Ju He
19 Papers
1 Citations
Ju He is an academic researcher. The author has contributed to research in topics: Computer science & Engineering. The author has an hindex of 1, co-authored 9 publications.
Chat about Author
Papers
Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP
TL;DR: This work proposes to build everything into a single-stage framework using a shared Frozen Convolutional CLIP backbone, which not only significantly simplifies the current two-stage pipeline, but also remarkably yields a better accuracy-cost trade-off.
81
NTIRE 2023 Video Colorization Challenge
Xiaoyang Kang,Xianhui Lin,Kai Zhang,Zheng Hui,Wangmeng Xiang,Ju He,Xiaoming Li,Peiran Ren,Xuansong Xie,Radu Timofte,Yixin Yang,Jinshan Pan,Zhong Zheng,Peng Qiyan,Jiangxin Zhang,Jinhui Dong,Jinjing Tang,Chichen Li,Lin Li,Qirong Liang,Ruipeng Gang,Xiaofeng Liu,Shuang Feng,Shuai Li,Hao Wang,Chaoyu Feng,Furui Bai,Yuqian Zhang,Guangqi Shao,Xiaotao Wang,Lei Lei,Si-Yu Chen,Yu Zhang,Hanning Xu,Zheyuan Liu,Zhao Zhang,Yan Luo,Zhichao Zuo +37 more
- 01 Jun 2023
TL;DR: This paper reviews the video colorization challenge on the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2023, and uses DeOldify-video as the baseline method for two tracks.
29
Optimal Proposal Learning for Deployable End-to-End Pedestrian Detection
Xiaolin Song,Binghui Chen,Pengyu Li,Ju He,Biao Wang,Yifeng Geng,Xuansong Xie,Honggang Zhang +7 more
- 01 Jun 2023
TL;DR: This paper uses CNN-based light detector and introduces two novel modules, including a Coarse-to-Fine (C2F) learning strategy for proposing precise positive proposals for the Ground-Truth (GT) instances by reducing the ambiguity of sample assignment/output in training/testing respectively, and a Completed Proposal Network (CPN) for producing extra information compensation to further recall the hard pedestrian samples.
14
Learning from Unique Perspectives: User-aware Saliency Modeling
Shi Chen,Nachiappan Valliappan,Shaolei Shen,Xinyu Ye,Kai Kohlhoff,Ju He +5 more
- 01 Jun 2023
TL;DR: The critical roles of visual preferences in attention modeling are identified, and for the first time the problem of user-aware saliency modeling is studied, to propose a principled learning method to understand visual attention in a progressive manner.
10
Tracking Anything in High Quality
Jiawen Zhu,Zhe Chen,Zeqi Hao,Shijie Chang,Lu Zhang,Dong Wang,Huchuan Lu,Bin Luo,Ju He,Jinpeng Lan,Hanyuan Chen,Chenyang Li +11 more
TL;DR: This report proposes HQTrack, a framework for High Quality Tracking anything in videos, which mainly consists of a video multi-object segmenter (VMOS) and a mask refiner (MR) and ranks the 2nd place in the Visual Object Tracking and Segmentation (VOTS2023) challenge.
7