Automated prediction of endometriosis using deep learning

سال انتشار: 1400
محل انتشار: مجله آنالیز غیر خطی و کاربردها، دوره: 12، شماره: 2
کد COI اختصاصی: JR_IJNAA-12-2_183
زبان مقاله: انگلیسی
تعداد مشاهده: 108

نویسندگان

Department of Computer Science and Engineering, Hindustan Institute of Technology and Science, Padur, Chennai, India.

چکیده

Endometriosis is the anomalous progress of cells at the outer part of the uterus. Generally, this endometrial tissue stripes the uterine cavity. The existence of endometriosis is identified through procedures known as Transvaginal Ultra Sound Scan (TVUS), Magnetic Resonance Imaging (MRI), Laparoscopic procedures, and Histopathological slides. Minimal Invasive Surgery (MIS) Laparo-scopic images are recorded in a small camera. To assist the surgeon in identifying their presence of endometriosis, image quality (characteristics) was enhanced for more visual clarity. Deep learning has the ability in recognising the images for classification. The Convolutional Neural Networks (CNNs) perform classification of images on large datasets. The proposed system evaluates the performance by a novel approach that implements the transfer learning model on a well-known architecture called ResNet۵۰. The proposed system train the model on ResNet۵۰ architecture and yielded a training accuracy of ۹۱%, validation accuracy of ۹۰%, precision of ۸۳%, and recall of ۸۲%, which can be applied for larger datasets with better performance. The presented system yields higher Area Under Curve (AUC) of about ۰.۷۸. The proposed method yields better performance using ResNet۵۰ compared to other transfer learning techniques.

کلیدواژه ها

TVUS, MRI, Laparoscopic images Deep Learning, Convolution neural network (CNN), Transfer learning, ResNet۵۰

اطلاعات بیشتر در مورد COI

COI مخفف عبارت CIVILICA Object Identifier به معنی شناسه سیویلیکا برای اسناد است. COI کدی است که مطابق محل انتشار، به مقالات کنفرانسها و ژورنالهای داخل کشور به هنگام نمایه سازی بر روی پایگاه استنادی سیویلیکا اختصاص می یابد.

کد COI به مفهوم کد ملی اسناد نمایه شده در سیویلیکا است و کدی یکتا و ثابت است و به همین دلیل همواره قابلیت استناد و پیگیری دارد.