Reproduce “Audio for Sports Highlight Detection: A Comparative Empirical Study” 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
Hao Xu, Meenakshi Sarkar, Vishnu Raj, David Gunawan
The authors test whether audio alone can rank sports highlights in broadcasts. They extract pretrained features from two-second clips and use a recurrent model to score each clip in temporal context, comparing audio, vision, and fused inputs. An audio-only GRU reaches 41.79 ± 0.68 mAP, above TR-DETR’s 37.77 ± 0.25, while fusion performs best; predictions remain fragmented and overinclusive on SV-Highlights.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.