Gender classification with KNN by extraction of Haar wavelet features from canny shape fingerprints

Fingerprints give a lot of information about various factors related to an individual. The main characteristic is that they are unique from person to person in many ways. The size, shape, pattern are some of the uniqueness factors seen, so they are area of research and study. Forensic science makes use of different evidences obtained out of which fingerprints are the one to be considered. Fingerprints play a vital role in getting details through the exact identification. Gender identification can also be done easily and efficiently through the fingerprints. Forensic anthropology has gender identification from fingerprints as an important part in order to identify the gender of a criminal and minimize the list of suspects search. Identification of fingerprints is studied and researched a lot in past and is continuously increasing day by day. The gender identification from fingerprints is carried in both spatial domain and frequency domain by applying different techniques. This paper studies frequency domain methods applied for gender identification from fingerprints. A survey of techniques show that DWT is widely used and also in combination with SVD and PCA for gender identification from fingerprints. An overall comparison of frequency domain techniques mainly focusing on DWT and its combinations is presented in this paper with a proposed canny edge detector and Haar DWT based fingerprint gender classification technique.