View-Invariant Recognition of Action Style Self-Dissimilarity
This work addresses the challenge of view-invariant action style recognition, which is incremental as it builds on self-similarity concepts for classification.
The paper tackled the problem of recognizing action styles across different viewing directions and camera parameters by introducing self-dissimilarity matrices, achieving remarkably good discriminant characteristics for gender recognition from video data on the IXMAS dataset.
Self-similarity was recently introduced as a measure of inter-class congruence for classification of actions. Herein, we investigate the dual problem of intra-class dissimilarity for classification of action styles. We introduce self-dissimilarity matrices that discriminate between same actions performed by different subjects regardless of viewing direction and camera parameters. We investigate two frameworks using these invariant style dissimilarity measures based on Principal Component Analysis (PCA) and Fisher Discriminant Analysis (FDA). Extensive experiments performed on IXMAS dataset indicate remarkably good discriminant characteristics for the proposed invariant measures for gender recognition from video data.