Search
Abstract
The paper presents an approach to localize human body joints in 3D coordinates based on a single low resolution depth image. First a framework to generate a database of 80k realistic depth images from a 3D body model is described. Then data preprocessing and normalization procedure, and DNN and MLP artificial neural networks architectures and training are presented. The robustness against camera distance and image noise is analysed. Localization accuracy for each joint is reported and application for low resolution and large distance pose estimation is proposed. A very fast regression on body joints locations in 3D space is achieved, even in case of sensor noise, large distance and reaching off the screen.
Citations
-
2
CrossRef
-
0
Web of Science
-
3
Scopus
Cite as
Full text
- Publication version
- Accepted or Published Version
- License
- Copyright (2017 IEEE)
Keywords
Details
- Category:
- Conference activity
- Type:
- materiały konferencyjne indeksowane w Web of Science
- Title of issue:
- 2017 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA) strony 354 - 359
- Language:
- English
- Publication year:
- 2017
- Bibliographic description:
- Szczuko P..: ANN for human pose estimation in low resolution depth images, W: 2017 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA), 2017, Institute of Electrical and Electronics Engineers (IEEE),.
- DOI:
- Digital Object Identifier (open in new tab) 10.23919/spa.2017.8166892
- Verified by:
- Gdańsk University of Technology
Referenced datasets
seen 82 times