A research team at MIT-SPARK has released DAAAM, an open-source system that builds hierarchical 4D scene graphs from RGB-D video in real time and attaches detailed natural-language descriptions to each element. Named for "Describe Anything, Anywhere, at Any Moment," the project is published on GitHub under a BSD-3-Clause license, and its companion paper, "Describe Anything Anywhere At Any Moment," is available on arXiv, with a presentation expected at CVPR 2026.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.