%2Fzip&Expires=1700769710&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcwMDc2OTcxMH19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy5odWdnaW5nZmFjZS5jby9yZXBvcy9kOS8wOS9kOTA5ZWY3NjY4YmI0MTdiNzA2NWEzMTFiZDU1YTMwODRjYzgzYTFmOTE4ZTEzY2I0MWM1NTAzMzI4NDMyZGIyLzQxOWZkZGM0ODk1OGNkMGY1NTk5OTM5ZWUwMjQ4ODUyYTM3Y2ViOGJiNzM4YzliOTUyNWU5NWIyNWE4OWRlOWE%7EcmVzcG9uc2UtY29udGVudC1kaXNwb3NpdGlvbj0qJnJlc3BvbnNlLWNvbnRlbnQtdHlwZT0qIn1dfQ__&Signature=VPqHlt0h6mUV7D3alDMIO61VSvUX498wZn5rIpo4u5yTYOu2s9CbO82xeGfrZguIuENVO6yiuoUAlZO4XXDsGC0Gc3MR3KIoTGuI9URA815nrdvFE616XBooGAW200KOUmVj2IoySAufi-7ORPuspaVJoKqWr8wytt0hDpNMeaWSg766kVMkJB1Aywq6yu5KHFGkqvOPDWNZZO6yfOtdX2XfbXVuiaiUlS03gRZ58H9pYn535TrE3BYP4W1u%7EehJ4OACpsRsnrsrXDr--PLH5RsxApOR2neFLySta3LiN9mtdjSpOKGn0oUapDfCWG7Ik5OMB5PGGzQBTB5J0b0O9g__&Key-Pair-Id=KVTP0A1DKRTAX [following] --2023-11-20 20:01:50-- https://cdn-lfs.huggingface.co/repos/d9/09/d909ef7668bb417b7065a311bd55a3084cc83a1f918e13cb41c5503328432db2/419fddc48958cd0f5599939ee0248852a37ceb8bb738c9b9525e95b25a89de9a?response-content-disposition=attachment%3B+filename*%3DUTF-8%27%27lol_dataset.zip%3B+filename%3D%22lol_dataset.zip%22%3B&response-content-type=application%2Fzip&Expires=1700769710&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcwMDc2OTcxMH19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy5odWdnaW5nZmFjZS5jby9yZXBvcy9kOS8wOS9kOTA5ZWY3NjY4YmI0MTdiNzA2NWEzMTFiZDU1YTMwODRjYzgzYTFmOTE4ZTEzY2I0MWM1NTAzMzI4NDMyZGIyLzQxOWZkZGM0ODk1OGNkMGY1NTk5OTM5ZWUwMjQ4ODUyYTM3Y2ViOGJiNzM4YzliOTUyNWU5NWIyNWE4OWRlOWE%7EcmVzcG9uc2UtY29udGVudC1kaXNwb3NpdGlvbj0qJnJlc3BvbnNlLWNvbnRlbnQtdHlwZT0qIn1dfQ__&Signature=VPqHlt0h6mUV7D3alDMIO61VSvUX498wZn5rIpo4u5yTYOu2s9CbO82xeGfrZguIuENVO6yiuoUAlZO4XXDsGC0Gc3MR3KIoTGuI9URA815nrdvFE616XBooGAW200KOUmVj2IoySAufi-7ORPuspaVJoKqWr8wytt0hDpNMeaWSg766kVMkJB1Aywq6yu5KHFGkqvOPDWNZZO6yfOtdX2XfbXVuiaiUlS03gRZ58H9pYn535TrE3BYP4W1u%7EehJ4OACpsRsnrsrXDr--PLH5RsxApOR2neFLySta3LiN9mtdjSpOKGn0oUapDfCWG7Ik5OMB5PGGzQBTB5J0b0O9g__&Key-Pair-Id=KVTP0A1DKRTAX Resolving cdn-lfs.huggingface.co (cdn-lfs.huggingface.co)... 108.138.94.122, 108.138.94.25, 108.138.94.14, ... Connecting to cdn-lfs.huggingface.co (cdn-lfs.huggingface.co)|108.138.94.122|:443... connected. HTTP request sent, awaiting response... 200 OK Length: 347171015 (331M) [application/zip] Saving to: ‘lol_dataset.zip’ </code></pre></div> </div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>lol_dataset.zip 100%[===================>] 331.09M 37.4MB/s in 9.5s </code></pre></div> </div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>2023-11-20 20:02:00 (34.9 MB/s) - ‘lol_dataset.zip’ saved [347171015/347171015] </code></pre></div> </div> <hr /> <h2 id="creating-a-tensorflow-dataset">Creating a TensorFlow Dataset</h2> We use 300 low-light images from the LoL Dataset training set for training, and we use the remaining 185 low-light images for validation. We resize the images to size <code>256 x 256</code> to be used for both training and validation. Note that in order to train the DCE-Net, we will not require the corresponding enhanced images. <div class="codehilite"><pre><code>IMAGE_SIZE = 256 BATCH_SIZE = 16 MAX_TRAIN_IMAGES = 400 def load_data(image_path): image = tf.io.read_file(image_path) image = tf.image.decode_png(image, channels=3) image = tf.image.resize(images=image, size=[IMAGE_SIZE, IMAGE_SIZE]) image = image / 255.0 return image def data_generator(low_light_images): dataset = tf.data.Dataset.from_tensor_slices((low_light_images)) dataset = dataset.map(load_data, num_parallel_calls=tf.data.AUTOTUNE) dataset = dataset.batch(BATCH_SIZE, drop_remainder=True) return dataset train_low_light_images = sorted(glob("./lol_dataset/our485/low/*"))[:MAX_TRAIN_IMAGES] val_low_light_images = sorted(glob("./lol_dataset/our485/low/*"))[MAX_TRAIN_IMAGES:] test_low_light_images = sorted(glob("./lol_dataset/eval15/low/*")) train_dataset = data_generator(train_low_light_images) val_dataset = data_generator(val_low_light_images) print("Train Dataset:", train_dataset) print("Validation Dataset:", val_dataset) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Train Dataset: <_BatchDataset element_spec=TensorSpec(shape=(16, 256, 256, 3), dtype=tf.float32, name=None)> Validation Dataset: <_BatchDataset element_spec=TensorSpec(shape=(16, 256, 256, 3), dtype=tf.float32, name=None)> </code></pre></div> </div> <hr /> <h2 id="the-zerodce-framework">The Zero-DCE Framework</h2> The goal of DCE-Net is to estimate a set of best-fitting light-enhancement curves (LE-curves) given an input image. The framework then maps all pixels of the input’s RGB channels by applying the curves iteratively to obtain the final enhanced image. <h3 id="understanding-lightenhancement-curves">Understanding light-enhancement curves</h3> A ligh-enhancement curve is a kind of curve that can map a low-light image to its enhanced version automatically, where the self-adaptive curve parameters are solely dependent on the input image. When designing such a curve, three objectives should be taken into account: <ul> <li>Each pixel value of the enhanced image should be in the normalized range <code>[0,1]</code>, in order to avoid information loss induced by overflow truncation.</li> <li>It should be monotonous, to preserve the contrast between neighboring pixels.</li> <li>The shape of this curve should be as simple as possible, and the curve should be differentiable to allow backpropagation.</li> </ul> The light-enhancement curve is separately applied to three RGB channels instead of solely on the illumination channel. The three-channel adjustment can better preserve the inherent color and reduce the risk of over-saturation. <img alt="" src="https://li-chongyi.github.io/Zero-DCE_files/framework.png" /> <h3 id="dcenet">DCE-Net</h3> The DCE-Net is a lightweight deep neural network that learns the mapping between an input image and its best-fitting curve parameter maps. The input to the DCE-Net is a low-light image while the outputs are a set of pixel-wise curve parameter maps for corresponding higher-order curves. It is a plain CNN of seven convolutional layers with symmetrical concatenation. Each layer consists of 32 convolutional kernels of size 3×3 and stride 1 followed by the ReLU activation function. The last convolutional layer is followed by the Tanh activation function, which produces 24 parameter maps for 8 iterations, where each iteration requires three curve parameter maps for the three channels. <img alt="" src="https://i.imgur.com/HtIg34W.png" /> <div class="codehilite"><pre><code>def build_dce_net(): input_img = keras.Input(shape=[None, None, 3]) conv1 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(input_img) conv2 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(conv1) conv3 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(conv2) conv4 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(conv3) int_con1 = layers.Concatenate(axis=-1)([conv4, conv3]) conv5 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(int_con1) int_con2 = layers.Concatenate(axis=-1)([conv5, conv2]) conv6 = layers.Conv2D( 32, (3, 3), strides=(1, 1), activation="relu", padding="same" )(int_con2) int_con3 = layers.Concatenate(axis=-1)([conv6, conv1]) x_r = layers.Conv2D(24, (3, 3), strides=(1, 1), activation="tanh", padding="same")( int_con3 ) return keras.Model(inputs=input_img, outputs=x_r) </code></pre></div> <hr /> <h2 id="loss-functions">Loss functions</h2> To enable zero-reference learning in DCE-Net, we use a set of differentiable zero-reference losses that allow us to evaluate the quality of enhanced images. <h3 id="color-constancy-loss">Color constancy loss</h3> The color constancy loss is used to correct the potential color deviations in the enhanced image. <div class="codehilite"><pre><code>def color_constancy_loss(x): mean_rgb = tf.reduce_mean(x, axis=(1, 2), keepdims=True) mr, mg, mb = ( mean_rgb[:, :, :, 0], mean_rgb[:, :, :, 1], mean_rgb[:, :, :, 2], ) d_rg = tf.square(mr - mg) d_rb = tf.square(mr - mb) d_gb = tf.square(mb - mg) return tf.sqrt(tf.square(d_rg) + tf.square(d_rb) + tf.square(d_gb)) </code></pre></div> <h3 id="exposure-loss">Exposure loss</h3> To restrain under-/over-exposed regions, we use the exposure control loss. It measures the distance between the average intensity value of a local region and a preset well-exposedness level (set to <code>0.6</code>). <div class="codehilite"><pre><code>def exposure_loss(x, mean_val=0.6): x = tf.reduce_mean(x, axis=3, keepdims=True) mean = tf.nn.avg_pool2d(x, ksize=16, strides=16, padding="VALID") return tf.reduce_mean(tf.square(mean - mean_val)) </code></pre></div> <h3 id="illumination-smoothness-loss">Illumination smoothness loss</h3> To preserve the monotonicity relations between neighboring pixels, the illumination smoothness loss is added to each curve parameter map. <div class="codehilite"><pre><code>def illumination_smoothness_loss(x): batch_size = tf.shape(x)[0] h_x = tf.shape(x)[1] w_x = tf.shape(x)[2] count_h = (tf.shape(x)[2] - 1) * tf.shape(x)[3] count_w = tf.shape(x)[2] * (tf.shape(x)[3] - 1) h_tv = tf.reduce_sum(tf.square((x[:, 1:, :, :] - x[:, : h_x - 1, :, :]))) w_tv = tf.reduce_sum(tf.square((x[:, :, 1:, :] - x[:, :, : w_x - 1, :]))) batch_size = tf.cast(batch_size, dtype=tf.float32) count_h = tf.cast(count_h, dtype=tf.float32) count_w = tf.cast(count_w, dtype=tf.float32) return 2 * (h_tv / count_h + w_tv / count_w) / batch_size </code></pre></div> <h3 id="spatial-consistency-loss">Spatial consistency loss</h3> The spatial consistency loss encourages spatial coherence of the enhanced image by preserving the contrast between neighboring regions across the input image and its enhanced version. <div class="codehilite"><pre><code>class SpatialConsistencyLoss(keras.losses.Loss): def __init__(self, **kwargs): super().__init__(reduction="none") self.left_kernel = tf.constant( [[[[0, 0, 0]], [[-1, 1, 0]], [[0, 0, 0]]]], dtype=tf.float32 ) self.right_kernel = tf.constant( [[[[0, 0, 0]], [[0, 1, -1]], [[0, 0, 0]]]], dtype=tf.float32 ) self.up_kernel = tf.constant( [[[[0, -1, 0]], [[0, 1, 0]], [[0, 0, 0]]]], dtype=tf.float32 ) self.down_kernel = tf.constant( [[[[0, 0, 0]], [[0, 1, 0]], [[0, -1, 0]]]], dtype=tf.float32 ) def call(self, y_true, y_pred): original_mean = tf.reduce_mean(y_true, 3, keepdims=True) enhanced_mean = tf.reduce_mean(y_pred, 3, keepdims=True) original_pool = tf.nn.avg_pool2d( original_mean, ksize=4, strides=4, padding="VALID" ) enhanced_pool = tf.nn.avg_pool2d( enhanced_mean, ksize=4, strides=4, padding="VALID" ) d_original_left = tf.nn.conv2d( original_pool, self.left_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_original_right = tf.nn.conv2d( original_pool, self.right_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_original_up = tf.nn.conv2d( original_pool, self.up_kernel, strides=[1, 1, 1, 1], padding="SAME" ) d_original_down = tf.nn.conv2d( original_pool, self.down_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_enhanced_left = tf.nn.conv2d( enhanced_pool, self.left_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_enhanced_right = tf.nn.conv2d( enhanced_pool, self.right_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_enhanced_up = tf.nn.conv2d( enhanced_pool, self.up_kernel, strides=[1, 1, 1, 1], padding="SAME" ) d_enhanced_down = tf.nn.conv2d( enhanced_pool, self.down_kernel, strides=[1, 1, 1, 1], padding="SAME", ) d_left = tf.square(d_original_left - d_enhanced_left) d_right = tf.square(d_original_right - d_enhanced_right) d_up = tf.square(d_original_up - d_enhanced_up) d_down = tf.square(d_original_down - d_enhanced_down) return d_left + d_right + d_up + d_down </code></pre></div> <h3 id="deep-curve-estimation-model">Deep curve estimation model</h3> We implement the Zero-DCE framework as a Keras subclassed model. <div class="codehilite"><pre><code>class ZeroDCE(keras.Model): def __init__(self, **kwargs): super().__init__(**kwargs) self.dce_model = build_dce_net() def compile(self, learning_rate, **kwargs): super().compile(**kwargs) self.optimizer = keras.optimizers.Adam(learning_rate=learning_rate) self.spatial_constancy_loss = SpatialConsistencyLoss(reduction="none") self.total_loss_tracker = keras.metrics.Mean(name="total_loss") self.illumination_smoothness_loss_tracker = keras.metrics.Mean( name="illumination_smoothness_loss" ) self.spatial_constancy_loss_tracker = keras.metrics.Mean( name="spatial_constancy_loss" ) self.color_constancy_loss_tracker = keras.metrics.Mean( name="color_constancy_loss" ) self.exposure_loss_tracker = keras.metrics.Mean(name="exposure_loss") @property def metrics(self): return [ self.total_loss_tracker, self.illumination_smoothness_loss_tracker, self.spatial_constancy_loss_tracker, self.color_constancy_loss_tracker, self.exposure_loss_tracker, ] def get_enhanced_image(self, data, output): r1 = output[:, :, :, :3] r2 = output[:, :, :, 3:6] r3 = output[:, :, :, 6:9] r4 = output[:, :, :, 9:12] r5 = output[:, :, :, 12:15] r6 = output[:, :, :, 15:18] r7 = output[:, :, :, 18:21] r8 = output[:, :, :, 21:24] x = data + r1 * (tf.square(data) - data) x = x + r2 * (tf.square(x) - x) x = x + r3 * (tf.square(x) - x) enhanced_image = x + r4 * (tf.square(x) - x) x = enhanced_image + r5 * (tf.square(enhanced_image) - enhanced_image) x = x + r6 * (tf.square(x) - x) x = x + r7 * (tf.square(x) - x) enhanced_image = x + r8 * (tf.square(x) - x) return enhanced_image def call(self, data): dce_net_output = self.dce_model(data) return self.get_enhanced_image(data, dce_net_output) def compute_losses(self, data, output): enhanced_image = self.get_enhanced_image(data, output) loss_illumination = 200 * illumination_smoothness_loss(output) loss_spatial_constancy = tf.reduce_mean( self.spatial_constancy_loss(enhanced_image, data) ) loss_color_constancy = 5 * tf.reduce_mean(color_constancy_loss(enhanced_image)) loss_exposure = 10 * tf.reduce_mean(exposure_loss(enhanced_image)) total_loss = ( loss_illumination + loss_spatial_constancy + loss_color_constancy + loss_exposure ) return { "total_loss": total_loss, "illumination_smoothness_loss": loss_illumination, "spatial_constancy_loss": loss_spatial_constancy, "color_constancy_loss": loss_color_constancy, "exposure_loss": loss_exposure, } def train_step(self, data): with tf.GradientTape() as tape: output = self.dce_model(data) losses = self.compute_losses(data, output) gradients = tape.gradient( losses["total_loss"], self.dce_model.trainable_weights ) self.optimizer.apply_gradients(zip(gradients, self.dce_model.trainable_weights)) self.total_loss_tracker.update_state(losses["total_loss"]) self.illumination_smoothness_loss_tracker.update_state( losses["illumination_smoothness_loss"] ) self.spatial_constancy_loss_tracker.update_state( losses["spatial_constancy_loss"] ) self.color_constancy_loss_tracker.update_state(losses["color_constancy_loss"]) self.exposure_loss_tracker.update_state(losses["exposure_loss"]) return {metric.name: metric.result() for metric in self.metrics} def test_step(self, data): output = self.dce_model(data) losses = self.compute_losses(data, output) self.total_loss_tracker.update_state(losses["total_loss"]) self.illumination_smoothness_loss_tracker.update_state( losses["illumination_smoothness_loss"] ) self.spatial_constancy_loss_tracker.update_state( losses["spatial_constancy_loss"] ) self.color_constancy_loss_tracker.update_state(losses["color_constancy_loss"]) self.exposure_loss_tracker.update_state(losses["exposure_loss"]) return {metric.name: metric.result() for metric in self.metrics} def save_weights(self, filepath, overwrite=True, save_format=None, options=None): """While saving the weights, we simply save the weights of the DCE-Net""" self.dce_model.save_weights( filepath, overwrite=overwrite, save_format=save_format, options=options, ) def load_weights(self, filepath, by_name=False, skip_mismatch=False, options=None): """While loading the weights, we simply load the weights of the DCE-Net""" self.dce_model.load_weights( filepath=filepath, by_name=by_name, skip_mismatch=skip_mismatch, options=options, ) </code></pre></div> <hr /> <h2 id="training">Training</h2> <div class="codehilite"><pre><code>zero_dce_model = ZeroDCE() zero_dce_model.compile(learning_rate=1e-4) history = zero_dce_model.fit(train_dataset, validation_data=val_dataset, epochs=100) def plot_result(item): plt.plot(history.history[item], label=item) plt.plot(history.history["val_" + item], label="val_" + item) plt.xlabel("Epochs") plt.ylabel(item) plt.title("Train and Validation {} Over Epochs".format(item), fontsize=14) plt.legend() plt.grid() plt.show() plot_result("total_loss") plot_result("illumination_smoothness_loss") plot_result("spatial_constancy_loss") plot_result("color_constancy_loss") plot_result("exposure_loss") </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Epoch 1/100 2/25 ━[37m━━━━━━━━━━━━━━━━━━━ 1s 85ms/step - color_constancy_loss: 0.0013 - exposure_loss: 3.0376 - illumination_smoothness_loss: 2.5211 - spatial_constancy_loss: 4.6834e-07 - total_loss: 5.5601 WARNING: All log messages before absl::InitializeLog() is called are written to STDERR I0000 00:00:1700510538.106578 3409375 device_compiler.h:187] Compiled cluster using XLA! This line is logged at most once for the lifetime of the process. 25/25 ━━━━━━━━━━━━━━━━━━━━ 16s 123ms/step - color_constancy_loss: 0.0029 - exposure_loss: 2.9968 - illumination_smoothness_loss: 2.1813 - spatial_constancy_loss: 1.8559e-06 - total_loss: 5.1810 - val_color_constancy_loss: 0.0023 - val_exposure_loss: 2.9489 - val_illumination_smoothness_loss: 2.7063 - val_spatial_constancy_loss: 5.0979e-06 - val_total_loss: 5.6575 Epoch 2/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0030 - exposure_loss: 2.9854 - illumination_smoothness_loss: 1.2876 - spatial_constancy_loss: 6.1811e-06 - total_loss: 4.2759 - val_color_constancy_loss: 0.0023 - val_exposure_loss: 2.9381 - val_illumination_smoothness_loss: 1.8299 - val_spatial_constancy_loss: 1.3742e-05 - val_total_loss: 4.7703 Epoch 3/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0031 - exposure_loss: 2.9746 - illumination_smoothness_loss: 0.8735 - spatial_constancy_loss: 1.6664e-05 - total_loss: 3.8512 - val_color_constancy_loss: 0.0024 - val_exposure_loss: 2.9255 - val_illumination_smoothness_loss: 1.3135 - val_spatial_constancy_loss: 3.1783e-05 - val_total_loss: 4.2414 Epoch 4/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0032 - exposure_loss: 2.9623 - illumination_smoothness_loss: 0.6259 - spatial_constancy_loss: 3.7938e-05 - total_loss: 3.5914 - val_color_constancy_loss: 0.0025 - val_exposure_loss: 2.9118 - val_illumination_smoothness_loss: 0.9835 - val_spatial_constancy_loss: 6.1902e-05 - val_total_loss: 3.8979 Epoch 5/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0033 - exposure_loss: 2.9493 - illumination_smoothness_loss: 0.4700 - spatial_constancy_loss: 7.2080e-05 - total_loss: 3.4226 - val_color_constancy_loss: 0.0026 - val_exposure_loss: 2.8976 - val_illumination_smoothness_loss: 0.7751 - val_spatial_constancy_loss: 1.0500e-04 - val_total_loss: 3.6754 Epoch 6/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0034 - exposure_loss: 2.9358 - illumination_smoothness_loss: 0.3693 - spatial_constancy_loss: 1.1878e-04 - total_loss: 3.3086 - val_color_constancy_loss: 0.0027 - val_exposure_loss: 2.8829 - val_illumination_smoothness_loss: 0.6316 - val_spatial_constancy_loss: 1.6075e-04 - val_total_loss: 3.5173 Epoch 7/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0036 - exposure_loss: 2.9219 - illumination_smoothness_loss: 0.2996 - spatial_constancy_loss: 1.7723e-04 - total_loss: 3.2252 - val_color_constancy_loss: 0.0028 - val_exposure_loss: 2.8660 - val_illumination_smoothness_loss: 0.5261 - val_spatial_constancy_loss: 2.3790e-04 - val_total_loss: 3.3951 Epoch 8/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0037 - exposure_loss: 2.9056 - illumination_smoothness_loss: 0.2486 - spatial_constancy_loss: 2.5932e-04 - total_loss: 3.1582 - val_color_constancy_loss: 0.0029 - val_exposure_loss: 2.8466 - val_illumination_smoothness_loss: 0.4454 - val_spatial_constancy_loss: 3.4372e-04 - val_total_loss: 3.2952 Epoch 9/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0039 - exposure_loss: 2.8872 - illumination_smoothness_loss: 0.2110 - spatial_constancy_loss: 3.6800e-04 - total_loss: 3.1025 - val_color_constancy_loss: 0.0031 - val_exposure_loss: 2.8244 - val_illumination_smoothness_loss: 0.3853 - val_spatial_constancy_loss: 4.8290e-04 - val_total_loss: 3.2132 Epoch 10/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0041 - exposure_loss: 2.8665 - illumination_smoothness_loss: 0.1846 - spatial_constancy_loss: 5.0693e-04 - total_loss: 3.0558 - val_color_constancy_loss: 0.0033 - val_exposure_loss: 2.8002 - val_illumination_smoothness_loss: 0.3395 - val_spatial_constancy_loss: 6.5965e-04 - val_total_loss: 3.1436 Epoch 11/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0044 - exposure_loss: 2.8440 - illumination_smoothness_loss: 0.1654 - spatial_constancy_loss: 6.8036e-04 - total_loss: 3.0145 - val_color_constancy_loss: 0.0035 - val_exposure_loss: 2.7749 - val_illumination_smoothness_loss: 0.3031 - val_spatial_constancy_loss: 8.6824e-04 - val_total_loss: 3.0824 Epoch 12/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0047 - exposure_loss: 2.8198 - illumination_smoothness_loss: 0.1512 - spatial_constancy_loss: 8.9387e-04 - total_loss: 2.9765 - val_color_constancy_loss: 0.0038 - val_exposure_loss: 2.7463 - val_illumination_smoothness_loss: 0.2753 - val_spatial_constancy_loss: 0.0011 - val_total_loss: 3.0265 Epoch 13/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0050 - exposure_loss: 2.7928 - illumination_smoothness_loss: 0.1408 - spatial_constancy_loss: 0.0012 - total_loss: 2.9398 - val_color_constancy_loss: 0.0041 - val_exposure_loss: 2.7132 - val_illumination_smoothness_loss: 0.2537 - val_spatial_constancy_loss: 0.0015 - val_total_loss: 2.9724 Epoch 14/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0054 - exposure_loss: 2.7600 - illumination_smoothness_loss: 0.1340 - spatial_constancy_loss: 0.0016 - total_loss: 2.9009 - val_color_constancy_loss: 0.0045 - val_exposure_loss: 2.6673 - val_illumination_smoothness_loss: 0.2389 - val_spatial_constancy_loss: 0.0021 - val_total_loss: 2.9129 Epoch 15/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0060 - exposure_loss: 2.7115 - illumination_smoothness_loss: 0.1314 - spatial_constancy_loss: 0.0022 - total_loss: 2.8512 - val_color_constancy_loss: 0.0055 - val_exposure_loss: 2.5820 - val_illumination_smoothness_loss: 0.2374 - val_spatial_constancy_loss: 0.0035 - val_total_loss: 2.8284 Epoch 16/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0075 - exposure_loss: 2.6129 - illumination_smoothness_loss: 0.1414 - spatial_constancy_loss: 0.0041 - total_loss: 2.7660 - val_color_constancy_loss: 0.0081 - val_exposure_loss: 2.3797 - val_illumination_smoothness_loss: 0.2453 - val_spatial_constancy_loss: 0.0083 - val_total_loss: 2.6414 Epoch 17/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0128 - exposure_loss: 2.3149 - illumination_smoothness_loss: 0.1766 - spatial_constancy_loss: 0.0148 - total_loss: 2.5190 - val_color_constancy_loss: 0.0286 - val_exposure_loss: 1.5060 - val_illumination_smoothness_loss: 0.3288 - val_spatial_constancy_loss: 0.0648 - val_total_loss: 1.9282 Epoch 18/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0505 - exposure_loss: 1.3386 - illumination_smoothness_loss: 0.2606 - spatial_constancy_loss: 0.1196 - total_loss: 1.7693 - val_color_constancy_loss: 0.0827 - val_exposure_loss: 0.6645 - val_illumination_smoothness_loss: 0.2964 - val_spatial_constancy_loss: 0.2687 - val_total_loss: 1.3123 Epoch 19/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0873 - exposure_loss: 0.8174 - illumination_smoothness_loss: 0.2378 - spatial_constancy_loss: 0.2577 - total_loss: 1.4002 - val_color_constancy_loss: 0.0861 - val_exposure_loss: 0.6856 - val_illumination_smoothness_loss: 0.2464 - val_spatial_constancy_loss: 0.2539 - val_total_loss: 1.2719 Epoch 20/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0753 - exposure_loss: 0.8584 - illumination_smoothness_loss: 0.1858 - spatial_constancy_loss: 0.2394 - total_loss: 1.3589 - val_color_constancy_loss: 0.0882 - val_exposure_loss: 0.6714 - val_illumination_smoothness_loss: 0.2195 - val_spatial_constancy_loss: 0.2620 - val_total_loss: 1.2410 Epoch 21/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0779 - exposure_loss: 0.8382 - illumination_smoothness_loss: 0.1706 - spatial_constancy_loss: 0.2486 - total_loss: 1.3354 - val_color_constancy_loss: 0.0886 - val_exposure_loss: 0.6648 - val_illumination_smoothness_loss: 0.2072 - val_spatial_constancy_loss: 0.2643 - val_total_loss: 1.2249 Epoch 22/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0784 - exposure_loss: 0.8337 - illumination_smoothness_loss: 0.1590 - spatial_constancy_loss: 0.2502 - total_loss: 1.3212 - val_color_constancy_loss: 0.0889 - val_exposure_loss: 0.6647 - val_illumination_smoothness_loss: 0.1934 - val_spatial_constancy_loss: 0.2653 - val_total_loss: 1.2122 Epoch 23/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0783 - exposure_loss: 0.8329 - illumination_smoothness_loss: 0.1498 - spatial_constancy_loss: 0.2508 - total_loss: 1.3118 - val_color_constancy_loss: 0.0897 - val_exposure_loss: 0.6602 - val_illumination_smoothness_loss: 0.1834 - val_spatial_constancy_loss: 0.2671 - val_total_loss: 1.2003 Epoch 24/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0787 - exposure_loss: 0.8283 - illumination_smoothness_loss: 0.1426 - spatial_constancy_loss: 0.2529 - total_loss: 1.3025 - val_color_constancy_loss: 0.0897 - val_exposure_loss: 0.6601 - val_illumination_smoothness_loss: 0.1754 - val_spatial_constancy_loss: 0.2671 - val_total_loss: 1.1923 Epoch 25/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0785 - exposure_loss: 0.8294 - illumination_smoothness_loss: 0.1365 - spatial_constancy_loss: 0.2524 - total_loss: 1.2968 - val_color_constancy_loss: 0.0902 - val_exposure_loss: 0.6562 - val_illumination_smoothness_loss: 0.1672 - val_spatial_constancy_loss: 0.2692 - val_total_loss: 1.1828 Epoch 26/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0793 - exposure_loss: 0.8229 - illumination_smoothness_loss: 0.1316 - spatial_constancy_loss: 0.2554 - total_loss: 1.2892 - val_color_constancy_loss: 0.0896 - val_exposure_loss: 0.6567 - val_illumination_smoothness_loss: 0.1606 - val_spatial_constancy_loss: 0.2699 - val_total_loss: 1.1768 Epoch 27/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0788 - exposure_loss: 0.8285 - illumination_smoothness_loss: 0.1238 - spatial_constancy_loss: 0.2534 - total_loss: 1.2845 - val_color_constancy_loss: 0.0906 - val_exposure_loss: 0.6519 - val_illumination_smoothness_loss: 0.1574 - val_spatial_constancy_loss: 0.2725 - val_total_loss: 1.1724 Epoch 28/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0794 - exposure_loss: 0.8247 - illumination_smoothness_loss: 0.1194 - spatial_constancy_loss: 0.2550 - total_loss: 1.2785 - val_color_constancy_loss: 0.0914 - val_exposure_loss: 0.6451 - val_illumination_smoothness_loss: 0.1542 - val_spatial_constancy_loss: 0.2783 - val_total_loss: 1.1689 Epoch 29/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0797 - exposure_loss: 0.8203 - illumination_smoothness_loss: 0.1139 - spatial_constancy_loss: 0.2577 - total_loss: 1.2715 - val_color_constancy_loss: 0.0914 - val_exposure_loss: 0.6468 - val_illumination_smoothness_loss: 0.1435 - val_spatial_constancy_loss: 0.2775 - val_total_loss: 1.1592 Epoch 30/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0795 - exposure_loss: 0.8199 - illumination_smoothness_loss: 0.1083 - spatial_constancy_loss: 0.2581 - total_loss: 1.2659 - val_color_constancy_loss: 0.0911 - val_exposure_loss: 0.6483 - val_illumination_smoothness_loss: 0.1336 - val_spatial_constancy_loss: 0.2768 - val_total_loss: 1.1498 Epoch 31/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0797 - exposure_loss: 0.8194 - illumination_smoothness_loss: 0.1037 - spatial_constancy_loss: 0.2589 - total_loss: 1.2617 - val_color_constancy_loss: 