CINXE.COM

<!DOCTYPE html> <html lang="en" dir="ltr"> <head>  <script async src="https://www.googletagmanager.com/gtag/js?id=G-P63WKM1TM1"></script> <script> window.dataLayer = window.dataLayer || []; function gtag(){dataLayer.push(arguments);} gtag('js', new Date()); gtag('config', 'G-P63WKM1TM1'); </script>  <script type="text/javascript" > (function(m,e,t,r,i,k,a){m[i]=m[i]||function(){(m[i].a=m[i].a||[]).push(arguments)}; m[i].l=1*new Date(); for (var j = 0; j < document.scripts.length; j++) {if (document.scripts[j].src === r) { return; }} k=e.createElement(t),a=e.getElementsByTagName(t)[0],k.async=1,k.src=r,a.parentNode.insertBefore(k,a)}) (window, document, "script", "https://mc.yandex.ru/metrika/tag.js", "ym"); ym(55165297, "init", { clickmap:false, trackLinks:true, accurateTrackBounce:true, webvisor:false }); </script> <noscript><div><img src="https://mc.yandex.ru/watch/55165297" style="position:absolute; left:-9999px;" alt="" /></div></noscript>    <title>Search results for: U-Net</title> <meta name="description" content="Search results for: U-Net"> <meta name="keywords" content="U-Net"> <meta name="viewport" content="width=device-width, initial-scale=1, minimum-scale=1, maximum-scale=1, user-scalable=no"> <meta charset="utf-8"> <link href="https://cdn.waset.org/favicon.ico" type="image/x-icon" rel="shortcut icon"> <link href="https://cdn.waset.org/static/plugins/bootstrap-4.2.1/css/bootstrap.min.css" rel="stylesheet"> <link href="https://cdn.waset.org/static/plugins/fontawesome/css/all.min.css" rel="stylesheet"> <link href="https://cdn.waset.org/static/css/site.css?v=150220211555" rel="stylesheet"> </head> <body> <header> <div class="container"> <nav class="navbar navbar-expand-lg navbar-light"> <a class="navbar-brand" href="https://waset.org"> <img src="https://cdn.waset.org/static/images/wasetc.png" alt="Open Science Research Excellence" title="Open Science Research Excellence" /> </a> <button class="d-block d-lg-none navbar-toggler ml-auto" type="button" data-toggle="collapse" data-target="#navbarMenu" aria-controls="navbarMenu" aria-expanded="false" aria-label="Toggle navigation"> <span class="navbar-toggler-icon"></span> </button> <div class="w-100"> <div class="d-none d-lg-flex flex-row-reverse"> <form method="get" action="https://waset.org/search" class="form-inline my-2 my-lg-0"> <input class="form-control mr-sm-2" type="search" placeholder="Search Conferences" value="U-Net" name="q" aria-label="Search"> <button class="btn btn-light my-2 my-sm-0" type="submit"><i class="fas fa-search"></i></button> </form> </div> <div class="collapse navbar-collapse mt-1" id="navbarMenu"> <ul class="navbar-nav ml-auto align-items-center" id="mainNavMenu"> <li class="nav-item"> <a class="nav-link" href="https://waset.org/conferences" title="Conferences in 2024/2025/2026">Conferences</a> </li> <li class="nav-item"> <a class="nav-link" href="https://waset.org/disciplines" title="Disciplines">Disciplines</a> </li> <li class="nav-item"> <a class="nav-link" href="https://waset.org/committees" rel="nofollow">Committees</a> </li> <li class="nav-item dropdown"> <a class="nav-link dropdown-toggle" href="#" id="navbarDropdownPublications" role="button" data-toggle="dropdown" aria-haspopup="true" aria-expanded="false"> Publications </a> <div class="dropdown-menu" aria-labelledby="navbarDropdownPublications"> <a class="dropdown-item" href="https://publications.waset.org/abstracts">Abstracts</a> <a class="dropdown-item" href="https://publications.waset.org">Periodicals</a> <a class="dropdown-item" href="https://publications.waset.org/archive">Archive</a> </div> </li> <li class="nav-item"> <a class="nav-link" href="https://waset.org/page/support" title="Support">Support</a> </li> </ul> </div> </div> </nav> </div> </header> <main> <div class="container mt-4"> <div class="row"> <div class="col-md-9 mx-auto"> <form method="get" action="https://publications.waset.org/abstracts/search"> <div id="custom-search-input"> <div class="input-group"> <i class="fas fa-search"></i> <input type="text" class="search-query" name="q" placeholder="Author, Title, Abstract, Keywords" value="U-Net"> <input type="submit" class="btn_search" value="Search"> </div> </div> </form> </div> </div> <div class="row mt-3"> <div class="col-sm-3"> <div class="card"> <div class="card-body"><strong>Commenced</strong> in January 2007</div> </div> </div> <div class="col-sm-3"> <div class="card"> <div class="card-body"><strong>Frequency:</strong> Monthly</div> </div> </div> <div class="col-sm-3"> <div class="card"> <div class="card-body"><strong>Edition:</strong> International</div> </div> </div> <div class="col-sm-3"> <div class="card"> <div class="card-body"><strong>Paper Count:</strong> 8</div> </div> </div> </div> <h1 class="mt-3 mb-3 text-center" style="font-size:1.6rem;">Search results for: U-Net</h1> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">8</span> A Distinct Method Based on Mamba-Unet for Brain Tumor Image Segmentation</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Djallel%20Bouamama">Djallel Bouamama</a>, <a href="https://publications.waset.org/abstracts/search?q=Yasser%20R.