Search papers, labs, and topics across Lattice.
University of North Carolina at Charlotte
2
0
3
High reconstruction errors in VQ-VAD reveal anomalies in human motion, achieving 81.83% accuracy in detecting video anomalies from normal behavior patterns.
MLLMs, despite their video understanding prowess, exhibit a surprisingly strong "better safe than sorry" bias when detecting anomalies in video, crippling their recall in real-world surveillance scenarios.
