원문정보
초록
영어
In computer vision related applications, video analysis of human walking motion is currently one of the most active research topics. The task of analyzing human walking can be divided into three distinct subtasks – human detection or segmentation, motion tracking and walking pose analysis. Typically, the analysis of the human walking starts with the extraction of motion information, detection of the presence of humans in the sequences of frames and then followed by analysis of events related to walking. This paper presents a survey of different methodologies used for human walking motion analysis, approaches used for human detection or segmentation, various tracking methods, approaches for pose estimation and pose analysis. The common data sets available for building robust, automatic and intelligent systems to understand “walking” motion are also included. Finally, uses of unsupervised techniques for analysing human walking are highlighted. Human walking motion is a subset of a broad topic of human motion analysis.
목차
1. Introduction
2. Context and Taxonomy
2.1 Video Sequences Obtained as Intensity Images and Depth Images
2.2 Marker-based and Markerless-based Applications
2.3 Action hierarchy Perspective
3. Human Motion Methodologies
4. Human Detection or Segmentation
4.1. Global-Top Down Approach
4.2. Local- Bottom–up Approach
5. Evaluation Methods for Segmentation
5.1. Supervised Methods
5.2. Unsupervised Methods
6. Tracking Human Motion
7. Pose analysis of Human Walking Motion
7.1. Pose Estimation:
7.2. Action Recognition:
8. Common Datasets
9. Conclusion Remarks
References