%20Haddadi"> Yasser R. Haddadi</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Accurate brain tumor segmentation is crucial for diagnosis and treatment planning, yet it remains a challenging task due to the variability in tumor shapes and intensities. This paper introduces a distinct approach to brain tumor image segmentation by leveraging an advanced architecture known as Mamba-Unet. Building on the well-established U-Net framework, Mamba-Unet incorporates distinct design enhancements to improve segmentation performance. Our proposed method integrates a multi-scale attention mechanism and a hybrid loss function to effectively capture fine-grained details and contextual information in brain MRI scans. We demonstrate that Mamba-Unet significantly enhances segmentation accuracy compared to conventional U-Net models by utilizing a comprehensive dataset of annotated brain MRI scans. Quantitative evaluations reveal that Mamba-Unet surpasses traditional U-Net architectures and other contemporary segmentation models regarding Dice coefficient, sensitivity, and specificity. The improvements are attributed to the method's ability to manage class imbalance better and resolve complex tumor boundaries. This work advances the state-of-the-art in brain tumor segmentation and holds promise for improving clinical workflows and patient outcomes through more precise and reliable tumor detection. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=brain%20tumor%20classification" title="brain tumor classification">brain tumor classification</a>, <a href="https://publications.waset.org/abstracts/search?q=image%20segmentation" title=" image segmentation"> image segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=CNN" title=" CNN"> CNN</a>, <a href="https://publications.waset.org/abstracts/search?q=U-NET" title=" U-NET"> U-NET</a> </p> <a href="https://publications.waset.org/abstracts/192073/a-distinct-method-based-on-mamba-unet-for-brain-tumor-image-segmentation" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/192073.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">34</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">7</span> COVID-19 Detection from Computed Tomography Images Using UNet Segmentation, Region Extraction, and Classification Pipeline</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Kenan%20Morani">Kenan Morani</a>, <a href="https://publications.waset.org/abstracts/search?q=Esra%20Kaya%20Ayana"> Esra Kaya Ayana</a> </p> <p class="card-text"><strong>Abstract:</strong></p> This study aimed to develop a novel pipeline for COVID-19 detection using a large and rigorously annotated database of computed tomography (CT) images. The pipeline consists of UNet-based segmentation, lung extraction, and a classification part, with the addition of optional slice removal techniques following the segmentation part. In this work, a batch normalization was added to the original UNet model to produce lighter and better localization, which is then utilized to build a full pipeline for COVID-19 diagnosis. To evaluate the effectiveness of the proposed pipeline, various segmentation methods were compared in terms of their performance and complexity. The proposed segmentation method with batch normalization outperformed traditional methods and other alternatives, resulting in a higher dice score on a publicly available dataset. Moreover, at the slice level, the proposed pipeline demonstrated high validation accuracy, indicating the efficiency of predicting 2D slices. At the patient level, the full approach exhibited higher validation accuracy and macro F1 score compared to other alternatives, surpassing the baseline. The classification component of the proposed pipeline utilizes a convolutional neural network (CNN) to make final diagnosis decisions. The COV19-CT-DB dataset, which contains a large number of CT scans with various types of slices and rigorously annotated for COVID-19 detection, was utilized for classification. The proposed pipeline outperformed many other alternatives on the dataset. