Richard Zhang wins 2022 Canadian Human-Computer Communications Society Achievement Award


Hao (Richard) Zhang, a distinguished professor at Simon Fraser University and an Amazon Scholar working with the Imaging Technology group, has received the 2022 Canadian Human-Computer Communications Society (CHCCS) Achievement Award “for his numerous high-impact contributions to computer graphics.”

Zhang has published more than 170 papers on various topics in visual computing, a subfield of computer science dealing with the analysis and creation of visual data that comprises images and 3D content, including online products and virtual environments used in AR/VR. Many of his papers have introduced groundbreaking approaches that have since become foundational tools in the field. The award citation notes “his sustained and impactful contributions to learning-based analysis and synthesis of visual data, especially 3D shapes and indoor scenes.”

“I am really thrilled to receive this award, especially at a time when we are all witnessing a tremendous growth in interests from both the industry and academia in 3D modeling and content creation,” said Zhang.

Related content

Reformulating the mapping problem to take advantage of sequence-to-sequence Transformers improves performance by an average of 15%.

Zhang’s work addresses challenges in 3D computer graphics and computer vision. One of his specialities is in 3D reconstruction: computing three-dimensional shapes from a variety of inputs, such as a laser scan to capture data points on a 3D object or only one or few photographs. The process often results in incomplete data requiring mathematical algorithms to fill in the missing pieces.

“The reconstruction problem is ill posed, meaning that there is not enough input to obtain a full and accurate 3D model,” said Zhang. “That’s why we need some priors, such as symmetry, other forms of regularity assumptions, or knowledge extracted or learned from existing data.”

Methods from three of Zhang’s papers, all related to 3D reconstruction, have been adopted by the Computational Geometry Algorithms Library (CGAL), a well-known open-source software project.

“Most of what I do today is data driven and learning based, involving neural networks,” said Zhang. “One really interesting question is what the best neural representation of 3D shapes would be. Unlike images and speech, 3D shapes are not confined to one standard representation.”

Related content

Integrating symbolic reasoning and learning efficiently from interactions with the world are two major remaining challenges, says vice president and distinguished scientist Nikko Ström.

With his student, Zhang published a foundational paper to introduce the use of implicit shape representations for geometric deep learning. While only three years old, this work has generated many follow-ups and triggered a shift in how modern neural networks are designed to reconstruct, render, and generate 3D data. A follow-up of this work from Zhang and collaborators won the Best Student Paper Award at CVPR 2020.

The CHCCS Achievement Award honors a Canadian researcher who has contributed significantly to the fields of computer graphics, visualization, or human-computer interaction. Zhang received the award and delivered a keynote talk at Graphics Interface 2022, an annual international conference devoted to computer graphics and human-computer interaction.

Zhang obtained his bachelor and master of mathematics degrees from the University of Waterloo, majoring in computer science. In 2002, he earned his PhD in computer science and computer graphics from the University of Toronto. He is a distinguished professor at Simon Fraser University, where he has been a full professor since 2014. He also directs the (GrUVi) lab, an interdisciplinary team of researchers who work in computer graphics and computer vision.

Zhang, who became an Amazon Scholar in November 2021, said he was attracted by an Amazon job description indicating the company’s desire to have a 3D model for every product it sells. As part of the Imaging Sciences group, Zhang is working on 3D modeling and content creation, with the goal of enhancing customers’ “3D experience” while shopping online: shoppers would not only view products in traditional ways but also interact with them as if they were real 3D objects in advance of making a purchase.

Related publication

We introduce Amazon Berkeley Objects (ABO), a new large-scale dataset designed to help bridge the gap between real and virtual 3D worlds. ABO contains product catalog images, metadata, and artist-created 3D models with complex geometries and physically-based materials that correspond to real, household objects. We derive challenging benchmarks that exploit the unique properties of ABO and measure the current

“We live in a 3D reality, so our online reality better be three dimensional so that people have the right and true experience,” said Zhang. “As far as I can see, there are still many technical challenges if we are to develop something that truly works—and at Amazon scale!”

In May 2022, Zhang took a one-year leave of absence from Simon Fraser to work full-time at Amazon.

“In terms of my knowledge and expertise, it’s a perfect match,” Zhang said. “I’m a 3D computer graphics person at heart while I also work on 3D vision and geometric deep learning. And these are all areas of interest and investment for Amazon.”

“I’m really excited about the opportunity to have a real-world impact,” Zhang added. “The opportunity to apply what I’ve learned from a theoretical perspective and have a big impact on people’s lives and experiences are extremely motivating.”

Source link


Please enter your comment!
Please enter your name here

Share post:


More like this

The Pixel 8 and Pixel 8 Pro: why Google’s innovative flagships should be your next buy

Google's brand new Pixel 8 and Pixel 8...

This Manhattan School Is Seeking To Change Education For Special Needs Students

For many special needs students and their families,...

To Counter Grade Inflation, We Need To Change The Way We Teach

Students are getting higher grades, but test scores...

5 ways to boost server efficiency

For AMD servers in particular, efficiency improves sharply...