Lassonde computer science PhD grad awarded the John Barron Doctoral Dissertation Award

Bergeron Centre
Bergeron Centre

Amir Rasouli, a recent PhD graduate from Professor John K. Tsotsos’s research group at the Lassonde School of Engineering at York University, was awarded the 2020 Canadian Image Processing and Pattern Recognition Society (CIPPRS) John Barron Doctoral Dissertation Award. His winning thesis, titled “The Role of Context in Understanding and Predicting Pedestrian Behavior in Urban Traffic Scenes,” focuses on understanding and predicting pedestrian behaviours from the perspective of drivers to develop systems for intelligent driving for both autonomous and assistive driving applications.

One of the barriers for autonomous, or self-driving, vehicles is being able to react, understand and adapt to the unpredictable nature of pedestrians. To address this barrier, Rasouli set out to investigate how drivers and pedestrians were interacting with each other. Rasouli, along with another student from Tsotsos’s lab, Iuliia Kotseruba, spent more than 60 hours driving around the Greater Toronto Area collecting information on pedestrian-vehicle interactions. With a dashboard camera inside the vehicle, they recorded a large collection of videos of pedestrians in traffic scenes.

“We went down several different types of streets – main, side and residential – at different times of the day to ensure that we captured as many interactions as possible,” said Rasouli.

These novel datasets were published in 2017 and 2019 in the Proceedings of the International Conference on Computer Vision (ICCV). Since this type of data is extremely limited in published literature, the resulting models and analysis offered by Rasouli and his colleague are both unique and high impact. The two papers published in the ICCV are titled “Are they going to cross? A benchmark dataset and baseline for pedestrian crosswalk behavior” (2017) and “PIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction” (2019). Over the course of his graduate studies, Rasouli produced 11 publications spread across multiple peer-reviewed journals and conference proceedings, as well as a number of preprints on open-access archives.

A world traveller, Rasouli originally completed his undergraduate degrees in computer systems engineering and business management at the Royal Melbourne Institute of Technology in Australia before relocating to Toronto and joining Tsotsos’s lab as a master’s student. In this graduate studies, Rasouli initially focused on developing active visual search capabilities for autonomous robots. He transitioned to the applications in the context of autonomous driving systems when he began his doctoral studies.

Rasouli spoke highly of his experience with Tsotsos. “He treats his lab like a family,” said Rasouli. “He let me discover what I liked to do and was always there to support me.”

Rasouli is now a senior engineer at Huawei Technologies, working at the company’s Markham-based research lab, Noah’s Ark. He completed an internship there at the end of 2019 and the connections that he built at the company encouraged him to seek out employment opportunities with them after completing his PhD. Rasouli is continuing the same type of research at Huawei as he was doing at Lassonde – studying pedestrian behaviour.

“I like what I’m able to do at Noah’s Ark,” said Rasouli. “It is truly a research group and we’re given autonomy to investigate the problems we are interested in.”

Rasouli was honoured with this award at a virtual banquet on May 28 at the Conference on Computer and Robot Vision. This award was recently renamed as the John Barron Doctoral Dissertation Award in memorial to John Barron, a former PhD graduate from Tsotsos’s research group at the University of Toronto and a faculty member at Western University.

This is the second year in a row that a PhD graduate from Tsotsos’s research group has won this prestigious award. Calden Wloka received the award in 2019 for his thesis titled “An Evaluation of Saliency and Its Limits.”