dinesh reddy

WALT3D has accepted as Oral at #cvpr (top 90 out of 12000) WALT3D:Generating Realistic Training Data from Time-Lapse Imagery for Reconstructing Dynamic Objects under Occlusion Project Page: cs.cmu.edu/~walt3d Key Idea: Convert you image to 3D under severe Occlusions

🎥 Most VLMs break when asked to process 2hr+ long-form videos – they can’t answer questions, can’t give commentary, and can’t do anything without reprocessing the entire file. 👨‍💻 With Video-RAG from @vlmrun, you can index hours-long videos just once – and get grounded answers…

🤯 Building computer-vision applications is never going to be the same again. As @karpathy says, "English is going to be the new programming-language", and we just made that happen for programming visual AI. 👨‍💻 This past month, the @vlmrun team cooked up something magical that…

Most LLMs are blind – they can’t see or understand your videos. 🧰 With our newly introduced @vlmrun MCP tools, we go well beyond video understanding. Bring your own LLM agent and allow it understand content, edit videos, trim clips, add subtitles, and prep social-ready content…

We just presented this is an oral session. For more details drop by poster 76 if you are #cvpr

At a university, for research in AI, we cannot compete with big tech companies in scale or in speed, but we can for sure compete whose idea/approach is more correct.

Made a Demo For WALT(walt.cs.cmu.edu) a vision method to tackle🔥insane occlusions on @huggingface Spaces using @Gradio and @ak92501 support. Code and Dataset are released. demo: huggingface.co/spaces/CVPR/WA… Join CVPR org to submit your #cvpr paper demo. @roboVisionCMU @CVPR

WALT: Watch And Learn 2D amodal representation from Time-lapse imagery paper: cs.cmu.edu/~walt/pdf/walt… project page: cs.cmu.edu/~walt/ a novel method to automatically learn complete(i.e. both visible and occluded) representation of objects using Time-Lapse Imagery

Introducing WALT(inspired from breaking bad), a new method to automatically learn complete(i.e. both visible and occluded) representation of objects using Time-Lapse based self-supervision. Code and Dataset are released. Checkout: walt.cs.cmu.edu #CVPR2022 #roboVisionCMU

Human-generated art is not going away anytime soon. What's lost in the hype is the impact the creative process has on the artist. Art is transformative and therapeutic. The final artifact is often just a byproduct of this process; most art in the world is not made to be sold.

We are glad to host Dinesh Reddy this Saturday, April 09 at 10:00AM EST/7:30PM IST to discuss 'Self-Supervising Occlusions For Vision'. Join us : lnkd.in/dvJ2CZcp

It is not the "smartest" people who succeed in science. It takes intrinsic motivation, perseverance, the drive to stay active and hands on, a strong support network of mentors, and the curiosity to keep learning. Nice summary in Nature: nature.com/articles/d4158…

Honored to have received the innovation fellowship. @CMU_Robotics @Qualcomm @bitspilaniindia @iiit_hyderabad

Robotics Institute Ph.D. Students Selected for Qualcomm Innovation Fellowship ri.cmu.edu/robotics-insti…

Won best paper award at IVS 21. I always hated waiting at a red signal when the roads were empty. This work reduces the wait time by reconstruction moving objects in 4D(space and time). #CMU #nvidia #iv2021 Award Link:lnkd.in/g5XtSGK Project Link:lnkd.in/gv43kWC

Checkout our new paper Tessetrack = Tesseract(Inspired from Avengers) + Tracking(3D pose of humans) at 10 PM EST on 24th June @CVPR2021 @AmazonScience @roboVisionCMU. #cvpr2021 Paper: dineshreddy91.github.io/papers/TesseTr… Talk: youtube.com/watch?v=gpNwWw… Webpage: cs.cmu.edu/~ILIM/projects…

cs.cmu.edu/news/cmu-richa…

Tessetrack: End-to-End Learnable Multi-Person Articulated 3D Pose Tracking #922 has been accepted to #CVPR2021 ... Stay tuned for paper and results in the coming months