0.0912 - val_exposure_loss: 0.6483 - val_illumination_smoothness_loss: 0.1289 - val_spatial_constancy_loss: 0.2772 - val_total_loss: 1.1456 Epoch 32/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0794 - exposure_loss: 0.8226 - illumination_smoothness_loss: 0.0982 - spatial_constancy_loss: 0.2578 - total_loss: 1.2580 - val_color_constancy_loss: 0.0923 - val_exposure_loss: 0.6421 - val_illumination_smoothness_loss: 0.1251 - val_spatial_constancy_loss: 0.2814 - val_total_loss: 1.1409 Epoch 33/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0801 - exposure_loss: 0.8188 - illumination_smoothness_loss: 0.0939 - spatial_constancy_loss: 0.2601 - total_loss: 1.2529 - val_color_constancy_loss: 0.0934 - val_exposure_loss: 0.6367 - val_illumination_smoothness_loss: 0.1261 - val_spatial_constancy_loss: 0.2853 - val_total_loss: 1.1416 Epoch 34/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0802 - exposure_loss: 0.8173 - illumination_smoothness_loss: 0.0889 - spatial_constancy_loss: 0.2611 - total_loss: 1.2475 - val_color_constancy_loss: 0.0941 - val_exposure_loss: 0.6326 - val_illumination_smoothness_loss: 0.1227 - val_spatial_constancy_loss: 0.2883 - val_total_loss: 1.1378 Epoch 35/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0807 - exposure_loss: 0.8134 - illumination_smoothness_loss: 0.0844 - spatial_constancy_loss: 0.2632 - total_loss: 1.2418 - val_color_constancy_loss: 0.0946 - val_exposure_loss: 0.6312 - val_illumination_smoothness_loss: 0.1180 - val_spatial_constancy_loss: 0.2893 - val_total_loss: 1.1330 Epoch 36/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0808 - exposure_loss: 0.8119 - illumination_smoothness_loss: 0.0798 - spatial_constancy_loss: 0.2644 - total_loss: 1.2368 - val_color_constancy_loss: 0.0941 - val_exposure_loss: 0.6351 - val_illumination_smoothness_loss: 0.1096 - val_spatial_constancy_loss: 0.2865 - val_total_loss: 1.1253 Epoch 37/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0807 - exposure_loss: 0.8127 - illumination_smoothness_loss: 0.0759 - spatial_constancy_loss: 0.2637 - total_loss: 1.2330 - val_color_constancy_loss: 0.0949 - val_exposure_loss: 0.6295 - val_illumination_smoothness_loss: 0.1088 - val_spatial_constancy_loss: 0.2904 - val_total_loss: 1.1237 Epoch 38/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0812 - exposure_loss: 0.8091 - illumination_smoothness_loss: 0.0732 - spatial_constancy_loss: 0.2658 - total_loss: 1.2293 - val_color_constancy_loss: 0.0946 - val_exposure_loss: 0.6313 - val_illumination_smoothness_loss: 0.1022 - val_spatial_constancy_loss: 0.2893 - val_total_loss: 1.1174 Epoch 39/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0810 - exposure_loss: 0.8100 - illumination_smoothness_loss: 0.0694 - spatial_constancy_loss: 0.2655 - total_loss: 1.2259 - val_color_constancy_loss: 0.0953 - val_exposure_loss: 0.6278 - val_illumination_smoothness_loss: 0.1015 - val_spatial_constancy_loss: 0.2918 - val_total_loss: 1.1164 Epoch 40/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0813 - exposure_loss: 0.8077 - illumination_smoothness_loss: 0.0668 - spatial_constancy_loss: 0.2668 - total_loss: 1.2226 - val_color_constancy_loss: 0.0951 - val_exposure_loss: 0.6294 - val_illumination_smoothness_loss: 0.0950 - val_spatial_constancy_loss: 0.2907 - val_total_loss: 1.1103 Epoch 41/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0814 - exposure_loss: 0.8074 - illumination_smoothness_loss: 0.0639 - spatial_constancy_loss: 0.2669 - total_loss: 1.2195 - val_color_constancy_loss: 0.0955 - val_exposure_loss: 0.6263 - val_illumination_smoothness_loss: 0.0946 - val_spatial_constancy_loss: 0.2930 - val_total_loss: 1.1093 Epoch 42/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0816 - exposure_loss: 0.8056 - illumination_smoothness_loss: 0.0613 - spatial_constancy_loss: 0.2684 - total_loss: 1.2168 - val_color_constancy_loss: 0.0950 - val_exposure_loss: 0.6304 - val_illumination_smoothness_loss: 0.0876 - val_spatial_constancy_loss: 0.2900 - val_total_loss: 1.1031 Epoch 43/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0813 - exposure_loss: 0.8074 - illumination_smoothness_loss: 0.0582 - spatial_constancy_loss: 0.2671 - total_loss: 1.2140 - val_color_constancy_loss: 0.0953 - val_exposure_loss: 0.6271 - val_illumination_smoothness_loss: 0.0859 - val_spatial_constancy_loss: 0.2925 - val_total_loss: 1.1008 Epoch 44/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0816 - exposure_loss: 0.8048 - illumination_smoothness_loss: 0.0564 - spatial_constancy_loss: 0.2687 - total_loss: 1.2115 - val_color_constancy_loss: 0.0956 - val_exposure_loss: 0.6266 - val_illumination_smoothness_loss: 0.0837 - val_spatial_constancy_loss: 0.2930 - val_total_loss: 1.0988 Epoch 45/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0816 - exposure_loss: 0.8045 - illumination_smoothness_loss: 0.0541 - spatial_constancy_loss: 0.2690 - total_loss: 1.2093 - val_color_constancy_loss: 0.0955 - val_exposure_loss: 0.6275 - val_illumination_smoothness_loss: 0.0796 - val_spatial_constancy_loss: 0.2923 - val_total_loss: 1.0949 Epoch 46/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0816 - exposure_loss: 0.8043 - illumination_smoothness_loss: 0.0517 - spatial_constancy_loss: 0.2691 - total_loss: 1.2067 - val_color_constancy_loss: 0.0959 - val_exposure_loss: 0.6245 - val_illumination_smoothness_loss: 0.0790 - val_spatial_constancy_loss: 0.2945 - val_total_loss: 1.0939 Epoch 47/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0819 - exposure_loss: 0.8025 - illumination_smoothness_loss: 0.0505 - spatial_constancy_loss: 0.2701 - total_loss: 1.2050 - val_color_constancy_loss: 0.0960 - val_exposure_loss: 0.6242 - val_illumination_smoothness_loss: 0.0764 - val_spatial_constancy_loss: 0.2949 - val_total_loss: 1.0914 Epoch 48/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0819 - exposure_loss: 0.8021 - illumination_smoothness_loss: 0.0482 - spatial_constancy_loss: 0.2706 - total_loss: 1.2027 - val_color_constancy_loss: 0.0957 - val_exposure_loss: 0.6262 - val_illumination_smoothness_loss: 0.0721 - val_spatial_constancy_loss: 0.2934 - val_total_loss: 1.0874 Epoch 49/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0818 - exposure_loss: 0.8027 - illumination_smoothness_loss: 0.0463 - spatial_constancy_loss: 0.2702 - total_loss: 1.2010 - val_color_constancy_loss: 0.0959 - val_exposure_loss: 0.6244 - val_illumination_smoothness_loss: 0.0712 - val_spatial_constancy_loss: 0.2947 - val_total_loss: 1.0863 Epoch 50/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0820 - exposure_loss: 0.8015 - illumination_smoothness_loss: 0.0446 - spatial_constancy_loss: 0.2711 - total_loss: 1.1992 - val_color_constancy_loss: 0.0959 - val_exposure_loss: 0.6248 - val_illumination_smoothness_loss: 0.0688 - val_spatial_constancy_loss: 0.2945 - val_total_loss: 1.0839 Epoch 51/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0819 - exposure_loss: 0.8019 - illumination_smoothness_loss: 0.0429 - spatial_constancy_loss: 0.2707 - total_loss: 1.1974 - val_color_constancy_loss: 0.0964 - val_exposure_loss: 0.6224 - val_illumination_smoothness_loss: 0.0677 - val_spatial_constancy_loss: 0.2964 - val_total_loss: 1.0829 Epoch 52/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0823 - exposure_loss: 0.7996 - illumination_smoothness_loss: 0.0416 - spatial_constancy_loss: 0.2721 - total_loss: 1.1955 - val_color_constancy_loss: 0.0958 - val_exposure_loss: 0.6240 - val_illumination_smoothness_loss: 0.0644 - val_spatial_constancy_loss: 0.2951 - val_total_loss: 1.0793 Epoch 53/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0822 - exposure_loss: 0.8004 - illumination_smoothness_loss: 0.0399 - spatial_constancy_loss: 0.2717 - total_loss: 1.1941 - val_color_constancy_loss: 0.0960 - val_exposure_loss: 0.6234 - val_illumination_smoothness_loss: 0.0633 - val_spatial_constancy_loss: 0.2957 - val_total_loss: 1.0785 Epoch 54/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0823 - exposure_loss: 0.7997 - illumination_smoothness_loss: 0.0382 - spatial_constancy_loss: 0.2723 - total_loss: 1.1924 - val_color_constancy_loss: 0.0959 - val_exposure_loss: 0.6242 - val_illumination_smoothness_loss: 0.0591 - val_spatial_constancy_loss: 0.2951 - val_total_loss: 1.0744 Epoch 55/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0822 - exposure_loss: 0.7999 - illumination_smoothness_loss: 0.0362 - spatial_constancy_loss: 0.2721 - total_loss: 1.1904 - val_color_constancy_loss: 0.0965 - val_exposure_loss: 0.6211 - val_illumination_smoothness_loss: 0.0603 - val_spatial_constancy_loss: 0.2974 - val_total_loss: 1.0754 Epoch 56/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0825 - exposure_loss: 0.7983 - illumination_smoothness_loss: 0.0351 - spatial_constancy_loss: 0.2732 - total_loss: 1.1890 - val_color_constancy_loss: 0.0960 - val_exposure_loss: 0.6237 - val_illumination_smoothness_loss: 0.0547 - val_spatial_constancy_loss: 0.2955 - val_total_loss: 1.0699 Epoch 57/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0823 - exposure_loss: 0.7987 - illumination_smoothness_loss: 0.0331 - spatial_constancy_loss: 0.2730 - total_loss: 1.1871 - val_color_constancy_loss: 0.0963 - val_exposure_loss: 0.6236 - val_illumination_smoothness_loss: 0.0540 - val_spatial_constancy_loss: 0.2956 - val_total_loss: 1.0694 Epoch 58/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0823 - exposure_loss: 0.7990 - illumination_smoothness_loss: 0.0319 - spatial_constancy_loss: 0.2727 - total_loss: 1.1859 - val_color_constancy_loss: 0.0965 - val_exposure_loss: 0.6210 - val_illumination_smoothness_loss: 0.0537 - val_spatial_constancy_loss: 0.2976 - val_total_loss: 1.0688 Epoch 59/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0826 - exposure_loss: 0.7969 - illumination_smoothness_loss: 0.0315 - spatial_constancy_loss: 0.2740 - total_loss: 1.1850 - val_color_constancy_loss: 0.0966 - val_exposure_loss: 0.6208 - val_illumination_smoothness_loss: 0.0530 - val_spatial_constancy_loss: 0.2978 - val_total_loss: 1.0682 Epoch 60/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0824 - exposure_loss: 0.7971 - illumination_smoothness_loss: 0.0304 - spatial_constancy_loss: 0.2740 - total_loss: 1.1840 - val_color_constancy_loss: 0.0966 - val_exposure_loss: 0.6206 - val_illumination_smoothness_loss: 0.0516 - val_spatial_constancy_loss: 0.2979 - val_total_loss: 1.0667 Epoch 61/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0825 - exposure_loss: 0.7969 - illumination_smoothness_loss: 0.0295 - spatial_constancy_loss: 0.2741 - total_loss: 1.1829 - val_color_constancy_loss: 0.0969 - val_exposure_loss: 0.6194 - val_illumination_smoothness_loss: 0.0506 - val_spatial_constancy_loss: 0.2988 - val_total_loss: 1.0657 Epoch 62/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7954 - illumination_smoothness_loss: 0.0287 - spatial_constancy_loss: 0.2749 - total_loss: 1.1817 - val_color_constancy_loss: 0.0967 - val_exposure_loss: 0.6203 - val_illumination_smoothness_loss: 0.0494 - val_spatial_constancy_loss: 0.2981 - val_total_loss: 1.0644 Epoch 63/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0825 - exposure_loss: 0.7966 - illumination_smoothness_loss: 0.0278 - spatial_constancy_loss: 0.2742 - total_loss: 1.1810 - val_color_constancy_loss: 0.0971 - val_exposure_loss: 0.6184 - val_illumination_smoothness_loss: 0.0491 - val_spatial_constancy_loss: 0.2996 - val_total_loss: 1.0642 Epoch 64/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 67ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7949 - illumination_smoothness_loss: 0.0268 - spatial_constancy_loss: 0.2753 - total_loss: 1.1797 - val_color_constancy_loss: 0.0969 - val_exposure_loss: 0.6199 - val_illumination_smoothness_loss: 0.0460 - val_spatial_constancy_loss: 0.2984 - val_total_loss: 1.0611 Epoch 65/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0826 - exposure_loss: 0.7957 - illumination_smoothness_loss: 0.0254 - spatial_constancy_loss: 0.2748 - total_loss: 1.1785 - val_color_constancy_loss: 0.0976 - val_exposure_loss: 0.6180 - val_illumination_smoothness_loss: 0.0464 - val_spatial_constancy_loss: 0.2998 - val_total_loss: 1.0618 Epoch 66/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7948 - illumination_smoothness_loss: 0.0249 - spatial_constancy_loss: 0.2753 - total_loss: 1.1777 - val_color_constancy_loss: 0.0975 - val_exposure_loss: 0.6189 - val_illumination_smoothness_loss: 0.0448 - val_spatial_constancy_loss: 0.2991 - val_total_loss: 1.0602 Epoch 67/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0825 - exposure_loss: 0.7954 - illumination_smoothness_loss: 0.0241 - spatial_constancy_loss: 0.2750 - total_loss: 1.1770 - val_color_constancy_loss: 0.0977 - val_exposure_loss: 0.6179 - val_illumination_smoothness_loss: 0.0441 - val_spatial_constancy_loss: 0.2998 - val_total_loss: 1.0595 Epoch 68/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7946 - illumination_smoothness_loss: 0.0231 - spatial_constancy_loss: 0.2757 - total_loss: 1.1761 - val_color_constancy_loss: 0.0973 - val_exposure_loss: 0.6198 - val_illumination_smoothness_loss: 0.0410 - val_spatial_constancy_loss: 0.2980 - val_total_loss: 1.0562 Epoch 69/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0826 - exposure_loss: 0.7947 - illumination_smoothness_loss: 0.0226 - spatial_constancy_loss: 0.2752 - total_loss: 1.1752 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6170 - val_illumination_smoothness_loss: 0.0435 - val_spatial_constancy_loss: 0.3003 - val_total_loss: 1.0587 Epoch 70/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7940 - illumination_smoothness_loss: 0.0224 - spatial_constancy_loss: 0.2758 - total_loss: 1.1749 - val_color_constancy_loss: 0.0976 - val_exposure_loss: 0.6182 - val_illumination_smoothness_loss: 0.0414 - val_spatial_constancy_loss: 0.2994 - val_total_loss: 1.0566 Epoch 71/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7941 - illumination_smoothness_loss: 0.0216 - spatial_constancy_loss: 0.2758 - total_loss: 1.1742 - val_color_constancy_loss: 0.0974 - val_exposure_loss: 0.6189 - val_illumination_smoothness_loss: 