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=classification" title="classification">classification</a>, <a href="https://publications.waset.org/abstracts/search?q=computed%20tomography" title=" computed tomography"> computed tomography</a>, <a href="https://publications.waset.org/abstracts/search?q=lung%20extraction" title=" lung extraction"> lung extraction</a>, <a href="https://publications.waset.org/abstracts/search?q=macro%20F1%20score" title=" macro F1 score"> macro F1 score</a>, <a href="https://publications.waset.org/abstracts/search?q=UNet%20segmentation" title=" UNet segmentation"> UNet segmentation</a> </p> <a href="https://publications.waset.org/abstracts/169737/covid-19-detection-from-computed-tomography-images-using-unet-segmentation-region-extraction-and-classification-pipeline" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/169737.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">131</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">6</span> Count of Trees in East Africa with Deep Learning</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Nubwimana%20Rachel">Nubwimana Rachel</a>, <a href="https://publications.waset.org/abstracts/search?q=Mugabowindekwe%20Maurice"> Mugabowindekwe Maurice</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Trees play a crucial role in maintaining biodiversity and providing various ecological services. Traditional methods of counting trees are time-consuming, and there is a need for more efficient techniques. However, deep learning makes it feasible to identify the multi-scale elements hidden in aerial imagery. This research focuses on the application of deep learning techniques for tree detection and counting in both forest and non-forest areas through the exploration of the deep learning application for automated tree detection and counting using satellite imagery. The objective is to identify the most effective model for automated tree counting. We used different deep learning models such as YOLOV7, SSD, and UNET, along with Generative Adversarial Networks to generate synthetic samples for training and other augmentation techniques, including Random Resized Crop, AutoAugment, and Linear Contrast Enhancement. These models were trained and fine-tuned using satellite imagery to identify and count trees. The performance of the models was assessed through multiple trials; after training and fine-tuning the models, UNET demonstrated the best performance with a validation loss of 0.1211, validation accuracy of 0.9509, and validation precision of 0.9799. This research showcases the success of deep learning in accurate tree counting through remote sensing, particularly with the UNET model. It represents a significant contribution to the field by offering an efficient and precise alternative to conventional tree-counting methods. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=remote%20sensing" title="remote sensing">remote sensing</a>, <a href="https://publications.waset.org/abstracts/search?q=deep%20learning" title=" deep learning"> deep learning</a>, <a href="https://publications.waset.org/abstracts/search?q=tree%20counting" title=" tree counting"> tree counting</a>, <a href="https://publications.waset.org/abstracts/search?q=image%20segmentation" title=" image segmentation"> image segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=object%20detection" title=" object detection"> object detection</a>, <a href="https://publications.waset.org/abstracts/search?q=visualization" title=" visualization"> visualization</a> </p> <a href="https://publications.waset.org/abstracts/177935/count-of-trees-in-east-africa-with-deep-learning" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/177935.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">71</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">5</span> Traffic Light Detection Using Image Segmentation</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Vaishnavi%20Shivde">Vaishnavi Shivde</a>, <a href="https://publications.waset.org/abstracts/search?q=Shrishti%20Sinha"> Shrishti Sinha</a>, <a href="https://publications.waset.org/abstracts/search?q=Trapti%20Mishra"> Trapti Mishra</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Traffic light detection from a moving vehicle is an important technology both for driver safety assistance functions as well as for autonomous driving in the city. This paper proposed a deep-learning-based traf铿乧 light recognition method that consists of a pixel-wise image segmentation technique and a fully convolutional network i.e., UNET architecture. This paper has used a method for detecting the position and recognizing the state of the traffic lights in video sequences is presented and evaluated using Traffic Light Dataset which contains masked traffic light image data. The first stage is the detection, which is accomplished through image processing (image