Research Scientist @ AIST | Visiting Postdoc @ UTN | HQ @ cvpaper.challenge

Joined June 2017
54 Photos and videos
Ryousuke Yamada retweeted
#CVPR2026 / @CVPR We've released the CVPR 2026 Report! hirokatsukataoka.net/temp/pr… Compiled during CVPR through a collaboration among LIMIT.Lab, cvpaper.challenge, and the Visual Geometry Group (VGG), this report provides meta-level insights into the key trends, research directions, emerging themes, and discussions that shaped this year's conference.
3
75
289
39,377
CVPR2026 Report is now publicly available!! It nicely covers the latest trends across SSL, generative models, robot learning, and many other topics 😄 Big thanks to the authors for putting this together so quickly!! hirokatsukataoka.net/temp/pr… @CVPR #CVPR
1
21
187
13,864
Web videos as scalable source data!! “Video” was one of the big keywords I kept hearing at CVPR 2026. I strongly feel that video will become a key data source in the next few years! Not in the old sense of simple action recognition, but as scalable supervision for vision models
1
364
If you are interested in research collaboration, please feel free to reach out to LIMIT.Lab!! limitlab.xyz/ Finally, I would like to thank the authors for putting together this excellent report so quickly!!
3
256
Yeeeees!!! Interested in an internship, full-time position, or research collaboration in Japan? Come talk to us at AIST booth 🇯🇵

ALT One Piece We Are Remix GIF

Come say hi our collaborators at @AIST_EN! They have a booth and can tell you all about it and the new cool ABCI 3.0 cluster with 6k H200 GPUs!! @gatheluck , @FragileGoodwill , @HirokatuKataoka
2
4
708
It seems everyone wants a Lambda cap 😄 #CVPR2026
1
1
256
Shall we do video generation models next?? 😛
154
Don’t forget to visit the AIST booth in #CVPR2026!! @AIST_EN @AIST_JP @AIResearchAIST @CVPR
1
2
8
1,057
Welcome remarks starting soon… #CVPR2026 @CVPR
1
6
485
Best Paper Honorable Mention 2/2 SAM 3D: 3Dfy Anything in Images ai.meta.com/research/publica… Congrats 👏
1
1
1
288
🥇Best Paper Efficiently Reconstructing Dynamic Scenes One D4RT at a Time d4rt-paper.github.io Super super congratulations 👏
1
5
351