Avatar

Tony Joseph

Research Assistant at VCLAB

University of Ontario Institute of Technology
tony.joseph_at_ontariotechu.net

I recently completed MSc. in Computer Science under the supervision of Prof. Faisal Qureshi and Prof. Kosta Derpanis. I am a member of Visual Computing Lab at UOIT. Over the past year, I have been working on the intersection between deep learning and computer vision. My research interests mainly lie in the areas of computer vision, machine learning, and generative models. More specifically, I have been working on attention models for convolutional architectures and image inpainting through edge learning. During my masters apart from my academic pursuits, I did an internship at SPXTRM Health Inc. where I worked on applied vision.

Interests

  • Computer Vision
  • Machine Learning
  • Generative Models
  • Reinforcement Learning

Education

  • MSc. in Computer Science, 2019

    University of Ontario Institute of Technology

  • Nanodegree in Self-Driving Car, 2017

    Udacity

  • BEng. in Electrical Engineering, 2016

    University of Ontario Institute of Technology

News

August 25, 2019
One paper accepted to the Advances in Image Manipulation (AIM) workshop at International Conference on Computer Vision (ICCV) 2019.
July 1, 2019
One paper accepted as spotlight to the British Machine Vision Conference (BMVC) 2019.
June 28, 2019
Attended SciNet-HPC summer school. It was great introduction to CUDA programming.
May 30, 2019
Successfully defended my M.Sc. thesis! Thank you to my advisors: Faisal Qureshi and Kosta Derpanis for their support and guidance.
May 3, 2019
Presented our work on Joint Spatial and Layer Attention at Southern Ontario Numerical Analysis Day (SONAD).
July 14, 2018
Attended International Computer Vision Summer School (ICVSS-2018). Thank you to all the organizers for making it a fun and educational event.

Publications

This work proposes a two stage adversarial model EdgeConnect that comprises of an edge generator followed by an image completion network. The edge generator hallucinates edges of the missing region (both regular and irregular) of the image, and the image completion network fills in the missing regions using hallucinated edges as a priori.
arXiv 2019 (accepted at ICCV-AIM workshop 2019)
                                @inproceedings{nazeri2019edgeconnect,
                                      title={EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning},
                                      author={Nazeri, Kamyar and Ng, Eric and Joseph, Tony and Qureshi, Faisal and Ebrahimi, Mehran},
                                      journal={arXiv preprint},
                                      year={2019},
                                }
                            

In this work, we propose a novel approach that learns to sequentially attend to different Convolutional Neural Networks (CNN) layers (i.e., “what” feature abstraction to attend to) and different spatial locations of the selected feature map (i.e., “where”) to perform the task at hand.
arXiv 2019 (accepted at BMVC 2019, Spotlight)
                                    @inproceedings{joseph2019JSLAN,
                                          title={Joint Spatial and Layer Attention for Convolutional Networks},
                                          author={Joseph, Tony and Derpanis, Konstantinos and Qureshi, Faisal},
                                          journal={Conference on the British Machine Vision Association (BMVC), 2019},
                                          year={2019},
                                    }