Recent Publications

  • Sucheng Ren, Chu Han, Xin Yang, Guoqiang Han, and Shengfeng He. TENet: Triple Excitation Network for Video Salient Object Detection. European Conference on Computer Vision (ECCV), Glasgow, 2020 (Spotlight Oral, Acceptance Rate 5.0%)

  • Sucheng Ren, Wenxi Liu, Jianbo Jiao, Guoqiang Han, and Shengfeng He. Edge Distraction-aware Salient Object Detection. In submission to IEEE Transaction on Image Processing (TIP).

Recent News

More News

  • Sep., 2017 South China University of Technology scholarship
  • Feb. - Mar., 2019 Exchange student in UC San Diego
  • Sep., 2019 South China University of Technology scholarship
  • July., 2019 A paper is accepted by ECCV as Spotlight Oral, Acceptance Rate 5.0%

Projects

  • Apr., 2019 - Mar., 2020 I am working on how to use GAN to retarget the saliency of an image. Manipulating image according to mask allows users to generate images with target and controllable saliency visually instead of just defrauding saliency in the detection model. This research is supported under National Training Programs of Innovation and Entrepreneurship for Undergraduates with 10000 RMB of national funding.

  • Mar., 2019 - Feb., 2020 Research on the prediction and optimization of traffic flow. We have two directions. The first is using reinforcement learning build a traffic flow and reward system by simulating in software, and interactive with the environment by DNQ or A3C, some reinforcement learning method. The other is an evolutionary algorithm. Due to the slow interaction with the environment, we also design a simple model for online quick decision and updating with reinforcement learning. This research is supported under National Training Programs of Innovation and Entrepreneurship for Undergraduates with 10000 RMB of national funding.

Researches

*

Generative Model

Use generative model to generative significant image or equivariance representation

Saliency Retarget

Manipulate image visually according to guided saliency(mask)

Contact