PhD Dissertation Defense Announcement of Yahya Alzahrani: "Convolutional Neural Networks for Breast Ultrasound Image Segmentation"

 

Date: Monday December 20th, 2021 

Time:  10:00 am – 1:00 pm 

Meeting URL:    https://us06web.zoom.us/j/83601278351?from=addon

Passcode: If interested in attending this event, contact the Graduate Secretary at csgradinfo@uwindsor.ca with sufficient notice before the event to obtain the passcode   

 

Medical image segmentation of anatomical structures in various types of 2D and 3D medical images plays an important role in the early stages detection of breast cancer. Different modalities, such as Magnetic Resonance Imaging (MRI), X-Rays, Positron Emission Tomography (PET), Computed Tomography (CT), and Ultrasound (US) can be used for diagnostics, planning, and treatment.  In this dissertation, we utilized US images as they are radiation free, cheap, and always available to detect breast masses. However, this is still a challenging task due to intensity inhomogeneity, shape variations, class imbalanced problems, etc. To solve the shortcomings of US, we proposed three encoder-decoder based neural networks for breast US images segmentation. First, a modified U-Net architecture equipped with pre-trained inception residual blocks as an encoder was presented along with a weighted loss function to handle the class imbalance challenge. Second, a U-Net architecture based residual blocks with convolution paths was introduced to resolve the problem of the vanishing gradient while downsampling the feature maps. Third, a novel attention based segmentation architecture was proposed comprising channel and spatial details; which can be employed efficiently for any segmentation task. Our contributions have proven their robustness and efficiency in the segmentation of breast US tumors.  

 

Keywords: Neural Networks, BUS Images, lesion Segmentation, Image Processing 

 

Internal Reader: Dr. Christie Ezeife 

Internal Reader: Dr. Dan Wu        

External Reader: Dr. Esam Abdel-Raheem 

External Examiner: Dr. Faisal Qureshi 

Advisor(s):  Dr. Boubakeur Boufama 

Chair:    Dr. Iris Xu (Dept. of Civil and Environmental Engineering)