Deep learning-based for human segmentation and tracking, 3D human pose estimation and action recognition on monocular video of MADS dataset.
Van-Hung LePublished in: Multim. Tools Appl. (2023)
Keyphrases
- action recognition
- deep learning
- video dataset
- human actions
- body parts
- human pose
- action detection
- static images
- human activities
- view invariant
- recognition of human actions
- human detection
- space time interest points
- action classification
- motion history images
- activity recognition
- robust tracking
- recognizing human actions
- computer vision
- bag of words
- unsupervised learning
- machine learning
- segmentation algorithm
- segmentation method
- image segmentation
- pose estimation
- weakly supervised
- multiscale
- spatio temporal
- video sequences
- image sequences
- particle filter
- multiple objects
- motion features
- video data
- optical flow
- object segmentation
- human motion
- video frames
- mean shift
- space time