Key Frame Proposal Network for Efficient Pose Estimation in Videos
| Autor: | Yin Wang, Octavia Camps, Yuexi Zhang, Mario Sznaier |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0209 industrial biotechnology
Sequence business.industry Computer science Motion blur Frame (networking) ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Context (language use) 02 engineering and technology 020901 industrial engineering & automation Robustness (computer science) 0202 electrical engineering electronic engineering information engineering Unsupervised learning Key frame 020201 artificial intelligence & image processing Computer vision Artificial intelligence business Pose |
| Zdroj: | Computer Vision – ECCV 2020 ISBN: 9783030585198 ECCV (17) |
| Popis: | Human pose estimation in video relies on local information by either estimating each frame independently or tracking poses across frames. In this paper, we propose a novel method combining local approaches with global context. We introduce a light weighted, unsupervised, key frame proposal network (K-FPN) to select informative frames and a learned dictionary to recover the entire pose sequence from these frames. The K-FPN speeds up the pose estimation and provides robustness to bad frames with occlusion, motion blur, and illumination changes, while the learned dictionary provides global dynamic context. Experiments on Penn Action and sub-JHMDB datasets show that the proposed method achieves state-of-the-art accuracy, with substantial speed-up. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|