Article
Design and Analysis of Pattern Recognition System Based on Sclera Using Machine Learning Algorithms
A Sclera recognition system based on Wavelet Packet Analysis is described in this paper. The intricate texture of each eye's iris is the most distinctive phenotypic characteristic apparent on a person's face. The apparent texture of a person's Sclera is encoded as a compact sequence of 2-D wavelet packet coefficients that produce a "Sclera code". Two different Sclera codes are compared using exclusively OR comparisons. In this study, we offer a novel multi-resolution technique based on Wavelet Packet Transform (WPT) for analyzing and recognizing the texture of the sclera. This method was inspired by the discovery that the predominant frequencies of the sclera texture are situated in the low- and middle-frequency channels. As a Sclera signature, WPT sub-image coefficients are quantified as 1, 0 or -1 using an adaptive threshold. This signature provides the regional details of several Sclera. By utilizing wavelet packets, the coding size of the Sclera signature is 640 bits. After calculating the signature of the new Sclera pattern, it is compared to the pattern that was previously saved. Identification is achieved by putting the test sclera code into a Multi-Layer Feed Forward Neural Network (MLFNN), which produces the identification result
Full Text Attachment
