PROJECT TITLE :

Low Resolution Face Recognition Across Variations in Pose and Illumination

ABSTRACT:

We propose a completely automatic approach for recognizing low resolution face pictures captured in uncontrolled surroundings. The approach uses multidimensional scaling to be told a common transformation matrix for the entire face that simultaneously transforms the facial features of the low resolution and therefore the high resolution training images such that the distance between them approximates the distance had both the images been captured under the same controlled imaging conditions. Stereo matching price is employed to get the similarity of two pictures in the reworked space. Though this provides terribly smart recognition performance, the time taken for computing the stereo matching cost is vital. To overcome this limitation, we propose a reference-primarily based approach in that each face image is represented by its stereo matching value from some reference images. Experimental analysis on the real world challenging databases and comparison with the state-of-the-art super-resolution, classifier based and cross modal synthesis techniques show the effectiveness of the proposed algorithm.