Абстрактный

Frequency Domain Approaches for Fingerprint Based Gender Classification

Suchita Tarare, Akhil Anjikar, Mukesh Raghuwanshi

Fingerprint based gender classification can be studied using frequency domain approaches like discrete wavelet transform (DWT), discrete cosine transform (DCT) and block-based discrete cosine transform (BBDCT). These give the energy based features of fingerprint. This paper is based on the “Frequency Domain Approaches for Fingerprint Based Gender Classification”, where fingerprint is used to identify gender of person. Dataset of some male and female fingerprints is divided into training and testing sample. All training sample images are pre-processed and feature database is created by extracting features of all images using frequency domain technique (dwt, dct, bbdct). Testing sample is used for testing purpose, testing fingerprint is processed in same way as training sample images to get feature vector. Using knn classifier testing fingerprint feature vector is compared with training sample feature database and classified as male or female fingerprint.