Back to projects

Disentangled Gradient Harmonizing Mechanism for Weakly-Supervised Medical Annotations

  • Date:10/19
  • Team: Yi Xu, Tiancheng Lin, Jiancheng Yang, Wen Gu, Guozheng Xu, Canqian Yang, Yuanfan Guo, Shaofei Qin

Brief Introduction

On October 13-17, at the 22nd MICCAI (22nd International Conference on Medical Image Computing & Computer Assisted Intervention) held in Shenzhen, the team of Associate Professor Xu Yi in Institute of Image Communication and Network Engineering has won good grades in DigestPath 2019(Digestive-System Pathological Detection and Segmentation Challenge 2019). They won the second place (1st Runner Up) in the signet cell detection and the fourth in colonoscopy tissue segmentation and classification. The award-winning team members under the guidance of Associate Professor Xu Yi are graduate students Lin Tiancheng, Yang Jiancheng, Gu Wen, Xu Guozhen and undergraduates Yang Canqian, Guo Yuanfan and Qin Shaofei, a total of 7 students.

Certificate and Medal

Competition overview

Digestpath 2019 challenge is hosted by Sensetime, Histo Pathology Diagnostic Center, Ruijin Hospital, Xijing Hospital and Shanghai Songjiang District Central Hospital. The goal of the challenge is to set up tasks for evaluating automatic algorithms on signet ring cell detection and colonoscopy tissue screening from digestive system pathological images.

This is the first challenge and first public dataset on signet ring cell detection and colonoscopy tissue screening. Releasing the large quantity of expert-level annotations on digestive-system pathological images will substantially advance the research on automatic pathological object detection and lesion segmentation.

Award-winning algorithm introduction

The detection model proposes a solution to the partial labeling problem in signet cell detection. A novel decoupled gradient harmonizing loss (DGHM-loss) is advanced to calculate the gradient harmonizing loss for different regions. Therefore, the model can solve the optimization deviation and model over-fitting problems incurred by partial annotation, and finally improve the recall rate and robustness in the weakly supervised problem which commonly exists in medical images.

Extended reading

MICCAI is a comprehensive academic conference organized by the Medical Image Computing and Computer Assisted Intervention Society, which is currently recognized as a comprehensive academic conference in the fields of medical imaging computing (MIC) and computer-assisted intervention (CAI). It is the top international conference in the fields of medical imaging computing, medical robots, artificial intelligence, assisted intervention, and computational biomedicine.