Comparison Analysis between implementation of Principal Components Analysis and Haar wavelet as feature extractors in palmprint recognition system
R. Rizal Isnanto, Risma Septiana, Ajub Ajulian Zahra, Ilina Khoirotun Khisan Iskandar, Galih Wasis Wicaksono · 2017
One type of biometrie system is human palmprint. The uniqueness and stability of the principle lines that makes characteristic of the palms reliable to be used as a means of recognition. The principle lines of palmprints are unique so that these can be used in recognition system. In this research, a recognition system using human palm based on Principal Component Analysis (PCA) and Haar wavelet transform for feature extraction. While for its identification, Euclidean distance was implemented. Both, individually, principal components and coefficient obtained from this extraction process then were implemented to determine minimum Euclidean distance. From the first test using PCA, when implementation of 150 training images of 40 respondents, it generated the lowest recognition rate. While, from the second test using Haar wavelet transform, it generated the highest recognition when the test using 7 training data. By these tests, it can be concluded that implementation of Haar wavelet transform gives better recognition rate rather than PCA. Two proposed methods produce good recognition rate but not the best when compared with other methods. The limited number of palms' specimens implemented is suspected of being the cause of its recognition rate is below other methods.