Zhengkai Jiang

Researcher
Tencent Youtu Lab
Email: zhengkjiang [at] tencent (dot) com

Google Scholar / Github / Zhihu / Twitter

Photo

I am currently a Researcher at Tencent Youtu Lab. My research interests are in computer vision and deep learning. Currently, I'm working on data-efficient learning for object detection, segmentation and tracking.

Previously, I obtained the master's degree at 2020 from National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. I received the bachelor's degree at 2017 from the Department of Automation, Northeastern University with honors.

I am looking for highly motivated research interns working on object detection, instance segmentation and data-efficient learning based in Shanghai.

News

Publications

 
p2pnet
Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework
Qingyu Song*, Changan Wang*, Zhengkai Jiang, Yabiao Wang, Ying Tai, Chengjie Wang, Jilin Li, Feiyue Huang, Yang Wu
ICCV 2021 / Paper
Oral Presentation

The first point-based framework for crowd counting and localization.
siamrcr
SiamRCR: Reciprocal Classification and Regression for Visual Object Tracking
Jinlong Peng*, Zhengkai Jiang*, Yueyang Gu*, Yang Wu, Yabiao Wang, Ying Tai, Chengjie Wang, Weiyao Lin
IJCAI 2021 / Paper

Reciprocal classification and regression was proposed for visual object tracking.
autoassign
AutoAssign: Differentiable Label Assignment for Dense Object Detection
Benjin Zhu, Jianfeng Wang, Zhengkai Jiang, Fuhang Zong, Songtao Liu, Zeming Li, Jian Sun
Arxiv 2020 / Paper / Code

Performing differentiable label assignment for dense object detection.
dynamichead
Fine-Grained Dynamic Head for Object Detection
Lin Song, Yanwei Li, Zhengkai Jiang, Zeming Li, Hongbin Sun, Jian Sun, Nanning Zheng
NeurIPS 2020 / Paper / Code

A fine-grained dynamic head is proposed to conditionally select a pixel-level combination of FPN features from different scales.
ltf-v2
Rethinking Learnable Tree Filter for Generic Feature Transform
Lin Song, Yanwei Li, Zhengkai Jiang, Zeming Li, Xiangyu Zhang, Hongbin Sun, Jian Sun, Nanning Zheng
NeurIPS 2020 / Paper / Code

Performing differentiable tree-filter with a learnable unary term for generic feature transform.
lsts
Learning Where to Focus for Efficient Video Object Detection
Zhengkai Jiang, Yu Liu, Ceyuan Yang, Jihao Liu, Peng Gao, Qian Zhang, Shiming Xiang Chunhong Pan
ECCV 2020 / Paper / Code

The offsets of sampling locations across videos are treated as parameters and would be learned through back-propagation guided by bounding box and classification loss.
cbgs
Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection
Benjin Zhu, Zhengkai Jiang, Xiangxin Zhou, Zeming Li, Gang Yu
Arxiv 2019 / Paper / Code

NuScenes winner of WAD 2019 3D Object Detection Challenges at CVPR 2019 workshop.
cross-modality
Dynamic Fusion with Intra and Inter-Modality Attention Flow for Visual Question Answering
Peng Gao, Zhengkai Jiang, Haoxuan You, Pan Lu, Steven CH Hoi, Xiaogang Wang, Hongsheng Li
CVPR 2019 / Paper
Oral Presentation

Cross modality self-attention was proposed to capture the high-level interactions between language and vision domains.
lwdn
Video Object Detection with Locally-Weightd Deformable Neighboors
Zhengkai Jiang, Peng Gao, Chaoxu Guo, Qian Zhang, Shiming Xiang, Chunhong Pan
AAAI 2019 / Paper

Locally-weighted deformable neighbors module was proposed to model motion for video object detection without utilizing time-consuming optical flow extraction networks.

Honors

Services

  • Conference Reviewer:
  • The International Conference on Learning Representations (ICLR), 2022
    Conference on Neural Information Processing Systems (NeurIPS), 2021
    International Conference on Computer Vision (ICCV), 2021
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021
    Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI), 2021
  • Journal Reviewer:
  • International Journal of Computer Vision (IJCV)