Reproduce “SEALECT: Scene-Conditioned Pose Priors for Plausible Maritime Vessel Retrieval” 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
Faisal Mohammad, Mohamed Younis
SEALECT selects vessel images for insertion into maritime scenes by combining a learned visual retriever with a training-free, scene-conditioned prior over vessel pose, vertical position, and scale. At K=20, its neural retriever reaches Recall@20 of 0.634, while adding the prior yields 0.627 and raises plausibility score from 2.278 to 2.418. The prior improves geometric fit, not demonstrated detection performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.