We propose the DeepPose-based pose estimation system that is flexible with the change of bounding-box range for top-view images. Our purpose is to link person detection system and pose estimation system. We introduce Bounding-box Curriculum Learning (BCL) and Recurrent Pose Estimation (RPE). BCL is a learning technique of CNN inspired from Curriculum Learning. RPE is a recurrent process of pose estimation that fixes the bounding-box range in response to the estimated results. We show the effect of proposed methods compared to normal learned CNN-based pose estimator on our original top-view dataset.
|Title of host publication
|2016 IEEE 5th Global Conference on Consumer Electronics, GCCE 2016
|Institute of Electrical and Electronics Engineers Inc.
|Published - 2016 Dec 27
|5th IEEE Global Conference on Consumer Electronics, GCCE 2016 - Kyoto, Japan
Duration: 2016 Oct 11 → 2016 Oct 14
|5th IEEE Global Conference on Consumer Electronics, GCCE 2016
|16/10/11 → 16/10/14
- Convolutional Neural Networks
- Pose Estimation
ASJC Scopus subject areas
- Signal Processing
- Electrical and Electronic Engineering
- Computer Science Applications
- Hardware and Architecture