Reproduce “Event-Aware Instructed Assistant for Referring Video Segmentation” 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
Jinyu Liu, Henghui Ding, Shuting He, Yu-Gang Jiang
EVIS addresses referring video object segmentation by organizing compound descriptions into text-guided events before combining information across a video. Learned event queries aggregate object queries, while pixel features support mask prediction; Object-Pixel-Hybrid Learning combines both for language-model decoding. On MeViS, EVIS reaches 46.8 J&F versus 39.4 for the baseline, but does not lead VideoLISA on ReasonSeg.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.