Global Pose Refinement using Bidirectional Long-Short Term Memory

Ibrahim Radwan; Akshay Asthana; Roland Goecke

Global Pose Refinement using Bidirectional Long-Short Term Memory

Ibrahim Radwan, Akshay Asthana, Roland Goecke

Research output: Contribution to conference › Paper › peer-review

Abstract

In This paper, a bi-directional long-short term memory (LSTM) framework is proposed to refine pose estimation and tracking for multiple people. The key idea of our algorithm is to learn the temporal consistencies of the human body shapes between subsequent frames. This helps removing the wrong sudden outliers and improve the general smoothness of the pose tracking. The proposed approach has been evaluated on PoseTrack dataset for both the validation and test subset sequences. The overall detection and tracking results have been improved over the frame-by-frame only baseline detection.

Original language	English
Pages	1-5
Publication status	Published - 2017
Event	IEEE International Conference on Computer Vision, PoseTrack Workshop and Challenge - Venice, Italy, Italy Duration: 1 Jan 2017 → …

Conference

Conference	IEEE International Conference on Computer Vision, PoseTrack Workshop and Challenge
Country/Territory	Italy
Period	1/01/17 → …
Other	October 22-29, 2017

Cite this

@conference{6750e76ce00e423fb2007930ec2e2283,

title = "Global Pose Refinement using Bidirectional Long-Short Term Memory",

abstract = "In This paper, a bi-directional long-short term memory (LSTM) framework is proposed to refine pose estimation and tracking for multiple people. The key idea of our algorithm is to learn the temporal consistencies of the human body shapes between subsequent frames. This helps removing the wrong sudden outliers and improve the general smoothness of the pose tracking. The proposed approach has been evaluated on PoseTrack dataset for both the validation and test subset sequences. The overall detection and tracking results have been improved over the frame-by-frame only baseline detection.",

author = "Ibrahim Radwan and Akshay Asthana and Roland Goecke",

year = "2017",

language = "English",

pages = "1--5",

note = "IEEE International Conference on Computer Vision, PoseTrack Workshop and Challenge ; Conference date: 01-01-2017",

}

TY - CONF

T1 - Global Pose Refinement using Bidirectional Long-Short Term Memory

AU - Radwan, Ibrahim

AU - Asthana, Akshay

AU - Goecke, Roland

PY - 2017

Y1 - 2017

N2 - In This paper, a bi-directional long-short term memory (LSTM) framework is proposed to refine pose estimation and tracking for multiple people. The key idea of our algorithm is to learn the temporal consistencies of the human body shapes between subsequent frames. This helps removing the wrong sudden outliers and improve the general smoothness of the pose tracking. The proposed approach has been evaluated on PoseTrack dataset for both the validation and test subset sequences. The overall detection and tracking results have been improved over the frame-by-frame only baseline detection.

AB - In This paper, a bi-directional long-short term memory (LSTM) framework is proposed to refine pose estimation and tracking for multiple people. The key idea of our algorithm is to learn the temporal consistencies of the human body shapes between subsequent frames. This helps removing the wrong sudden outliers and improve the general smoothness of the pose tracking. The proposed approach has been evaluated on PoseTrack dataset for both the validation and test subset sequences. The overall detection and tracking results have been improved over the frame-by-frame only baseline detection.

M3 - Paper

SP - 1

EP - 5

T2 - IEEE International Conference on Computer Vision, PoseTrack Workshop and Challenge

Y2 - 1 January 2017

ER -

Global Pose Refinement using Bidirectional Long-Short Term Memory

Abstract

Conference

Fingerprint

Cite this