Ruiheng Chang (常瑞恒)

Hi, my name is Ruiheng Chang(常瑞恒), a senior undergraduate student from Department of Computer Science, Shanghai Jiao Tong University. I am also a member of Zhiyuan Honor's Program of Engineering.

I am a member of Machine Vision and Inteligence Group directed by Prof. Cewu Lu. My reserach interests lie in the field of computer vision and deep learning.

Email  /  CV /  GitHub  /  LinkedIn


Shanghai Jiao Tong University, China

  • B.Sc., Department of Computer Science & Zhiyuan College
  • Sept. 2015 to June. 2019 (Expected)


Annotation-Free and One-Shot Learning for Instance Segmentation of Homogeneous Object Clusters
Zheng Wu, Ruiheng Chang, Jiaxu Ma, Cewu Lu, Chi Keung Tang
In the proceedings of IJCAI 2018, the Twenty-Seventh International Joint Conference on Artificial Intelligence

  • We propose an efficient framework for instance segmentation of homogeneous object cluster (HOC). Our proposed method significantly reduces the cost of data collection and labeling.
  • We propose an efficient method to generate realistic synthetic training data which significantly improves instance segmentation performance.
  • We build a dataset consisting of pixel-level annotated images of HOC.

  • National Scholarship, highest honor for undergraduates in China, top 0.2% nationwide  
  • Academic Excellence Scholarship (A-Class), awarded to top 1% students in SJTU  
  • Zhiyuan College Honored Scholarship, awarded to members of Zhiyuan Honors Program, top 5%  
  • Excellent Undergraduate Scholarship of Yang Yuanqing Education Fund, awarded to 4 students in CS department of SJTU  

Honors and Awards
  • Outstanding Winner & Vilfredo Pareto Award of Interdisciplinary Contest In Modeling (ICM), top 0.1% worldwide  
  • 6th place in Microsoft COCO Keypoints Challenge 2017, ICCV 2017 workshop, core developer  
  • 1th place in ICRA Tidy Up My Room Challenge, ICRA 2018 competition, core developer  
  • Most Commercial Award in Microsoft Penta Hackathon, top 5 of all teams  

Selected Projects

Alpha Pose, an Accurate Multi-person Pose Estimation System

Alpha Pose is an accurate multi-person pose estimation system.

  • It is the first open-sourced system that can achieve 70+ mAP (72.3 mAP) on COCO dataset and 80+ mAP (82.1 mAP) on MPII dataset.
  • To associate poses that indicates the same person across frames, we also provide an efficient online pose tracker called Pose Flow. It is also the first open-sourced online pose tracker that can both satisfy 60+ mAP (66.5 mAP) and 50+ MOTA (58.3 MOTA) on PoseTrack Challenge dataset.
  • View it on github:


SJTU Bot, a Question Answer System

The project was completed for Beauty Of Programming 2017 Competition.

  • We propose a sequence to sequence approach to judge the Matching Degree of question-answer pair.
  • We propose a algorithm to answer a question based on seqtoseq approach and information distilling technique.
  • We develop a bot service based on Microsoft Azure for answering questions of everything about SJTU.
  • The project rank 13/1198 in Microsoft Beauty of Programming 2017.
  • View it on github:

This page has been visited for   HTML Hit Counters   times