RenHuan1999/CVPR2023_P-MIL

The official implementation of 'Proposal-based Multiple Instance Learning for Weakly-supervised Temporal Action Localization' (CVPR 2023)

32
/ 100
Emerging

This project helps video analysts and researchers automatically identify and precisely locate specific actions within long video clips, even when only general descriptions of the video content are available. It takes video features and existing action proposals as input, then outputs detailed temporal boundaries for each action. This is ideal for professionals working with large volumes of video footage for content analysis or surveillance.

No commits in the last 6 months.

Use this if you need to pinpoint the exact start and end times of actions in videos, but only have access to video-level labels for training.

Not ideal if you already have precisely timestamped action labels for every action in your training videos, as this method is designed for weaker supervision.

video-analysis action-recognition content-moderation surveillance event-detection
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 8 / 25

How are scores calculated?

Stars

42

Forks

3

Language

Python

License

MIT

Last pushed

Jun 01, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/RenHuan1999/CVPR2023_P-MIL"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.