27983

قطعه بندي تصاوير پزشكي با استفاده از يادگيري خودنظارتي

كارشناسي ارشد

مهندسي كامپيوتر - هوش مصنوعي و رباتيكز

1398

1401/8/14

دكتر محسن سرياني

دكتر محمدرضا محمدي

مهندسي كامپيوتر

امروزه پزشكان متخص ص با استفاده از تصاوير راديوگراف ي قادر به تشخيص نوع و محل دقيق بيماري ها هستند. از انواع تصاوير پزشكي ميتوان به تصاوير MRI ، CT و Ultrusound اشاره كرد. با اين حال تشخيص سريع و به موقع بيماري يكي از عوامل حياتي و تعيين كننده در روند درمان و پيشگيري از پيشرفت بيماري است. سيست مهاي هوشمند تشخيص بيماري با استفاده از تصاوير پزشكي كمك شاياني در سرعت بخشيدن به تشخيص نوع و حتي مكان دقيق بيماري كرد ها ند . براي آموزش مدل هاي هوشمند به منظور تشخيص بيماري، نياز به تعداد زيادي تصاوير پزشكي برچسب خورده است تا بتوان با استفاده از آ نها يك مدل را آموزش داد. با اين حال جمع آوري و برچسب زني تعداد زيادي داده در اين حوزه هزينه بسيار بالايي دارد. به همين دليل استفاده از روش هاي خودنظارتي جهت استفاده از تصاوير بدون برچسب ميتواند تاثير به سزايي داشته باشد . در اين پايان نامه انواع روش هاي خودنظارتي براي آموزش يك شبكه عصبي عميق را بررسي كرد هايم. اين روشها ويژگيهاي غني و كاربردي را از تصاوير بدون برچسب استخراج كرده تا با استفاده از داده كمتر برچسب خورده به خروجي مناسبتري جهت تشخيص بيماري برسيم. وظيفه مورد بررسي در اين پايان نامه قطعه بند ي 1 تصاوير پزشكي جهت استخراج تومور مغزي است كه بر روي تصاوير MRI كار شده است . شبكه عصبي عميق استفاده شده در اين پايان نامه [1]Unet است كه به عنوان يك شبكه قطعهبند در

1401/12/10

Medical Image Segmentation Using Self-Supervise Learning

11/5/2023 12:00:00 AM

Today, doctors are able to diagnose the exact type and location of diseases by using radiography images. Among the types of medical images, we can mention MRI, CT and Ultrasound images. However, the quick diagnosis of the disease is one of the vital and decisive factors in the process of treatment and prevention of disease progression. Intelligent disease diagnosis systems using medical images are a great help in speeding up the diagnosis of the type and even the exact location of the disease. In order to train smart models for disease diagnosis, many labeled medical images are needed so that a model can be trained using them. However, collecting and labeling a large amount of data in this area is very expensive. For this reason, using self-supervised methods to use images without labels can have a significant effect. In this work, we have examined various self-supervised methods for training a deep neural network. These methods extract rich and useful features from unlabeled images to reach a more suitable output for disease diagnosis by using less labeled data. The task investigated in this thesis is the segmentation of medical images for brain tumor extraction, which is done on MRI images. The deep neural network used in this thesis is Unet[1], which was introduced as a segmentation network in 2015 and has undergone many changes in the past years to improve the results. In the first part, by using auxiliary tasks such as rotation and inpainting, it has been shown that the use of self-supervised tasks can increase the Dice score by 7% to 10% compared to the time when these tasks were not used. In the second part, with inspired by contrastive learning, a new method has been introduced in the network training, which has a significant effect on learning the segmentation of medical images. There are 1000 brain volumes in the data used in this work. When we use all these 1000 volumes without self-supervise learning, the Dice score of the segmentation is 86%. Using this method, using only 300 volumes, the accuracy of the segmentation network has reached 84%.

قطعه بندي , توده , يادگيري خودنظارتي

segmentation , Brain Tumor , Self-supervised learning

Sadegh Ranjbar

Dr. Mohsen Soryani