Tianyu (Steve) Wang
@VisionSteveResearch Scientist @AdobeResearch | Ph.D. @ CUHK | Prev. Research Intern @AdobeResearch | Photographer | Opinions are my own
Similar User
@yining_hong
@yuewang314
@ziqi_huang_
@stevenpg8
@siyuanhuang95
@JingkangY
@j1h0u
@ge_yixiao
@kaiyangzhou
@RchalYang
@zhaohang0124
@JiahuiLei1998
@wjqdev
@zhijianliu_
@ShenhanQian
I will present two papers at the CVPR 2023. I am looking forward to see you in person on Thursday in poster section and discuss our new ideas in this field. #CVPR2023
Ads in AI. Copilot. Microsoft. My thoughts: Noooooooooooooooooooooooooooooooooooooo. No. No no no.
It’s lucky for this era that I can own some old film cameras to capture the life.
“Oh your model is based on previous model and you did not propose a new method. Then the paper lacks of novelty!” #ECCV2024
Amazing review verbatim - I'm speechless. The computer vision community should be ashamed that this is the state of our peer review system "Strengths: A good analysis is performed and some conclusion are been drawn. "Weaknesses: These more of an incremenatal work no detailed…
Great initiative！
Really interest paper that uses our SOBA dataset!
What Does Stable Diffusion Know about the 3D Scene? Our recent work @Oxford_VGG with @ChuanxiaZ @WeidiXie and Andrew Zisserman introduces a protocol to answer questions of 3D physical understanding via features of SD and other large models, enabling potential future applications.
If you guys persist in this, I will reject the review request from NeurIPS.
This year, we invite high school students to submit research papers on the topic of machine learning for social impact! See our call for high school research project submissions below. buff.ly/43TiTdD
I reviewed 7 papers for CVPR-24 (4 normal +3 emergency) while ensuring full engagement at all stages. For my paper, 2 reviewers (positive leaning) out of 3 disappeared in the final rating stage and we were left with abusive R3, who single handedly overruled the expected decision.
I'm not sure if this is the best possible outcome. I thought a lot about if I should write about this or not, but I hope there is some value in thinking about this a bit more. A longer 🧵 ...
Glad to share that our project Relightful Harmonization: Lighting-aware portrait background replacement has been accepted to #CVPR2024. 🧵 Project page: mengweiren.com/research/relig… Preprint: arxiv.org/abs/2312.06886
Hope it can be adopted to a immersive version for Vision Pro.
Really excellent demo: current-exhibition.com/laboratorio31/. Rock-solid and fast interactive 3DGS viewer with WASD+mouse controls for an "IRL" art gallery experience, very impressive and polished.
Make-A-Shape a Ten-Million-scale 3D Shape Model paper page: huggingface.co/papers/2401.11… shape model, known as Make-A-Shape, has been trained on 10 million diverse 3D shapes. It exhibits the capability to unconditionally generate a wide range of 3D shapes, featuring intricate…
The novelty in dataset paper can't only be the "data". There should be novelty in the collection approach/pipeline, annotations (better, faster,...), and potentially novelty in the benchmarks. Just collecting a bigger amount of data for a known pipeline cannot result in a paper.
Meet Luis Figueroa, Research Engineer at @AdobeResearch! His work focuses on teaching computers how to understand images, a crucial step in building new imaging editing tools. adobe.ly/49TMx4S
Glad to see that our proposed instance shadow detection (object-shadow association) and SSISv2 are proving beneficial in identifying unrealistic generated images. Great work!!! @anand_bhattad
Using just basic geometry cues like object-shadow associations, perspective fields, and line segments, we can accurately identify generated images. Our classifiers exclusively analyze these geometric features, without any access to the actual image pixels. 2/4
United States Trends
- 1. Clemson 9.995 posts
- 2. Travis Hunter 22,7 B posts
- 3. Feds 26,7 B posts
- 4. Heisman 10,4 B posts
- 5. Colorado 76,9 B posts
- 6. Chris Wright 5.893 posts
- 7. #iubb 1.154 posts
- 8. Zepeda 3.847 posts
- 9. Tyler Warren 1.037 posts
- 10. Dabo 1.789 posts
- 11. Ty Simpson N/A
- 12. Ceyair Wright N/A
- 13. Pentagon 120 B posts
- 14. Nebraska 6.900 posts
- 15. #Huskers N/A
- 16. Arkansas 26,8 B posts
- 17. FAMU 1.479 posts
- 18. Lagway 1.624 posts
- 19. Bill Nye 3.010 posts
- 20. #mnwildfirst N/A
Who to follow
-
Yining Hong
@yining_hong -
Yue Wang
@yuewang314 -
Ziqi Huang
@ziqi_huang_ -
Shaowei Liu
@stevenpg8 -
Siyuan Huang
@siyuanhuang95 -
Jingkang Yang @NTU🇸🇬
@JingkangY -
Ji Hou
@j1h0u -
Yixiao Ge
@ge_yixiao -
Kaiyang Zhou
@kaiyangzhou -
Ruihan Yang
@RchalYang -
Hang Zhao
@zhaohang0124 -
Jiahui
@JiahuiLei1998 -
Jiaqi Wang
@wjqdev -
Zhijian Liu
@zhijianliu_ -
Shenhan Qian
@ShenhanQian
Something went wrong.
Something went wrong.