xi's moments
Home | Education

Interest in 3D, cars drives graduate student's future

By CAO CHEN in Shanghai | CHINA DAILY | Updated: 2022-07-07 09:28

Chen Hansheng sits in the cockpit of the TR19, a racing car he and other racing club members designed at Tongji University in 2020. CHINA DAILY

In an era where career choices can be driven by practical concerns, there are still many who stick to pursuing their passions and interests. Chen Hansheng, a postgraduate from Tongji University who won the Best Student Paper Award at the world's premier computer vision event recently, is one such person.

The honor was awarded by the IEEE Conference on Computer Vision and Pattern Recognition 2022 on June 22. This year's conference received 8,161 submissions and around one-quarter were accepted. The Best Student Paper Honorable Mention went to Harvard University and Google Research.

The paper, titled "EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation", is a collaborative work between Chen and professors from Tongji University and experts from Alibaba Damo Academy, a science and technology research institute under Alibaba Group, where Chen was an intern last year.

Integrating methods of geometric algorithms and deep learning, the team discussed in the paper a model able to identify the 3D positions of objects in reality based on 2D images captured by cameras, which may have applications in studies of autonomous driving, robotics and drones.

"It's a classic topic in basic research on 3D vision. We explored an integrated research approach to increase a positioning system's 3D accuracy," said the 25-year-old, a master's student in automotive studies and lead author of the paper.

The award has brought Chen fame and many job offers in the industry, but he is determined to stick to his path in academia.

"Working with amazing experts in the field while writing the paper was enlightening for me, sharing insights on scientific problems and utilizing massive computing resources," he said.

"I'm not ready to go deep into applications now and would prefer a quiet, pure environment that is highly research and discovery focused for learning computer vision, which I like. It's fun to explore the unknown," said Chen, who plans to pursue a doctoral degree.

Like many boys, Chen has grown up with a passion for computers and motor racing. His parents bought him toy cars and magazines, and he played racing video games and watched motorsport.

An interest in computer science has also motivated him to learn how to modify game parameters and design customized cars in games.

Chen was a member of a racing club in Tongji for three years until 2019, where his skills in computers, engineering and automobiles were polished while club members designed and built their own racing cars for participating in Formula SAE, a global student design competition organized by United States-based SAE International.

The team members from the club regarded Chen as being dedicated to research and who often puts his ideas into practice in hands-on experiments, according to China Science Daily.

The TR-18, a racing car developed by the club for Formula SAE in China and Japan in 2018, was Chen's first design and his proudest to date.

The black, white and red car with a carbon fiber monocoque body is light weight with high torsional rigidity and so ensures good handling performance during cornering. The application of electronic throttle technology improves the stability of the vehicle.

"My computer skills helped a lot in the process when I did simulation design on a computer to improve the accuracy and make it as close as possible to the data of a real car," he said.

Chen's TR-18 is regularly put on display in exhibitions at Tongji University.

Global Edition
BACK TO THE TOP
Copyright 1995 - . All rights reserved. The content (including but not limited to text, photo, multimedia information, etc) published in this site belongs to China Daily Information Co (CDIC). Without written authorization from CDIC, such content shall not be republished or used in any form. Note: Browsers with 1024*768 or higher resolution are suggested for this site.
License for publishing multimedia online 0108263

Registration Number: 130349