This paper proposes a method for human pose estimation in still images. The proposed method achieves occlusion-aware appearance modeling. Appearance modeling with less accurate appearance data is problematic because it adversely affects the entire training process. The proposed method evaluates the effectiveness of mitigating the influence of occluded body parts in training sample images. In order to improve occlusion evaluation by a discriminatively-trained model, occlusion images are synthesized and employed with non-occlusion images for discriminative modeling. The score of this discriminative model is used for weighting each sample in the training process. Experimental results demonstrate that our approach improves the performance of human pose estimation in contrast to base models.
Yuki KAWANA
Nara Institute of Science and Technology
Norimichi UKITA
Nara Institute of Science and Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yuki KAWANA, Norimichi UKITA, "Occluded Appearance Modeling with Sample Weighting for Human Pose Estimation" in IEICE TRANSACTIONS on Information,
vol. E100-D, no. 10, pp. 2627-2634, October 2017, doi: 10.1587/transinf.2017EDP7088.
Abstract: This paper proposes a method for human pose estimation in still images. The proposed method achieves occlusion-aware appearance modeling. Appearance modeling with less accurate appearance data is problematic because it adversely affects the entire training process. The proposed method evaluates the effectiveness of mitigating the influence of occluded body parts in training sample images. In order to improve occlusion evaluation by a discriminatively-trained model, occlusion images are synthesized and employed with non-occlusion images for discriminative modeling. The score of this discriminative model is used for weighting each sample in the training process. Experimental results demonstrate that our approach improves the performance of human pose estimation in contrast to base models.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2017EDP7088/_p
Copy
@ARTICLE{e100-d_10_2627,
author={Yuki KAWANA, Norimichi UKITA, },
journal={IEICE TRANSACTIONS on Information},
title={Occluded Appearance Modeling with Sample Weighting for Human Pose Estimation},
year={2017},
volume={E100-D},
number={10},
pages={2627-2634},
abstract={This paper proposes a method for human pose estimation in still images. The proposed method achieves occlusion-aware appearance modeling. Appearance modeling with less accurate appearance data is problematic because it adversely affects the entire training process. The proposed method evaluates the effectiveness of mitigating the influence of occluded body parts in training sample images. In order to improve occlusion evaluation by a discriminatively-trained model, occlusion images are synthesized and employed with non-occlusion images for discriminative modeling. The score of this discriminative model is used for weighting each sample in the training process. Experimental results demonstrate that our approach improves the performance of human pose estimation in contrast to base models.},
keywords={},
doi={10.1587/transinf.2017EDP7088},
ISSN={1745-1361},
month={October},}
Copy
TY - JOUR
TI - Occluded Appearance Modeling with Sample Weighting for Human Pose Estimation
T2 - IEICE TRANSACTIONS on Information
SP - 2627
EP - 2634
AU - Yuki KAWANA
AU - Norimichi UKITA
PY - 2017
DO - 10.1587/transinf.2017EDP7088
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2017
AB - This paper proposes a method for human pose estimation in still images. The proposed method achieves occlusion-aware appearance modeling. Appearance modeling with less accurate appearance data is problematic because it adversely affects the entire training process. The proposed method evaluates the effectiveness of mitigating the influence of occluded body parts in training sample images. In order to improve occlusion evaluation by a discriminatively-trained model, occlusion images are synthesized and employed with non-occlusion images for discriminative modeling. The score of this discriminative model is used for weighting each sample in the training process. Experimental results demonstrate that our approach improves the performance of human pose estimation in contrast to base models.
ER -