Infrared face recognition based on compressive sensing and PCA

Conventional signal sampling requires that a signal must sample at least two times faster than the signal bandwidth to avoid losing information. But in the recent years, an emerging theory of “compressive sensing” which shows that it can capture and represent compressible signal at a rate below the Nyquist rate, and it is possible to reconstruct signals from far fewer data than what is usually considered necessary. So in this paper, we presents a new infrared face recognition approach combining compressive sensing and PCA, which gains measurements y via compressive sensing and then applies the PCA to y to get the features of the infrared facial images. We compare the performance of this method with some of existing approaches. The experiment shows that this new approach is invariant to illumination, shadow and facial expression variation, and it has higher classification accuracy and performance.