News

[2022-06-19] We won the 5th place of Video object segmentation in The 4th Large-scale Video Object Segmentation Challenge (Youtube-VOS).

[2021-10-14] Code for IJCAI'21 SGNet are released [Code].

[2021-10-14] Code for T-IP'21 PurNet are released [Code].

[2021-10-13] Dataset and code for ICME'21 task-driven SOD (TSOD) are released [Project].

[2021-07-04] One paper on foreground object segmentation is accepted by T-IP 2021.

[2021-04-30] One paper on lane detection is accepted by IJCAI 2021.

[2021-03-06] One paper on task-driven SOD is accepted as oral paper by ICME 2021.

[2020-02-17] I join Meituan as an algorithm engineer.

[2019-12-10] I pass my Master Dissertation defense.

[2019-11-19] One paper on 360$^\circ$ SOD is accepted by J-STSP 2020.

[2019-07-22] One paper on boundary-aware SOD is accepted by ICCV 2019.

About

I currently work as an algorithm engineer at Meituan.

I am broadly interested in solving real-world visual understanding problems with optimization/learning-based algorithms. Recently, I focus on developing learning and optimization frameworks for scene understanding and 3D Reconstruction.

I obtained my Master's degree (CS) in State Key Laboratory of Virtual Reality Technology and Systems, SCSE, Beihang University (BUAA), adviced by Prof. Jia Li [CVTEAM], in Jan. 2020. From 2018 to 2019, I worked as research assistant at Peng Cheng Laboratory with Dr. Changqun Xia. Before that, I received Bachelor's degree (CS) from SCSE, Northeastern University (NEU), in Jul. 2017.

This is my [CV].

Research Interest

Technical Report [Google Scholar]

5th Place Solution for YouTube-VOS Challenge 2022: Video Object Segmentation.

Wangwang Yang*, Jinming Su*, Yiting Duan, Tingyi Guo and Junfeng Luo
[PDF][arXiv]

InsCon:Instance Consistency Feature Representation via Self-Supervised Learning.

Junwei Yang, Ke Zhang, Zhaolin Cui, Jinming Su, Junfeng Luo and Xiaolin Wei
[arXiv]

Exploring Reciprocal Attention for Salient Object Detection by Cooperative Learning.

Changqun Xia, Jia Li, Jinming Su and Yonghong Tian.
[arXiv]

Learning a saliency evaluation metric using crowdsourced perceptual judgments.

Changqun Xia, Jia Li, Jinming Su and Ali Borji.
[arXiv]

A Novel Method for Video Semantic Segmentation.

A Novel Method for Instance Segmentation.

A Big Dataset and Novel Method for 360$^\circ$ SOD.

Publications [Google Scholar]

Salient Object Detection with Purificatory Mechanism and Structural Similarity Loss.

Jia Li, Jinming Su, Mingcan, Ma, Changqun Xia and Yonghong Tian.
IEEE Transactions on Image Processing (T-IP), 30, pp.6855-6868, 2021. [PDF] [arXiv] [Github]

Structure Guided Lane Detection.

Jinming Su, Chao Chen, Ke Zhang, Junfeng Luo, Xiaoming Wei and Xiaolin Wei.
International Joint Conference on Artificial Intelligence (IJCAI), 2021. [arXiv] [中文解读] [Github]

Exploring Driving-aware Salient Object Detection via Knowledge Transfer.

Jinming Su, Changqun Xia and Jia Li.
IEEE International Conference on Multimedia and Expo (ICME), 2021. (oral) [PDF] [arXiv] [Project]

Distortion-adaptive Salient Object Detection in 360$^\circ$ Omnidirectional Images.

Jia Li, Jinming Su, Changqun Xia and Yonghong Tian.
IEEE Journal of Selected Topics in Signal Processing (J-STSP), 14(1), pp.38-48, 2020. [PDF] [Arxiv] [Project]

Selectivity or Invariance: Boundary-aware Salient Object Detection.

Jinming Su, Jia Li, Yu Zhang, Changqun Xia and Yonghong Tian.
IEEE International Conference on Computer Vision (ICCV), 2019. [PDF] [Project]

Patents [SooPAT]

Projects

I specialize in computer vision with deep/machine learning learning. In addition, I am interested in some engineering practices, especially some open source projects related to deep/machine learning. Moreover, I am also interested in basic computer algorithm, such as data structure, program optimization, etc.

Selected Honors

Services and Activities