Human pose estimation via convolutional part heatmap regression
This paper is on human pose estimation using Convolutional Neural Networks. Our main contribution is a CNN cascaded architecture specifically designed for learning part relationships and spatial context, and robustly inferring pose even for the case of severe part occlusions. To this end, we propose...
| Main Authors: | , |
|---|---|
| Format: | Conference or Workshop Item |
| Published: |
2016
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/36395/ |
| _version_ | 1848795279585705984 |
|---|---|
| author | Bulat, Adrian Tzimiropoulos, Georgios |
| author_facet | Bulat, Adrian Tzimiropoulos, Georgios |
| author_sort | Bulat, Adrian |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | This paper is on human pose estimation using Convolutional Neural Networks. Our main contribution is a CNN cascaded architecture specifically designed for learning part relationships and spatial context, and robustly inferring pose even for the case of severe part occlusions. To this end, we propose a detection-followed-by-regression CNN cascade. The first part of our cascade outputs part detection heatmaps and the second part performs regression on these heatmaps. The benefits of the proposed architecture are multi-fold: It guides the network where to focus in the image and effectively encodes part constraints and context. More importantly, it can effectively cope with occlusions because part detection heatmaps for occluded parts provide low confidence scores which subsequently guide the regression part of our net-work to rely on contextual information in order to predict the location of these parts. Additionally, we show that the proposed cascade is flexible enough to readily allow the integration of various CNN architectures for both detection and regression, including recent ones based on residual learning. Finally, we illustrate that our cascade achieves top performance on the MPII and LSP data sets. Code can be downloaded from http://www.cs.nott.ac.uk/~psxab5/ |
| first_indexed | 2025-11-14T19:29:34Z |
| format | Conference or Workshop Item |
| id | nottingham-36395 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| last_indexed | 2025-11-14T19:29:34Z |
| publishDate | 2016 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-363952020-05-04T18:17:38Z https://eprints.nottingham.ac.uk/36395/ Human pose estimation via convolutional part heatmap regression Bulat, Adrian Tzimiropoulos, Georgios This paper is on human pose estimation using Convolutional Neural Networks. Our main contribution is a CNN cascaded architecture specifically designed for learning part relationships and spatial context, and robustly inferring pose even for the case of severe part occlusions. To this end, we propose a detection-followed-by-regression CNN cascade. The first part of our cascade outputs part detection heatmaps and the second part performs regression on these heatmaps. The benefits of the proposed architecture are multi-fold: It guides the network where to focus in the image and effectively encodes part constraints and context. More importantly, it can effectively cope with occlusions because part detection heatmaps for occluded parts provide low confidence scores which subsequently guide the regression part of our net-work to rely on contextual information in order to predict the location of these parts. Additionally, we show that the proposed cascade is flexible enough to readily allow the integration of various CNN architectures for both detection and regression, including recent ones based on residual learning. Finally, we illustrate that our cascade achieves top performance on the MPII and LSP data sets. Code can be downloaded from http://www.cs.nott.ac.uk/~psxab5/ 2016-10-07 Conference or Workshop Item PeerReviewed Bulat, Adrian and Tzimiropoulos, Georgios (2016) Human pose estimation via convolutional part heatmap regression. In: 14th European Conference on Computer Vision (EECV 2016), 8-16 October 2016, Amsterdam, Netherlands. Human pose estimation Part heatmap regression Convolutional Neural Networks http://link.springer.com/book/10.1007%2F978-3-319-46478-7 |
| spellingShingle | Human pose estimation Part heatmap regression Convolutional Neural Networks Bulat, Adrian Tzimiropoulos, Georgios Human pose estimation via convolutional part heatmap regression |
| title | Human pose estimation via convolutional part heatmap regression |
| title_full | Human pose estimation via convolutional part heatmap regression |
| title_fullStr | Human pose estimation via convolutional part heatmap regression |
| title_full_unstemmed | Human pose estimation via convolutional part heatmap regression |
| title_short | Human pose estimation via convolutional part heatmap regression |
| title_sort | human pose estimation via convolutional part heatmap regression |
| topic | Human pose estimation Part heatmap regression Convolutional Neural Networks |
| url | https://eprints.nottingham.ac.uk/36395/ https://eprints.nottingham.ac.uk/36395/ |