Paper Title
Robust Human Action Recognition based on Scale, Rotation and Shift Invariant Features

Abstract
Human Action Recognition from complex environments is challenging task due to several constraints like camera motion, cluttered background, and pose appearance variations. To achieve an efficient recognition performance in such conditions, features are needed to extract from an action image such that the recognition system will become more resilient. Towards such objective, in this paper, we have proposed a new hybrid feature extraction technique which is composed of Scale, Rotation and shift invariant features. We have applied Gabor filter for scale and rotational invariant feature and Complex Wavelet Transform for shift invariant features. Further to reduce the dimensionality, we have employed Principal Component Analysis. Simulation experiments are conducted over a standard benchmark dataset, i.e., UCF YouTube Action dataset. The performance of developed HAR system is analyzed through Recognition Accuracy. Keywords - Action Recognition, Difference of Gaussian, Complex Wavelet, UCF Action Dataset, Accuracy.