segmentation) techniques such as image cropping, color transformation, segmentation of possible traffic lights. The second stage is the recognition, which means identifying the color of the traffic light or knowing the state of traffic light which is achieved by using a Convolutional Neural Network (UNET architecture). <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=traffic%20light%20detection" title="traffic light detection">traffic light detection</a>, <a href="https://publications.waset.org/abstracts/search?q=image%20segmentation" title=" image segmentation"> image segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=machine%20learning" title=" machine learning"> machine learning</a>, <a href="https://publications.waset.org/abstracts/search?q=classification" title=" classification"> classification</a>, <a href="https://publications.waset.org/abstracts/search?q=convolutional%20neural%20networks" title=" convolutional neural networks"> convolutional neural networks</a> </p> <a href="https://publications.waset.org/abstracts/137254/traffic-light-detection-using-image-segmentation" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/137254.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">173</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">4</span> Deep Supervision Based-Unet to Detect Buildings Changes from VHR Aerial Imagery</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Shimaa%20Holail">Shimaa Holail</a>, <a href="https://publications.waset.org/abstracts/search?q=Tamer%20Saleh"> Tamer Saleh</a>, <a href="https://publications.waset.org/abstracts/search?q=Xiongwu%20Xiao"> Xiongwu Xiao</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Building change detection (BCD) from satellite imagery is an essential topic in urbanization monitoring, agricultural land management, and updating geospatial databases. Recently, methods for detecting changes based on deep learning have made significant progress and impressive results. However, it has the problem of being insensitive to changes in buildings with complex spectral differences, and the features being extracted are not discriminatory enough, resulting in incomplete buildings and irregular boundaries. To overcome these problems, we propose a dual Siamese network based on the Unet model with the addition of a deep supervision strategy (DS) in this paper. This network consists of a backbone (encoder) based on ImageNet pre-training, a fusion block, and feature pyramid networks (FPN) to enhance the step-by-step information of the changing regions and obtain a more accurate BCD map. To train the proposed method, we created a new dataset (EGY-BCD) of high-resolution and multi-temporal aerial images captured over New Cairo in Egypt to detect building changes for this purpose. The experimental results showed that the proposed method is effective and performs well with the EGY-BCD dataset regarding the overall accuracy, F1-score, and mIoU, which were 91.6 %, 80.1 %, and 73.5 %, respectively. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=building%20change%20detection" title="building change detection">building change detection</a>, <a href="https://publications.waset.org/abstracts/search?q=deep%20supervision" title=" deep supervision"> deep supervision</a>, <a href="https://publications.waset.org/abstracts/search?q=semantic%20segmentation" title=" semantic segmentation"> semantic segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=EGY-BCD%20dataset" title=" EGY-BCD dataset"> EGY-BCD dataset</a> </p> <a href="https://publications.waset.org/abstracts/154993/deep-supervision-based-unet-to-detect-buildings-changes-from-vhr-aerial-imagery" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/154993.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">120</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">3</span> NFResNet: Multi-Scale and U-Shaped Networks for Deblurring</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Tanish%20Mittal">Tanish Mittal</a>, <a href="https://publications.waset.org/abstracts/search?q=Preyansh%20Agrawal"> Preyansh Agrawal</a>, <a href="https://publications.waset.org/abstracts/search?q=Esha%20Pahwa"> Esha Pahwa</a>, <a href="https://publications.waset.org/abstracts/search?q=Aarya%20Makwana"> Aarya Makwana</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Multi-Scale and U-shaped Networks are widely used in various image restoration problems, including deblurring. Keeping in mind the wide range of applications, we present a comparison of these architectures and their effects on image deblurring. We also introduce a new block called as NFResblock. It consists of a Fast Fourier Transformation layer and a series of modified Non-Linear Activation Free Blocks. Based on these architectures and additions, we introduce NFResnet and NFResnet+, which are modified multi-scale and U-Net architectures, respectively. We also use three differ-ent loss functions to train these architectures: Charbonnier Loss, Edge Loss, and Frequency Reconstruction Loss. Extensive experiments on the Deep Video Deblurring dataset, along with ablation studies for each component, have been presented in this paper. The proposed architectures achieve a considerable increase in Peak Signal to Noise (PSNR) ratio and Structural Similarity Index (SSIM) value. