http://yjxiong.me/feeds.xml This is a list of Yuanjun's recent academic activities. http://www.wider-challenge.org/ http://www.wider-challenge.org/ Tuesday, 15 Nov 2020 00:00:00 GMT can found here. The challenge centers around the problem of precise localization of human faces and bodies, and accurate association of identities. This year the challenge is sponsored by SenseTime and Amazon. The winning prize will be appealing. The challenge comprises of three tracks: WIDER Face, aims at soliciting new approaches to advance the state-of- the-art in face detection. WIDER Pedestrian, has the goal of gathering effective and efficient approaches to address the problem of pedestrian detection in unconstrained environments. WIDER Person Search, presents an exciting challenge of searching persons across 192 movies. Welcome to participate in the challenge! ]]> http://www.wider-challenge.org/ http://www.wider-challenge.org/ Tuesday, 15 May 2018 00:00:00 GMT can found here. The challenge centers around the problem of precise localization of human faces and bodies, and accurate association of identities. This year the challenge is sponsored by SenseTime and Amazon. The winning prize will be appealing. The challenge comprises of three tracks: WIDER Face, aims at soliciting new approaches to advance the state-of- the-art in face detection. WIDER Pedestrian, has the goal of gathering effective and efficient approaches to address the problem of pedestrian detection in unconstrained environments. WIDER Person Search, presents an exciting challenge of searching persons across 192 movies. Welcome to participate in the challenge! ]]> https://github.com/zhirongw/lemniscate.pytorch https://github.com/zhirongw/lemniscate.pytorch Sunday, 6 May 2018 00:00:00 GMT
Download the paper at [Arxiv Preprint].
Checkout the code and models this [Github link]. ]]> https://github.com/yysijie/st-gcn https://github.com/yysijie/st-gcn Tuesday, 23 Jan 2018 00:00:00 GMT
Download the paper at [Arxiv Preprint].
Checkout the code and models this [Github link]. ]]> https://github.com/yjxiong/action-detection https://github.com/yjxiong/action-detection Wednesday, 20 Sep 2017 00:00:00 GMT
Check out the code and models this [Github link]. ]]> http://yjxiong.me/others/kinetics_action/ http://yjxiong.me/others/kinetics_action/ Tuesday, 15 Aug 2017 00:00:00 GMT Kinetics Human Action Dataset is a large-scale human action recogntion dataset comprising over 300,000 videos. We are releasing the TSN model weights pretrained on this dataset. Extensive transfer learning experiments confirmed that TSN models pretrained on Kinetics leads to significant performance boost compared with sole ImageNet pretraining. The model weights with CNN architectures BNInception and InceptionV3, as well as transfer learning results can be found on the website below.

[TSN on Kinetics Website]. ]]> https://github.com/yjxiong/tsn-pytorch https://github.com/yjxiong/tsn-pytorch Tuesday, 15 Aug 2017 00:00:00 GMT The reimplementation process is quite smooth thanks to the design of PyTorch. The coding takes less than one day. I have tested the classification performance on UCF101 and Kinetics to confirm that it matches the original Caffe implementation of TSN. So, happy experimenting!

[Github link]. ]]> http://yjxiong.me/others/ssn/ http://yjxiong.me/others/ssn/ Sunday, 30 Jul 2017 00:00:00 GMT
The code and models are released to foster future research efforts.
See you in Venice! ]]> http://openaccess.thecvf.com/content_cvpr_2017/html/Wang_UntrimmedNets_for_Weakly_CVPR_2017_paper.html http://openaccess.thecvf.com/content_cvpr_2017/html/Wang_UntrimmedNets_for_Weakly_CVPR_2017_paper.html Tuesday, 30 Jul 2017 00:00:00 GMT
The code and models are released here. ]]> http://yjxiong.me http://yjxiong.me Tuesday, 30 Jul 2017 00:00:00 GMT I have made a RSS feed to reflect my activities. Subscribe if you are interested!