Recurrent human pose estimation
We propose a novel ConvNet model for predicting 2D human body poses in an image. The model regresses a heatmap representation for each body keypoint, and is able to learn and represent both the part appearances and the context of the part configuration. We make the following three contributions: (i)...
| Hauptverfasser: | , |
|---|---|
| Format: | Internet publication |
| Sprache: | English |
| Veröffentlicht: |
arXiv
2016
|
| _version_ | 1826315376778018816 |
|---|---|
| author | Belagiannis, V Zisserman, A |
| author_facet | Belagiannis, V Zisserman, A |
| author_sort | Belagiannis, V |
| collection | OXFORD |
| description | We propose a novel ConvNet model for predicting 2D human body poses in an image. The model regresses a heatmap representation for each body keypoint, and is able to learn and represent both the part appearances and the context of the part configuration. We make the following three contributions: (i) an architecture combining a feed forward module with a recurrent module, where the recurrent module can be run iteratively to improve the performance, (ii) the model can be trained end-to-end and from scratch, with auxiliary losses incorporated to improve performance, (iii) we investigate whether keypoint visibility can also be predicted. The model is evaluated on two benchmark datasets. The result is a simple architecture that achieves performance on par with the state of the art, but without the complexity of a graphical model stage (or layers). |
| first_indexed | 2024-12-09T03:26:46Z |
| format | Internet publication |
| id | oxford-uuid:c3b9eafb-eae4-4b05-b42d-2e92611fa841 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-12-09T03:26:46Z |
| publishDate | 2016 |
| publisher | arXiv |
| record_format | dspace |
| spelling | oxford-uuid:c3b9eafb-eae4-4b05-b42d-2e92611fa8412024-11-28T13:05:47ZRecurrent human pose estimationInternet publicationhttp://purl.org/coar/resource_type/c_7ad9uuid:c3b9eafb-eae4-4b05-b42d-2e92611fa841EnglishSymplectic ElementsarXiv2016Belagiannis, VZisserman, AWe propose a novel ConvNet model for predicting 2D human body poses in an image. The model regresses a heatmap representation for each body keypoint, and is able to learn and represent both the part appearances and the context of the part configuration. We make the following three contributions: (i) an architecture combining a feed forward module with a recurrent module, where the recurrent module can be run iteratively to improve the performance, (ii) the model can be trained end-to-end and from scratch, with auxiliary losses incorporated to improve performance, (iii) we investigate whether keypoint visibility can also be predicted. The model is evaluated on two benchmark datasets. The result is a simple architecture that achieves performance on par with the state of the art, but without the complexity of a graphical model stage (or layers). |
| spellingShingle | Belagiannis, V Zisserman, A Recurrent human pose estimation |
| title | Recurrent human pose estimation |
| title_full | Recurrent human pose estimation |
| title_fullStr | Recurrent human pose estimation |
| title_full_unstemmed | Recurrent human pose estimation |
| title_short | Recurrent human pose estimation |
| title_sort | recurrent human pose estimation |
| work_keys_str_mv | AT belagiannisv recurrenthumanposeestimation AT zissermana recurrenthumanposeestimation |