Facial acknowledgment is a complex multidimensional building that demands sophisticated computing methods for authentication purpose. In this paper, we present the Integral Normalized Gradient Image (INGI) algorithm with differing normalizing stages. The system includes a novel illumination indifferent preprocessing method, a hybrid Fourier located feature extraction and corresponding process. The Pre-processing method is based in the analysis of the first imaging model, taking everything in mind intrinsic and extrinsic determinants of the human face. Feature extraction contains hybrid Fourier features gleaned from different repetitiveness bands and multiple face models. By deriving Fourier looks from three Fourier domains and three apparent frequency bandwidths, we collected additional complementary lineaments. These features are separately classified using Principal Component and Linear Discriminant Analysis (PCLDA). This approach allows in analyzing a face representation from the various directions for identity recognition. Furthermore, we suggest multiple face models established different eye positions with a unchanging image length. This contributes considerably to enhancing the performance of the projected system. Recognition is realized through Euclidean Distance and Neural Network based classifier, resulting in a acknowledgment accuracy of nearly 89.23% for the Euclidean Distance classifier-based model and 93.40% for Back Propagation Neural Network Classifier.
Author(s) Details:
V. Vijaya Kumari,
Department
of Electronics and Communication Engineering, The Oxford College of
Engineering, Bangalore, India.
Please see the link here: https://stm.bookpi.org/ACST-V9/article/view/12625
