Can I read TS2-Net: Advanced Text-Video Retrieval on EtoBox?
TS2-Net: Advanced Text-Video Retrieval by imuyangyang is a document available to read on EtoBox.
What is TS2-Net: Advanced Text-Video Retrieval about?
The document presents TS2-Net, a novel transformer architecture designed for text-video retrieval, which enhances spatial-temporal video representation through a token shift and selection mechanism. The token shift module captures subtle movements by adjusting the token sequence across adjacent frames, while the token selection module identifies and retains the most informative tokens for improved semantic modeling. TS2-Net achieves state-of-the-art performance on major benchmarks, demonstrating its effecti
- Author
- imuyangyang
- Language
- EN