Yes! That is very scary! Just trying to stay positive and survive 🙈
Btw, NIH already was having a political appointee review all grants before NOA. This just solidifies the whole thing.
I was freaking out too. But read somewhere that the grant is technically awarded to an institution and not the PI. So hopefully it won’t come to that.
🤣🤣🤣🤣
Why Fall colors in April? 🤣
Don’t joke this can be true! Like how they banned transgenic mice 🤦🏽♀️
#CVPR2025 offers registration and travel support to students from underrepresented communities. Awards are based on need, contribution, travel distance, identity, and advisor support.
Information and form: forms.gle/uDR2Q74drC4V...
Fueling up for #ICCV2025 deadline with some local Pizza!
PS. I kind of like hanging around in the lab on the deadline day! So much energy all around!
Paper rejected from #CVPR2025, paper ready for #ICCV2025 💪
My students will present two papers, including one oral paper, in #WACV2025, which centers on personalized generative face models.
Thanks @luchaoqi.bsky.social for leading this direction.
Unfortunately, couldn't attend WACV as my other students needed me for MICCAI and ICCV deadlines!
I don’t know how this will be implemented, but may be time to search for opportunities outside the US!
Congratulations Andrew! This is amazing news! 🍾
We did found one! We say that these techniques can be used as data augmentation technique for creating ai models that can measure heart rate from videos, especially one with motion: motion-matters.github.io
Kind of surprising that these techniques do not destroy the heart rate of original video!
🎉 Congratulations to Prof. @mohitbansal.bsky.social for receiving the Presidential #PECASE Award by @WhiteHouse, which is the highest honor bestowed by US govt. on outstanding scientists/engineers who show exceptional potential for leadership early in their careers!
whitehouse.gov/ostp/news-up...
X is like this cringy Desi TV show, that you hate but continue watching … where I saw how trans hatred changed to Indians in tech hatred in matter of days - not sure how to feel about that! 😒
🚨 We have postdoc openings at UNC 🙂
Exciting+diverse NLP/CV/ML topics**, freedom to create research agenda, competitive funding, very strong students, mentorship for grant writing, collabs w/ many faculty+universities+companies, superb quality of life/weather.
Please apply + help spread the word 🙏
This is awesome! Me and @silviasellan.bsky.social has recently started exploring how gender-agnostics 3D face models can help tracking faces during HRT and help in FFS, to realize this can also help in other facial reconstructive surgery and tracking weight gain-loss + aging.
Thats a great idea! We were also looking into other SLAM applications in robotics which also face the same challenge of textureless surface with dynamic lighting.
Our NFL-BA loss relies on the fact that points closer to the camera (depth map) with orientations facing the camera (normal map) reflect more light. NFL-BA loss is minimized during the Tracking & Mapping phase of existing SLAM frameworks, results shown on 3DGS SLAMs.
SLAM algorithms struggle on endoscopy videos - specularity, textureless, & dynamic lighting.
We introduce a Near-Field Light Bundle Adjst. loss (NFL-BA): improves performance of SOTA SLAM, e.g. MonoGS (⬆️35% in tracking, ⬆️48% in mapping).
See asdunnbe.github.io/NFL-BA/
Led by Andrea & Daniel
Actually in 2.5yrs that’s my biggest green flag - passion!
Its almost time to recruit PhD students, any advice on how people identify the best students for their labs? This feels like the hardest and possibly the most consequential decision I have to make (so far I have been reasonably lucky). What qualities do you look for and any red flags to watch out?
I know 🤣
Why this bias against cats? 😭
Funny cat pics = huge mental health boost! 🤣
This is my first time and I am surprised to see how good the recommendations were and what a good quality of reviewers we have in the pool!
A few key insights:
❌ CtrlNet struggles to preserve input color & texture during relighting.
✅ Albedo-cond. Stable Diffusion to the rescue.
❌ Only scribble input in CtrlNet doesn't work at all.
✅ Use latent code from a denoising autoencoder that predicts shading map from scribble+normal map.
💡Introducing ScribbleLight, a generative model that can relight an image with simple scribbles - making it easier for virtual staging and interior design of homes.
Led by Jun-Myeong, collabs with Anand (TTIC), Pieter (W&M), and Annie.
@unccs.bsky.social
👉 chedgekorea.github.io/ScribbleLight/
SOTA virtual face re-aging techniques often struggle to preserve identity for large age changes.
We present MyTimeMachine, a personalized virtual aging gen. model, trained with ~50 images across 20-40 years.
Check out more cool results here: mytimemachine.github.io
1/2
⬇️👂
We are in the process of editing a SLAM handbook, to be published by Cambridge University Press, with many *stellar* contributors. Part 1 is available as an online draft for public comments. Help us find bugs/problems!
Link to release repo is here: lnkd.in/gZhTkaxb
Holiday party with the UNC Vision & Graphics Lab students … where they get to see more relaxed version of me after the CVPR deadline! 🤣
'Surprisingly', good-old StyleGANv2 is more effective for identity-preserving personalized aging than Diffusion, which struggles to make subtle changes - something to fix in the future!
Credits to @luchaoqi.bsky.social for leading the work, in collabs with folks from @unccs.bsky.social and umd
2/2
