Human Action Recognition Based on Radon Transform

Springer-Verlag Berlin / Heidelberg
Publication Type:
Studies in Computational Intelligence vol 346. Multimedia Analysis, Processing and Communications, 2011, Studies in Computational Intelligence, pp. 369 - 389
Issue Date:
Full metadata record
Files in This Item:
Filename Description Size
Thumbnail2011002116OK.pdf527.16 kB
Adobe PDF
A new feature description is used for human action representation and recognition. Features are extracted from the Radon transforms of silhouette images. Using the features, key postures are selected. Key postures are combined to construct an action template for each action sequence. Linear Discriminant Analysis (LDA) is applied to obtain low dimensional feature vectors. Different classification methods are used for human action recognition. Experiments are carried out based on a publicly available human action database.
Please use this identifier to cite or link to this item: