Wenxiu Sun
SenseTime
124 Papers
575 Citations
Wenxiu Sun is an academic researcher from SenseTime. The author has contributed to research in topics: Computer science & Convolutional neural network. The author has an hindex of 21, co-authored 104 publications. Previous affiliations of Wenxiu Sun include Hong Kong University of Science and Technology & University of Hong Kong.
Chat about Author
Papers
•Posted Content
Polarized Reflection Removal with Perfect Alignment in the Wild
TL;DR: This work identifies the misalignment issues of existing reflection removal datasets where the collected reflection-free images are not perfectly aligned with input mixed images due to glass refraction and proposes a polarized reflection removal model with a two-stage architecture.
6
MIPI 2022 Challenge on RGB+ToF Depth Completion: Dataset and Report
Wenxiu Sun,Qingpeng Zhu,Chongyi Li,Ruicheng Feng,Shangchen Zhou,Jun Jiang,Qingyu Yang,Chen Change Loy,Jianzhong Xu +8 more
- 15 Sep 2022
TL;DR: RGB+ToF Depth Completion, one of the five tracks, work-ing on the fusion of RGB sensor and ToF sensor (with spot illumina-tion), is introduced in the first MIPI challenge including five tracks focusing on novel image sensors and imaging algorithms.
6
MIPI 2022 Challenge on Quad-Bayer Re-mosaic: Dataset and Report
Qingyu Yang,Guang Yang,Jun Jiang,Chongyi Li,Ruicheng Feng,Shangchen Zhou,Wenxiu Sun,Qingpeng Zhu,Chen Change Loy,Jianzhong Xu +9 more
- 15 Sep 2022
TL;DR: The first MIPI challenge, including five tracks focusing on novel image sensors and imaging algorithms, is introduced, with Quad Joint Remosaic and Denoise, one of the five tracks, work-ing on the interpolation of Quad CFA to Bayer at full resolution.
•Posted Content
Robust Tracking Using Region Proposal Networks.
TL;DR: This paper discovered that the internal structure of Region Proposal Network (RPN)'s top layer feature can be utilized for robust visual tracking and showed that such property has to be unleashed by a novel loss function which simultaneously considers classification accuracy and bounding box quality.
5
•Posted Content
Zoom and Learn: Generalizing Deep Stereo Matching to Novel Domains
TL;DR: This work proposes a self-adaptation approach for CNN training, utilizing both synthetic training data and stereo pairs in the new domain (without ground-truths), and forms an iterative optimization problem with graph Laplacian regularization.
5