Face recognition using a new compressive sensing-based feature extraction method

Mehdi Banitalebi Dehkordi, Amin Banitalebi-Dehkordi, Jamshid Abouei, Konstantinos N. Plataniotis

Abstract: This paper proposes a novel face recognition algorithm that utilizes a sparse Fast Fourier Transform (FFT)-based feature extraction method. In our algorithm, we use Compressive Sampling (CS) theory two times. First, in the feature extraction process for extracting the feature vectors from a face images, and second, in the classification process where the CS reconstruction is used for selecting true classes. As a result, a significant reduction in the dimensionality of the signals is achieved. Extensive and comparative experiments have been conducted to evaluate the performance of the proposed scheme. The experiment results show that the combined Compressive Sensing and Sparse Representation Classification (SRC) achieves a high recognition accuracy, while maintaining a reasonable computational complexity.