.png){kind=icon}
.png){kind=icon}
Return
T5-based anomaly-behavior video captioning using semantic relation mining
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2025.113923.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We propose a text-to-text transfer transformer (T5)-based abnormal behavior video capturing. • A semantic relation mining module is implemented. • A frequency matrix represents the objects and actions appearing in each frame. • Complex semantic relationships are among object locations, visual relationships, interactions. • The module is generalizable and can be provided insights into abnormal behavior patterns.