Nathan Jacobs
@jacobsn.bsky.social
professor at WashU; computer vision, remote sensing, etc.
"GenStereo: Towards Open-World Generation of Stereo Images and Unsupervised Matching" @ #ICCV2025
TL;DR: Is it possible to hallucinate a stereo pair for an image? Is that useful for training unsupervised stereo matching methods? Yes and YES!
qjizhi.github.io/genstereo/
TL;DR: Is it possible to hallucinate a stereo pair for an image? Is that useful for training unsupervised stereo matching methods? Yes and YES!
qjizhi.github.io/genstereo/
September 30, 2025 at 9:54 PM
"GenStereo: Towards Open-World Generation of Stereo Images and Unsupervised Matching" @ #ICCV2025
TL;DR: Is it possible to hallucinate a stereo pair for an image? Is that useful for training unsupervised stereo matching methods? Yes and YES!
qjizhi.github.io/genstereo/
TL;DR: Is it possible to hallucinate a stereo pair for an image? Is that useful for training unsupervised stereo matching methods? Yes and YES!
qjizhi.github.io/genstereo/
Global and Local Entailment Learning for Natural World Imagery at #ICCV2025. See you in a few weeks!
TL;DR: If you're using BioCLIP for something, give this a try. We make the learned embeddings pay more attention to the taxonomic hierarchy, and it often works better.
vishu26.github.io/RCME/
TL;DR: If you're using BioCLIP for something, give this a try. We make the learned embeddings pay more attention to the taxonomic hierarchy, and it often works better.
vishu26.github.io/RCME/
September 30, 2025 at 9:50 PM
Global and Local Entailment Learning for Natural World Imagery at #ICCV2025. See you in a few weeks!
TL;DR: If you're using BioCLIP for something, give this a try. We make the learned embeddings pay more attention to the taxonomic hierarchy, and it often works better.
vishu26.github.io/RCME/
TL;DR: If you're using BioCLIP for something, give this a try. We make the learned embeddings pay more attention to the taxonomic hierarchy, and it often works better.
vishu26.github.io/RCME/
Reposted by Nathan Jacobs
I'm giving a virtual talk tomorrow at 12pm EST about the work that I've done with @pless.bsky.social @rsouvenir19.bsky.social and @jacobsn.bsky.social on building image search tools to support the National Center for Missing and Exploited Children. Feel free to tune in. cina.gmu.edu/event/cina-d...
CINA Distinguished Speaker Series with Abby Stylianou
TraffickCam: Identifying Where Victims of Human Trafficking and Child Sexual Abuse are Photographed Join us on Wednesday, January 29 for the upcoming virtual CINA Distinguished Speaker Series event…
cina.gmu.edu
January 29, 2025 at 2:44 AM
I'm giving a virtual talk tomorrow at 12pm EST about the work that I've done with @pless.bsky.social @rsouvenir19.bsky.social and @jacobsn.bsky.social on building image search tools to support the National Center for Missing and Exploited Children. Feel free to tune in. cina.gmu.edu/event/cina-d...
Reposted by Nathan Jacobs
Our CVPR Workshop EarthVision, bridging computer vision and Earth observation, was accepted at #CVPR2025 for its 8th edition! Submit your papers by March 3 and stay tuned for more info : grss-ieee.org/events/earth...
January 13, 2025 at 4:24 PM
Our CVPR Workshop EarthVision, bridging computer vision and Earth observation, was accepted at #CVPR2025 for its 8th edition! Submit your papers by March 3 and stay tuned for more info : grss-ieee.org/events/earth...
This was a fun collaboration with Anindya, Srikumar, @aleksispi.bsky.social, Chongjie, and Eugene. It brings together our recent work on using reinforcement learning for geospatial search tasks and a long history of work on geolocalization. I encourage you to stop by to chat with Anindya!
December 4, 2024 at 7:35 PM
This was a fun collaboration with Anindya, Srikumar, @aleksispi.bsky.social, Chongjie, and Eugene. It brings together our recent work on using reinforcement learning for geospatial search tasks and a long history of work on geolocalization. I encourage you to stop by to chat with Anindya!