We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: A CLIP-Enhanced Method for Video-Language Understanding

Abstract: This technical report summarizes our method for the Video-And-Language Understanding Evaluation (VALUE) challenge (this https URL). We propose a CLIP-Enhanced method to incorporate the image-text pretrained knowledge into downstream video-text tasks. Combined with several other improved designs, our method outperforms the state-of-the-art by $2.4\%$ ($57.58$ to $60.00$) Meta-Ave score on VALUE benchmark.
Comments: 3 pages, 1 figure, 2 tables. Technical report
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Cite as: arXiv:2110.07137 [cs.CV]
  (or arXiv:2110.07137v1 [cs.CV] for this version)

Submission history

From: Guohao Li [view email]
[v1] Thu, 14 Oct 2021 03:50:23 GMT (89kb,D)

Link back to: arXiv, form interface, contact.