Unsupervised human action retrieval using salient points in 3D mesh sequences.
Christos VeinidisIoannis PratikakisTheoharis TheoharisPublished in: Multim. Tools Appl. (2019)
Keyphrases
- salient points
- human actions
- d mesh
- action recognition
- spatio temporal
- space time
- human motion
- activity recognition
- human activities
- video sequences
- motion history images
- three dimensional
- d objects
- information retrieval
- action classification
- feature vectors
- visual features
- image retrieval
- shape descriptors
- geometric properties
- semi supervised
- motion recognition
- global features
- salient features
- object recognition
- computer vision
- knn
- image segmentation
- feature space
- image registration