Reproduce “MemoryFusion: Cross-Temporal Memory Learning for Multimodal Video Fusion” 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
Meiqi Gong, Hao Zhang, Jiayi Ma
MemoryFusion stabilizes infrared-visible video fusion by retaining useful features from recent and more distant frames, then smoothing and refining the fused sequence in two stages. On M3SVD, it reports VIF 0.4926 and flowD 2.6759, compared with TemCoCo’s 0.4770 and 3.1882. It leads on five of six M3SVD metrics and all six on HDO, but not every metric or setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.