Harry Thasarathan

Harry Thasarathan

Computer Vision Research Fellow

Ontario Tech University

Hi there! I’m Harry, a fourth year computer science undergraduate student at Ontario Tech University. I’ve been fortunate enough to get involved in research early on in my undergrad career. Previously, I was supervised by Mehran Ebrahimi at the Imaging Lab. It was here where I collaborated with Kamyar Nazeri to incorporate temporal coherence into the GAN framework for automatic video colorization and single-image super resolution.

For my undergraduate honors thesis, I’ve been working on problems at the intersection of computer vision and graphics. The main focus of my research is on structured generative models that learn physics constrained representations of human locomotion in 3D. I’m very fortunate to be working at the Visual Computing Lab under the supervision of Faisal Qureshi and Ken Pu.

Interests

  • Deep Learning applied to Computer Vision
  • Computer Graphics
  • Computational Photography

Education

  • BSc in Computer Science, 2021

    University of Ontario Institute of Technology

Skills

python

Python

c-plusplus

C++

tensorflow

Tensorflow

pytorch

Pytorch

flutter

Flutter

sql

SQL

Publications

Artist-Guided Semiautomatic Animation Colorization

Artist-Guided Semiautomatic Animation Colorization

A continuation of our work on automatic animation colorization, we extend our method to keep artist’s in the loop to simultaneously preserve artistic vision and reduce tedious animation workloads. By incorporating color hints and temporal information to an adversarial image-to-image framework, we show that it is possible to meet the balance between automation and authenticity through artist’s input to generate colored frames with temporal consistency.

Automatic Temporally Coherent Video Colorization

Automatic Temporally Coherent Video Colorization

Greyscale image colorization for applications in image restoration has seen significant improvements in recent years. Many of these techniques that use learning-based methods struggle to effectively colorize sparse inputs. With the consistent growth of the anime industry, the ability to colorize sparse input such as line art can reduce significant cost and redundant work for production studios by eliminating the in-between frame colorization process. Simply using existing methods yields inconsistent colors between related frames resulting in a flicker effect in the final video. In order to successfully automate key areas of large-scale anime production, the colorization of line arts must be temporally consistent between frames. This paper proposes a method to colorize line art frames in an adversarial setting, to create temporally coherent video of large anime by improving existing image to image translation methods. We show that by adding an extra condition to the generator and discriminator, we can effectively create temporally consistent video sequences from anime line arts.

Edge-Informed Single Image Super-Resolution

Edge-Informed Single Image Super-Resolution

The recent increase in the extensive use of digital imaging technologies has brought with it a simultaneous demand for higher-resolution images. We develop a novel edge-informed approach to single image super-resolution (SISR). The SISR problem is reformulated as an image inpainting task. We use a two-stage inpainting model as a baseline for super-resolution and show its effectiveness for different scale factors (x2, x4, x8) compared to basic interpolation schemes. This model is trained using a joint optimization of image contents (texture and color) and structures (edges). Quantitative and qualitative comparisons are included and the proposed model is compared with current state-of-the-art techniques. We show that our method of decoupling structure and texture reconstruction improves the quality of the final reconstructed high-resolution image.

Experience

 
 
 
 
 

Research Assistant

Visual Computing Lab, Ontario Tech

May 2020 – Present

Responsibilities include:

  • Analysing and mining various 3D motion capture datasets and related techniques to incorporate 3D information into the deep learning pipeline
  • Modelling the temporal dynamics of human motion using the SMPL parameterized body mesh representation
  • Leveraging the dynamics model to make trajectory predictions and edit motions for animation tasks
  • Exploring differentiable rendering for real-time applications
 
 
 
 
 

Research Fellow

Imaging Lab, Ontario Tech

May 2018 – Jan 2020

Responsibilities included:

  • Conducting original research in the area of GANs for assisting humans in animation colorization and image editing
  • Developing tools for animators to leverage state of the art models to speed up workflows creating interest from animation studios
  • Delivering multiple presentations to both technical and non-technical crowds on deep learning and deep generative modelling
  • Collaborating with graduate students at the imaging lab at Ontario Tech with their research involving automatic edge guided image in-painting and super resolution.
  • Developing Amazon Mechanical Turk surveys to measure human perceptual plausibility of generated images via crowd source (2AFC)
  • Conducting multiple research surveys of machine learning techniques for industrial applications of computer vision to develop methods for addressing new use cases

Recent Posts

Stay Tuned!

Contact