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=multi-scale" title="multi-scale">multi-scale</a>, <a href="https://publications.waset.org/abstracts/search?q=Unet" title=" Unet"> Unet</a>, <a href="https://publications.waset.org/abstracts/search?q=deblurring" title=" deblurring"> deblurring</a>, <a href="https://publications.waset.org/abstracts/search?q=FFT" title=" FFT"> FFT</a>, <a href="https://publications.waset.org/abstracts/search?q=resblock" title=" resblock"> resblock</a>, <a href="https://publications.waset.org/abstracts/search?q=NAF-block" title=" NAF-block"> NAF-block</a>, <a href="https://publications.waset.org/abstracts/search?q=nfresnet" title=" nfresnet"> nfresnet</a>, <a href="https://publications.waset.org/abstracts/search?q=charbonnier" title=" charbonnier"> charbonnier</a>, <a href="https://publications.waset.org/abstracts/search?q=edge" title=" edge"> edge</a>, <a href="https://publications.waset.org/abstracts/search?q=frequency%20reconstruction" title=" frequency reconstruction"> frequency reconstruction</a> </p> <a href="https://publications.waset.org/abstracts/159306/nfresnet-multi-scale-and-u-shaped-networks-for-deblurring" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/159306.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">136</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">2</span> Network Conditioning and Transfer Learning for Peripheral Nerve Segmentation in Ultrasound Images</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Harold%20Mauricio%20D%C3%ADaz-Vargas">Harold Mauricio D铆az-Vargas</a>, <a href="https://publications.waset.org/abstracts/search?q=Cristian%20Alfonso%20Jimenez-Casta%C3%B1o"> Cristian Alfonso Jimenez-Casta帽o</a>, <a href="https://publications.waset.org/abstracts/search?q=David%20Augusto%20C%C3%A1rdenas-Pe%C3%B1a"> David Augusto C谩rdenas-Pe帽a</a>, <a href="https://publications.waset.org/abstracts/search?q=Guillermo%20Alberto%20Ortiz-G%C3%B3mez"> Guillermo Alberto Ortiz-G贸mez</a>, <a href="https://publications.waset.org/abstracts/search?q=Alvaro%20Angel%20Orozco-Gutierrez"> Alvaro Angel Orozco-Gutierrez</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Precise identification of the nerves is a crucial task performed by anesthesiologists for an effective Peripheral Nerve Blocking (PNB). Now, anesthesiologists use ultrasound imaging equipment to guide the PNB and detect nervous structures. However, visual identification of the nerves from ultrasound images is difficult, even for trained specialists, due to artifacts and low contrast. The recent advances in deep learning make neural networks a potential tool for accurate nerve segmentation systems, so addressing the above issues from raw data. The most widely spread U-Net network yields pixel-by-pixel segmentation by encoding the input image and decoding the attained feature vector into a semantic image. This work proposes a conditioning approach and encoder pre-training to enhance the nerve segmentation of traditional U-Nets. Conditioning is achieved by the one-hot encoding of the kind of target nerve a the network input, while the pre-training considers five well-known deep networks for image classification. The proposed approach is tested in a collection of 619 US images, where the best C-UNet architecture yields an 81% Dice coefficient, outperforming the 74% of the best traditional U-Net. Results prove that pre-trained models with the conditional approach outperform their equivalent baseline by supporting learning new features and enriching the discriminant capability of the tested networks. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=nerve%20segmentation" title="nerve segmentation">nerve segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=U-Net" title=" U-Net"> U-Net</a>, <a href="https://publications.waset.org/abstracts/search?q=deep%20learning" title=" deep learning"> deep learning</a>, <a href="https://publications.waset.org/abstracts/search?q=ultrasound%20imaging" title=" ultrasound imaging"> ultrasound imaging</a>, <a href="https://publications.waset.org/abstracts/search?q=peripheral%20nerve%20blocking" title=" peripheral nerve blocking"> peripheral nerve blocking</a> </p> <a href="https://publications.waset.org/abstracts/152338/network-conditioning-and-transfer-learning-for-peripheral-nerve-segmentation-in-ultrasound-images" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/152338.