Skip to main content
Oral defences & examinations, Thesis defences

Masters Thesis Defense: Shima Shahfar


Date & time
Wednesday, December 15, 2021
9 a.m. – 11 a.m.
Cost

This event is free

Where

Online

Candidate:

Shima Shahfar

   
             

Thesis Title:

Unsupervised Structure-Consistent Image-to-Image Translation

             

Date & Time:

Wednesday, December 15th, 2021 @ 9:00 AM

   
             

Location:

Zoom

   
             

Examining Committee:

         
             
 

Dr. Yiming Xiao

(Chair)

   
             
 

Dr. Charalambos Poullis

(Supervisor)

   
             
 

Dr. Adam Krzyzak

(Examiner)

 
             
 

Dr. Yiming Xiao

(Examiner)

 
             
             

 

 

 

Abstract:

           

 

There have been significant advances in designing deep networks for complex computer vision tasks. One that is of considerable importance is image understanding through pixel-wise classification, i.e. semantic segmentation. Despite the advances, self-supervised algorithms have many limitations and challenges, with perhaps the most significant being generalization. This thesis introduces a method based on generative models as a practical approach for addressing these shortcomings. First, we analyze several semantic segmentation methods to gain insight into their limitations. We investigate the effectiveness of one of the state-of-the-art methods on two different problem settings. The latter part of the thesis introduces an alternative approach using generative adversarial networks and autoencoders for image-to-image translation. The main idea is encoding an image into two latent codes to represent structure and style. We propose a new approach to enforce structure-consistency without requiring semantic labels to disentangle the two latent codes. We further show how this would result in a more detailed style transfer and image manipulation. Finally, we present results on multiple datasets and discuss how our approach can be practical in real-world applications. Our experiments demonstrate that our approach performs better than the baselines -or, in the worst-case, gives comparable results- while solving some of the shortcomings in tasks requiring a semantic mask.

Back to top

© Concordia University