Coming to campus? Visit this page for important information.

PhD Comprehensive Examination Announcement of Abdala Nour:"Deep learning approach for MRI Breast tumor segmentation and classification by using active contour model and image processing techniques"

Friday, December 3, 2021 - 11:00 to 12:30


The School of Computer Science would like to present…  

PhD. Comprehensive Exam by:  Abdala Nour 

Date: Friday December 3, 2021 
Time: 11:00am -12:30pm 
Passcode: If interested in attending this event, contact the Graduate Secretary at with sufficient notice before the event to obtain the passcode   


Breast Image segmentation is one of the most important techniques in the medical image processing field for diagnosis and treatment follow up. Fully automatic segmentation of the region of interest from medical images has been researched for years to enhance the efficiency and accuracy of understanding such images. Currently, with the advance of deep learning based image segmentation, it has gained great success in medical image fields especially in semantic segmentations and classifications. Various deep convolutional neural networks were evaluated including U-Net, Mask R-CNN, and U-Net++. However, it is difficult to find a specific universal segmentation model suitable for most MRI breast cancer images. This is due to the large variation of the mass in term of texture distribution and location in MRI images with low contrast, ambiguous boundaries and different types of noises that corrupt the images. Moreover, the scarcity of high-quality annotated training data has a difficulty in handling irregularities of mass shape, size, and density. Therefore, the combination of deep learning and classical image processing techniques has a significant effect to improve the accuracy of lesion detection and mass segmentation results. In this paper, we propose a framework which integrates fully convolutional neural network(FCN) with a region-based active contour model. Particularly, we want to define a new loss function which incorporates area and size information and integrates this into a dense deep learning model for enhancing segmentation and detection methods. Thus, the objects can be localized, distinguished and/or measured to show reasonable results. Moreover, to evaluate our method, we built different types of deep convolutional neural network models used for medical image segmentation process, namely: FCN, U-Net, VGG-16, Resnet50, and Mask-RCNN to measure the performance of each method.  
Keywords: Image segmentation, Convolutional neural network, U-Net architecture, active contour model, Mask R-CNN. 

PhD Dissertation Committee: 

Internal Reader:  Dr. Imran Ahmad  
Internal Reader: Dr. Dima Alhadidi 
External Reader: Dr. Mohamed Belalia (Dept. of Mathematics and Statistics) 
Advisor: Dr. Boubaker Boufama and Dr. Sherif Saad  

PhD Comprehensive Exam Announcement   Vector Institute in Artificial Intelligence artificial intelligence approved topic logo




5113 Lambton Tower 401 Sunset Ave. Windsor ON, N9B 3P4 (519) 253-3000 Ext. 3716 (working remotely)