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">106</span> </span> </div> </div> <div class="card paper-listing mb-3 mt-3"> <h5 class="card-header" style="font-size:.9rem"><span class="badge badge-info">1</span> Image Segmentation with Deep Learning of Prostate Cancer Bone Metastases on Computed Tomography</h5> <div class="card-body"> <p class="card-text"><strong>Authors:</strong> <a href="https://publications.waset.org/abstracts/search?q=Joseph%20M.%20Rich">Joseph M. Rich</a>, <a href="https://publications.waset.org/abstracts/search?q=Vinay%20A.%20Duddalwar"> Vinay A. Duddalwar</a>, <a href="https://publications.waset.org/abstracts/search?q=Assad%20A.%20Oberai"> Assad A. Oberai</a> </p> <p class="card-text"><strong>Abstract:</strong></p> Prostate adenocarcinoma is the most common cancer in males, with osseous metastases as the commonest site of metastatic prostate carcinoma (mPC). Treatment monitoring is based on the evaluation and characterization of lesions on multiple imaging studies, including Computed Tomography (CT). Monitoring of the osseous disease burden, including follow-up of lesions and identification and characterization of new lesions, is a laborious task for radiologists. Deep learning algorithms are increasingly used to perform tasks such as identification and segmentation for osseous metastatic disease and provide accurate information regarding metastatic burden. Here, nnUNet was used to produce a model which can segment CT scan images of prostate adenocarcinoma vertebral bone metastatic lesions. nnUNet is an open-source Python package that adds optimizations to deep learning-based UNet architecture but has not been extensively combined with transfer learning techniques due to the absence of a readily available functionality of this method. The IRB-approved study data set includes imaging studies from patients with mPC who were enrolled in clinical trials at the University of Southern California (USC) Health Science Campus and Los Angeles County (LAC)/USC medical center. Manual segmentation of metastatic lesions was completed by an expert radiologist Dr. Vinay Duddalwar (20+ years in radiology and oncologic imaging), to serve as ground truths for the automated segmentation. Despite nnUNet鈥檚 success on some medical segmentation tasks, it only produced an average Dice Similarity Coefficient (DSC) of 0.31 on the USC dataset. DSC results fell in a bimodal distribution, with most scores falling either over 0.66 (reasonably accurate) or at 0 (no lesion detected). Applying more aggressive data augmentation techniques dropped the DSC to 0.15, and reducing the number of epochs reduced the DSC to below 0.1. Datasets have been identified for transfer learning, which involve balancing between size and similarity of the dataset. Identified datasets include the Pancreas data from the Medical Segmentation Decathlon, Pelvic Reference Data, and CT volumes with multiple organ segmentations (CT-ORG). Some of the challenges of producing an accurate model from the USC dataset include small dataset size (115 images), 2D data (as nnUNet generally performs better on 3D data), and the limited amount of public data capturing annotated CT images of bone lesions. Optimizations and improvements will be made by applying transfer learning and generative methods, including incorporating generative adversarial networks and diffusion models in order to augment the dataset. Performance with different libraries, including MONAI and custom architectures with Pytorch, will be compared. In the future, molecular correlations will be tracked with radiologic features for the purpose of multimodal composite biomarker identification. Once validated, these models will be incorporated into evaluation workflows to optimize radiologist evaluation. Our work demonstrates the challenges of applying automated image segmentation to small medical datasets and lays a foundation for techniques to improve performance. As machine learning models become increasingly incorporated into the workflow of radiologists, these findings will help improve the speed and accuracy of vertebral metastatic lesions detection. <p class="card-text"><strong>Keywords:</strong> <a href="https://publications.waset.org/abstracts/search?q=deep%20learning" title="deep learning">deep learning</a>, <a href="https://publications.waset.org/abstracts/search?q=image%20segmentation" title=" image