Reproduce “VISTA: Differentiable Viewpoint Planning for Long-Horizon Task Visualization” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Donggang Jia, Da Li, Zainab Alsuwaykit, Sebastian Hartwig, Dominik Engel, Timo Ropinski, Ivan Viola
VISTA plans action-conditioned camera views by optimizing differentiable image-space objectives for target visibility, surface diversity, functional orientation, context focus, and continuity. In VISTA-Bench, its adaptive views achieve 65.7% Task 3 closed-loop success averaged across three vision-language models, compared with 30.6% for the best static-camera average. This result is limited to the benchmark tasks and evaluated models.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.