Gender Recognition From Face Images With Dyadic Wavelet Transform and Local Binary Pattern

Abstract

Gender recognition from facial images plays an important role in biometric applications. Employing Dyadic wavelet Transform (DyWT) and Local Binary Pattern (LBP), we propose a new feature descriptor DyWT-LBP for gender recognition. DyWT is a multi-scale image transformation technique that decomposes an image into a number of sub-bands which separate the features at different scales. DyWT is a kind of translation invariant wavelet transform that has a better potential for detection than Discrete Wavelet Transform (DWT). On the other hand, LBP is a texture descriptor and is known to be the best for representing texture micro-patterns, which play a key role in the discrimination of different objects in an image. For DyWT, we used spline dyadic wavelets (SDW). There exist many types of SDW; we investigated a number of SDWs for finding the best SDW for gender recognition. The dimension of the feature space generated by DyWT-LBP descriptor becomes excessively high. To tackle this problem, we apply a feature subset selection (FSS) technique that not only reduces the number of features significantly but also improves the recognition accuracy. Through a large number of experiments performed on FERET and Multi-PIE databases, we report for DyWT-LBP descriptor the parameter settings, which result in the best accuracy. The proposed system outperforms the stat of the art gender recognition approaches; it achieved a recognition rate of 99.25% and 99.09% on FERET and Multi-PIE databases, respectively.

Nonsubsampled Contourlet Transform Based Descriptors for Gender Recognition

“…We perform comparison with some recent gender recognition systems: Local Gabor Binary Pattern with LDA and SVMAC (LGBP-SVMAC) [1], Local Gabor Binary Pattern with LDA and SVM (LGBP-SVM) [1], Multi-resolution Decision Fusion method (MDF) [7], and the method based on dyadic wavelet transform and LBP (DyWT-LBP) [4]. The results are shown in Figure 11.…”

“…For NSCT-WLD, WLD involves two parameters types: (T, M, S) and the number of non-overlapping rectangular blocks (b) into which an image is divided. For (T, M, S), we tested (4,4,5), (4,6,5), (12,6,15) and (12,4,5) and for b we examined 4x4, 5x5, 6x6, 7x7, and 8x8 blocks. We found that (12,4,5) and 6x6 blocks gave the best results.…”

“…Zheng and Lu [1] gave a comparison of 6 types of features using three classifiers and showed that for FERET database the best accuracy (99.1%) was obtained with features based on local Gabor binary pattern, LAD and SVM with automatic confidence. Ihsan et al [4] used NN classifier and multiscale texture features based on dyadic wavelet trans-form and LBP and reported 99.25% accuracy on FERET.…”