The overview framework of the coarse-to-fine alignment network.

The overview framework of the coarse-to-fine alignment network.

<p>The CFAN consists of a video alignment module to learn the common visual-semantic space, a cross-modal interaction module to explore the fine-grained alignment among frames, proposals and video. Also, the multi-level coarse-to-fine alignment information flows between modules to make full us...

Full description

Saved in:
Bibliographic Details
Main Author: Lingwen Meng (8968106) (author)
Other Authors: Fangyuan Liu (1438045) (author), Mingyong Xin (15185747) (author), Siqi Guo (355869) (author), Fu Zou (21370430) (author)
Published: 2025
Subjects:
Marine Biology
Science Policy
Plasma Physics
Infectious Diseases
Biological Sciences not elsewhere classified
sophisticated network architecture
perform sufficient experiments
evaluation results demonstrate
grained alignment information
global contextual information
fine alignment network
video alignment module
fine alignment information
numerous videos according
modal interaction module
level information
video retrieval
video moment
video collections
single video
two subtasks
temporal boundary
task due
original problem
moment retrieval
existing methods
directly scale
Tags: Add Tag
No Tags, Be the first to tag this record!

Similar Items