Gender Classification using Central Fibonacci Weighted Neighborhood Pattern Flooding Binary Matrix (CFWNP_FBM) Shape Primitive Features

Abstract

Gender Classification from facial images is an open research area with wide range of computer vision applications like security, biometrics and human computer interaction applications. In the proposed method the LL band image of facial image is obtained by using wavelet then on this image Fibonacci Weighted Neighborhood Central pixel Flood binary Matrix is computed and then shape features are evaluated. SVM method uses these shape features for gender classification. The proposed approach has been experimented on FG NET database. The experimental results has shown the more accuracy compared to with other existing methods.