0.0389 - val_spatial_constancy_loss: 0.2986 - val_total_loss: 1.0538 Epoch 72/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7941 - illumination_smoothness_loss: 0.0211 - spatial_constancy_loss: 0.2755 - total_loss: 1.1734 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6166 - val_illumination_smoothness_loss: 0.0420 - val_spatial_constancy_loss: 0.3005 - val_total_loss: 1.0571 Epoch 73/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7935 - illumination_smoothness_loss: 0.0214 - spatial_constancy_loss: 0.2759 - total_loss: 1.1735 - val_color_constancy_loss: 0.0977 - val_exposure_loss: 0.6172 - val_illumination_smoothness_loss: 0.0401 - val_spatial_constancy_loss: 0.3001 - val_total_loss: 1.0551 Epoch 74/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7935 - illumination_smoothness_loss: 0.0205 - spatial_constancy_loss: 0.2760 - total_loss: 1.1727 - val_color_constancy_loss: 0.0978 - val_exposure_loss: 0.6168 - val_illumination_smoothness_loss: 0.0395 - val_spatial_constancy_loss: 0.3005 - val_total_loss: 1.0546 Epoch 75/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7924 - illumination_smoothness_loss: 0.0204 - spatial_constancy_loss: 0.2764 - total_loss: 1.1721 - val_color_constancy_loss: 0.0977 - val_exposure_loss: 0.6176 - val_illumination_smoothness_loss: 0.0385 - val_spatial_constancy_loss: 0.2997 - val_total_loss: 1.0536 Epoch 76/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7933 - illumination_smoothness_loss: 0.0198 - spatial_constancy_loss: 0.2760 - total_loss: 1.1718 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6166 - val_illumination_smoothness_loss: 0.0376 - val_spatial_constancy_loss: 0.3002 - val_total_loss: 1.0524 Epoch 77/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7925 - illumination_smoothness_loss: 0.0195 - spatial_constancy_loss: 0.2763 - total_loss: 1.1710 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6170 - val_illumination_smoothness_loss: 0.0384 - val_spatial_constancy_loss: 0.2999 - val_total_loss: 1.0532 Epoch 78/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0827 - exposure_loss: 0.7929 - illumination_smoothness_loss: 0.0196 - spatial_constancy_loss: 0.2761 - total_loss: 1.1713 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6170 - val_illumination_smoothness_loss: 0.0369 - val_spatial_constancy_loss: 0.3000 - val_total_loss: 1.0518 Epoch 79/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7922 - illumination_smoothness_loss: 0.0192 - spatial_constancy_loss: 0.2763 - total_loss: 1.1704 - val_color_constancy_loss: 0.0981 - val_exposure_loss: 0.6157 - val_illumination_smoothness_loss: 0.0380 - val_spatial_constancy_loss: 0.3009 - val_total_loss: 1.0527 Epoch 80/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7918 - illumination_smoothness_loss: 0.0191 - spatial_constancy_loss: 0.2766 - total_loss: 1.1703 - val_color_constancy_loss: 0.0980 - val_exposure_loss: 0.6159 - val_illumination_smoothness_loss: 0.0373 - val_spatial_constancy_loss: 0.3004 - val_total_loss: 1.0516 Epoch 81/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7917 - illumination_smoothness_loss: 0.0190 - spatial_constancy_loss: 0.2764 - total_loss: 1.1699 - val_color_constancy_loss: 0.0981 - val_exposure_loss: 0.6153 - val_illumination_smoothness_loss: 0.0373 - val_spatial_constancy_loss: 0.3009 - val_total_loss: 1.0516 Epoch 82/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 66ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7915 - illumination_smoothness_loss: 0.0187 - spatial_constancy_loss: 0.2766 - total_loss: 1.1697 - val_color_constancy_loss: 0.0979 - val_exposure_loss: 0.6170 - val_illumination_smoothness_loss: 0.0348 - val_spatial_constancy_loss: 0.2996 - val_total_loss: 1.0493 Epoch 83/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7918 - illumination_smoothness_loss: 0.0182 - spatial_constancy_loss: 0.2763 - total_loss: 1.1691 - val_color_constancy_loss: 0.0980 - val_exposure_loss: 0.6158 - val_illumination_smoothness_loss: 0.0358 - val_spatial_constancy_loss: 0.3004 - val_total_loss: 1.0500 Epoch 84/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7911 - illumination_smoothness_loss: 0.0184 - spatial_constancy_loss: 0.2766 - total_loss: 1.1689 - val_color_constancy_loss: 0.0982 - val_exposure_loss: 0.6146 - val_illumination_smoothness_loss: 0.0366 - val_spatial_constancy_loss: 0.3010 - val_total_loss: 1.0505 Epoch 85/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7907 - illumination_smoothness_loss: 0.0185 - spatial_constancy_loss: 0.2767 - total_loss: 1.1687 - val_color_constancy_loss: 0.0980 - val_exposure_loss: 0.6154 - val_illumination_smoothness_loss: 0.0361 - val_spatial_constancy_loss: 0.3006 - val_total_loss: 1.0501 Epoch 86/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0828 - exposure_loss: 0.7910 - illumination_smoothness_loss: 0.0182 - spatial_constancy_loss: 0.2765 - total_loss: 1.1685 - val_color_constancy_loss: 0.0982 - val_exposure_loss: 0.6145 - val_illumination_smoothness_loss: 0.0356 - val_spatial_constancy_loss: 0.3009 - val_total_loss: 1.0492 Epoch 87/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7902 - illumination_smoothness_loss: 0.0181 - spatial_constancy_loss: 0.2767 - total_loss: 1.1680 - val_color_constancy_loss: 0.0981 - val_exposure_loss: 0.6149 - val_illumination_smoothness_loss: 0.0357 - val_spatial_constancy_loss: 0.3007 - val_total_loss: 1.0494 Epoch 88/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7904 - illumination_smoothness_loss: 0.0180 - spatial_constancy_loss: 0.2766 - total_loss: 1.1679 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6133 - val_illumination_smoothness_loss: 0.0359 - val_spatial_constancy_loss: 0.3015 - val_total_loss: 1.0491 Epoch 89/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0830 - exposure_loss: 0.7893 - illumination_smoothness_loss: 0.0181 - spatial_constancy_loss: 0.2770 - total_loss: 1.1674 - val_color_constancy_loss: 0.0981 - val_exposure_loss: 0.6148 - val_illumination_smoothness_loss: 0.0350 - val_spatial_constancy_loss: 0.3006 - val_total_loss: 1.0484 Epoch 90/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7901 - illumination_smoothness_loss: 0.0178 - spatial_constancy_loss: 0.2765 - total_loss: 1.1673 - val_color_constancy_loss: 0.0984 - val_exposure_loss: 0.6128 - val_illumination_smoothness_loss: 0.0358 - val_spatial_constancy_loss: 0.3017 - val_total_loss: 1.0487 Epoch 91/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0831 - exposure_loss: 0.7886 - illumination_smoothness_loss: 0.0181 - spatial_constancy_loss: 0.2771 - total_loss: 1.1669 - val_color_constancy_loss: 0.0981 - val_exposure_loss: 0.6142 - val_illumination_smoothness_loss: 0.0351 - val_spatial_constancy_loss: 0.3007 - val_total_loss: 1.0481 Epoch 92/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0829 - exposure_loss: 0.7895 - illumination_smoothness_loss: 0.0177 - spatial_constancy_loss: 0.2766 - total_loss: 1.1668 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6133 - val_illumination_smoothness_loss: 0.0349 - val_spatial_constancy_loss: 0.3011 - val_total_loss: 1.0476 Epoch 93/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0831 - exposure_loss: 0.7884 - illumination_smoothness_loss: 0.0179 - spatial_constancy_loss: 0.2770 - total_loss: 1.1664 - val_color_constancy_loss: 0.0984 - val_exposure_loss: 0.6125 - val_illumination_smoothness_loss: 0.0355 - val_spatial_constancy_loss: 0.3014 - val_total_loss: 1.0478 Epoch 94/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 65ms/step - color_constancy_loss: 0.0831 - exposure_loss: 0.7882 - illumination_smoothness_loss: 0.0181 - spatial_constancy_loss: 0.2769 - total_loss: 1.1663 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6128 - val_illumination_smoothness_loss: 0.0349 - val_spatial_constancy_loss: 0.3012 - val_total_loss: 1.0473 Epoch 95/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0831 - exposure_loss: 0.7881 - illumination_smoothness_loss: 0.0179 - spatial_constancy_loss: 0.2770 - total_loss: 1.1660 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6130 - val_illumination_smoothness_loss: 0.0341 - val_spatial_constancy_loss: 0.3009 - val_total_loss: 1.0462 Epoch 96/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0832 - exposure_loss: 0.7874 - illumination_smoothness_loss: 0.0179 - spatial_constancy_loss: 0.2771 - total_loss: 1.1656 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6125 - val_illumination_smoothness_loss: 0.0353 - val_spatial_constancy_loss: 0.3010 - val_total_loss: 1.0471 Epoch 97/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0830 - exposure_loss: 0.7882 - illumination_smoothness_loss: 0.0181 - spatial_constancy_loss: 0.2765 - total_loss: 1.1658 - val_color_constancy_loss: 0.0984 - val_exposure_loss: 0.6120 - val_illumination_smoothness_loss: 0.0346 - val_spatial_constancy_loss: 0.3014 - val_total_loss: 1.0464 Epoch 98/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 63ms/step - color_constancy_loss: 0.0832 - exposure_loss: 0.7869 - illumination_smoothness_loss: 0.0180 - spatial_constancy_loss: 0.2772 - total_loss: 1.1653 - val_color_constancy_loss: 0.0984 - val_exposure_loss: 0.6118 - val_illumination_smoothness_loss: 0.0344 - val_spatial_constancy_loss: 0.3012 - val_total_loss: 1.0458 Epoch 99/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0832 - exposure_loss: 0.7863 - illumination_smoothness_loss: 0.0182 - spatial_constancy_loss: 0.2772 - total_loss: 1.1650 - val_color_constancy_loss: 0.0983 - val_exposure_loss: 0.6120 - val_illumination_smoothness_loss: 0.0343 - val_spatial_constancy_loss: 0.3007 - val_total_loss: 1.0453 Epoch 100/100 25/25 ━━━━━━━━━━━━━━━━━━━━ 2s 64ms/step - color_constancy_loss: 0.0831 - exposure_loss: 0.7873 - illumination_smoothness_loss: 0.0180 - spatial_constancy_loss: 0.2765 - total_loss: 1.1649 - val_color_constancy_loss: 0.0984 - val_exposure_loss: 0.6115 - val_illumination_smoothness_loss: 0.0341 - val_spatial_constancy_loss: 0.3011 - val_total_loss: 1.0451 </code></pre></div> </div> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_21_3.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_21_4.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_21_5.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_21_6.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_21_7.png" /> <hr /> <h2 id="inference">Inference</h2> <div class="codehilite"><pre><code>def plot_results(images, titles, figure_size=(12, 12)): fig = plt.figure(figsize=figure_size) for i in range(len(images)): fig.add_subplot(1, len(images), i + 1).set_title(titles[i]) _ = plt.imshow(images[i]) plt.axis("off") plt.show() def infer(original_image): image = keras.utils.img_to_array(original_image) image = image.astype("float32") / 255.0 image = np.expand_dims(image, axis=0) output_image = zero_dce_model(image) output_image = tf.cast((output_image[0, :, :, :] * 255), dtype=np.uint8) output_image = Image.fromarray(output_image.numpy()) return output_image </code></pre></div> <h3 id="inference-on-test-images">Inference on test images</h3> We compare the test images from LOLDataset enhanced by MIRNet with images enhanced via the <code>PIL.ImageOps.autocontrast()</code> function. You can use the trained model hosted on <a href="https://huggingface.co/keras-io/low-light-image-enhancement">Hugging Face Hub</a> and try the demo on <a href="https://huggingface.co/spaces/keras-io/low-light-image-enhancement">Hugging Face Spaces</a>. <div class="codehilite"><pre><code>for val_image_file in test_low_light_images: original_image = Image.open(val_image_file) enhanced_image = infer(original_image) plot_results( [original_image, ImageOps.autocontrast(original_image), enhanced_image], ["Original", "PIL Autocontrast", "Enhanced"], (20, 12), ) </code></pre></div> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_25_0.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_25_1.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_25_2.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_25_3.png" /> <img alt="png" src="/img/examples/vision/zero_dce/zero_dce_25_4.png" /> </div> <div class='k-outline'> <div class='k-outline-depth-1'> <a href='#zerodce-for-lowlight-image-enhancement'>Zero-DCE for low-light image enhancement</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#introduction'>Introduction</a> </div> <div class='k-outline-depth-3'> <a href='#references'>References</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#downloading-loldataset'>Downloading LOLDataset</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#creating-a-tensorflow-dataset'>Creating a TensorFlow Dataset</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#the-zerodce-framework'>The Zero-DCE Framework</a> </div> <div class='k-outline-depth-3'> <a href='#understanding-lightenhancement-curves'>Understanding light-enhancement curves</a> </div> <div class='k-outline-depth-3'> <a href='#dcenet'>DCE-Net</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#loss-functions'>Loss functions</a> </div> <div class='k-outline-depth-3'> <a href='#color-constancy-loss'>Color constancy loss</a> </div> <div class='k-outline-depth-3'> <a href='#exposure-loss'>Exposure loss</a> </div> <div class='k-outline-depth-3'> <a href='#illumination-smoothness-loss'>Illumination smoothness loss</a> </div> <div class='k-outline-depth-3'> <a href='#spatial-consistency-loss'>Spatial consistency loss</a> </div> <div class='k-outline-depth-3'> <a href='#deep-curve-estimation-model'>Deep curve estimation model</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#training'>Training</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#inference'>Inference</a> </div> <div class='k-outline-depth-3'> <a href='#inference-on-test-images'>Inference on test images</a> </div> </div> </div> </div> </div> </body> <footer style="float: left; width: 100%; padding: 1em; border-top: solid 1px #bbb;"> <a href="https://policies.google.com/terms">Terms</a> | <a href="https://policies.google.com/privacy">Privacy</a> </footer> </html>