A fast face recognition based on image gradient compensation for feature description

Abstract To improve the efficiency of traditional face recognition techniques, this paper proposes a novel algorithm called Image Gradient Feature Compensation (IGFC). Based on gradients along four directions in an image, fusion and compensation method are implemented to obtain features original image. In study, gradient magnitude maps image calculated directions. Fusion differential produced by fusing multiple ways, they used as variables compensate appropriate coefficients build IGFC feature Subsequently, divided into several blocks calculate concatenated histogram over all blocks, which is turn utilized descriptor for recognition. Principal component analysis (PCA) cut down number dimensions high-dimensional features, recognized Support Vector Machine (SVM) classifier. Finally, proposed superior methods suggested verification studies YALE, ORL, CMU_PIE, FERET databases. When LibSVM parameter was set ‘-s 0 -t 2 -c 16 -g 0.0009765625’, achieved 100% Yale ORL data sets, 92.16% CMU_PIE 74.3% sets. The average time simultaneous completion sets examined 1.93 s, demonstrated 70.71% higher running than CLBP algorithm. Therefore, highly efficient with excellent computational efficiency.