segmentation"> image segmentation</a>, <a href="https://publications.waset.org/abstracts/search?q=medicine" title=" medicine"> medicine</a>, <a href="https://publications.waset.org/abstracts/search?q=nnUNet" title=" nnUNet"> nnUNet</a>, <a href="https://publications.waset.org/abstracts/search?q=prostate%20carcinoma" title=" prostate carcinoma"> prostate carcinoma</a>, <a href="https://publications.waset.org/abstracts/search?q=radiomics" title=" radiomics"> radiomics</a> </p> <a href="https://publications.waset.org/abstracts/162601/image-segmentation-with-deep-learning-of-prostate-cancer-bone-metastases-on-computed-tomography" class="btn btn-primary btn-sm">Procedia</a> <a href="https://publications.waset.org/abstracts/162601.pdf" target="_blank" class="btn btn-primary btn-sm">PDF</a> <span class="bg-info text-light px-1 py-1 float-right rounded"> Downloads <span class="badge badge-light">96</span> </span> </div> </div> </div> </main> <footer> <div id="infolinks" class="pt-3 pb-2"> <div class="container"> <div style="background-color:#f5f5f5;" class="p-3"> <div class="row"> <div class="col-md-2"> <ul class="list-unstyled"> About <li><a href="https://waset.org/page/support">About Us</a></li> <li><a href="https://waset.org/page/support#legal-information">Legal</a></li> <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/WASET-16th-foundational-anniversary.pdf">WASET celebrates its 16th foundational anniversary</a></li> </ul> </div> <div class="col-md-2"> <ul class="list-unstyled"> Account <li><a href="https://waset.org/profile">My Account</a></li> </ul> </div> <div class="col-md-2"> <ul class="list-unstyled"> Explore <li><a href="https://waset.org/disciplines">Disciplines</a></li> <li><a href="https://waset.org/conferences">Conferences</a></li> <li><a href="https://waset.org/conference-programs">Conference Program</a></li> <li><a href="https://waset.org/committees">Committees</a></li> <li><a href="https://publications.waset.org">Publications</a></li> </ul> </div> <div class="col-md-2"> <ul class="list-unstyled"> Research <li><a href="https://publications.waset.org/abstracts">Abstracts</a></li> <li><a href="https://publications.waset.org">Periodicals</a></li> <li><a href="https://publications.waset.org/archive">Archive</a></li> </ul> </div> <div class="col-md-2"> <ul class="list-unstyled"> Open Science <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/Open-Science-Philosophy.pdf">Open Science Philosophy</a></li> <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/Open-Science-Award.pdf">Open Science Award</a></li> <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/Open-Society-Open-Science-and-Open-Innovation.pdf">Open Innovation</a></li> <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/Postdoctoral-Fellowship-Award.pdf">Postdoctoral Fellowship Award</a></li> <li><a target="_blank" rel="nofollow" href="https://publications.waset.org/static/files/Scholarly-Research-Review.pdf">Scholarly Research Review</a></li> </ul> </div> <div class="col-md-2"> <ul class="list-unstyled"> Support <li><a href="https://waset.org/page/support">Support</a></li> <li><a href="https://waset.org/profile/messages/create">Contact Us</a></li> <li><a href="https://waset.org/profile/messages/create">Report Abuse</a></li> </ul> </div> </div> </div> </div> </div> <div class="container text-center"> <hr style="margin-top:0;margin-bottom:.3rem;"> <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank" class="text-muted small">Creative Commons Attribution 4.0 International License</a> <div id="copy" class="mt-2">© 2024 World Academy of Science, Engineering and Technology</div> </div> </footer> <a href="javascript:" id="return-to-top"><i class="fas fa-arrow-up"></i></a> <div class="modal" id="modal-template"> <div class="modal-dialog"> <div class="modal-content"> <div class="row m-0 mt-1"> <div class="col-md-12"> <button type="button" class="close" data-dismiss="modal" aria-label="Close"><span aria-hidden="true">×</span></button> </div> </div> <div class="modal-body"></div> </div> </div> </div> <script src="https://cdn.waset.org/static/plugins/jquery-3.3.1.min.js"></script> <script src="https://cdn.waset.org/static/plugins/bootstrap-4.2.1/js/bootstrap.bundle.min.js"></script> <script src="https://cdn.waset.org/static/js/site.js?v=150220211556"></script> <script> jQuery(document).ready(function() { /*jQuery.get("https://publications.waset.org/xhr/user-menu", function (response) { jQuery('#mainNavMenu').append(response); });*/ jQuery.get({ url: "https://publications.waset.org/xhr/user-menu", cache: false }).then(function(response){ jQuery('#mainNavMenu').append(response); }); }); </script> </body> </html>