Reproduce “Future Frame Prediction for Anomaly Detection -- A New Baseline” 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

Wen Liu, Weixin Luo, Dongze Lian, Shenghua Gao

The method predicts the next frame from four observed frames using normal training video. They compare the prediction with the next frame and constrain predicted motion to match actual motion. A mismatch is a possible warning, not proof of abnormality. It reaches 72.8% frame-level AUC on ShanghaiTech versus 68.0% for Stacked RNN, with scores relative within each test video.

Summary by alphaXiv.

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