Abstract: Face recognition is one of the challenging areas in Image Processing, especially in unconstrained environments. The challenges arise due to the variations in face pose, expressions, illuminations, occlusions etc. This paper introduces a novel enhanced face recognition algorithm using RGB-D images. The 3D face recognition algorithms are developed to achieve higher accuracy. While it is challenging to use specialized 3D sensors due to their high cost. RGB-D images can be captured by low-cost sensors such as Kinect. The performance and applicability of existing face recognition algorithms is bound by the information content or cost implications. This existing paper uses a RISE algorithm that utilizes the depth information along with RGB images. This algorithm uses a combination of entropy, visual saliency, and depth information of the images with HOG for feature extraction, identification and random decision forest for classification. The proposed algorithm using DCT with DPA instead of RISE algorithm. The DCT is applied to the entire image to obtain the DCT coefficients, and then only some of the coefficients are selected to construct feature vectors and the DPA method is used to select the coefficients with the highest discriminant power. Finally, the recognition step is executed using a SVM classifier. Further, the ADM algorithm is proposed to extract and match geometric attributes. Geometric facial attributes is extracted from the depth image and face recognition is performed by fusing both of these. Here ADM is then combined with the DCT with DPA for identification.
Keywords: Face recognition, saliency, entropy, RGB-D, Kinect, Discrete Cosine Transform, and Discriminant Power Analysis.