Reproduce “CrossWeave: Cross-Modal Scene and Instance Retrieval from Sparse 2D-3D Alignment” 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

Aadith Warrier, Gnana Prakash Punnavajhala, Siddharth Tourani, Muhammad Haris Khan, Avinash Sharma, Madhava Krishna

CrossWeave trains a shared model for scene and instance retrieval from sparse image-patch/3D-point correspondences. At retrieval, it uses aggregated features for scenes and mean-pooled local features for instances. On ScanNet, under the 10-view protocol, it reaches 61.54% Top-1 recall versus 40.38% for CrossOver (Updated). Instance retrieval is within a known scene and mainly evaluated with ground-truth masks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.