fansunqi/AKeyS
Agentic Keyframe Search for Video Question Answering
This project helps video analysts, researchers, and content managers quickly find answers to specific questions about video content. It takes a video and a question as input, then identifies only the most relevant frames to provide an accurate answer, reducing the need to manually review long videos. The primary users are those who need to extract concise information from extensive video datasets.
No commits in the last 6 months.
Use this if you need to efficiently answer questions about the content of long videos without watching the entire footage.
Not ideal if your main goal is general video summarization or transcription rather than answering specific queries about events or objects within the video.
Stars
16
Forks
—
Language
Python
License
MIT
Category
Last pushed
Apr 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/fansunqi/AKeyS"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
open-mmlab/mmpretrain
OpenMMLab Pre-training Toolbox and Benchmark
facebookresearch/mmf
A modular framework for vision & language multimodal research from Facebook AI Research (FAIR)
adambielski/siamese-triplet
Siamese and triplet networks with online pair/triplet mining in PyTorch
HuaizhengZhang/Awsome-Deep-Learning-for-Video-Analysis
Papers, code and datasets about deep learning and multi-modal learning for video analysis
KaiyangZhou/pytorch-vsumm-reinforce
Unsupervised video summarization with deep reinforcement learning (AAAI'18)