Low-frequency DD-DWT based Illumination Compensated and Color Intensified features for face recognition

The efficacy of face recognition systems is significantly affected by uneven light deviations, incidents over images from different directions. Many prior techniques have excluded color information for normalizing such effects in color images, that thereby compromise recognition performance. However, consideration of color information is significant in extracting color facial features to yield high recognition rates. Thus, a new light invariable color face recognition approach is presented in this work that efficiently normalizes illumination variances and improves color intensities along with appropriate extraction of textural features. The proposed approach includes following steps: Initially, segregation of images into primary spectral components is performed to effectively balance contrast and evenly amplify color intensities. The varying illumination effects are due to low frequency component of image; therefore, Difference of Gaussian high-pass filter based homomorphic filtering is further utilized in HSI (Hue, Saturation, and Intensity) space. Subsequently, extraction of multiscale facial coefficients is performed in double density discrete wavelet transform for vigilant selection of discriminative high-frequency components while compensating low-frequency based illumination effects. The large feature vector space is hereafter projected over eigen subspace to reduce its dimension. The effectiveness of proposed method has been assessed on AR and CMU-PIE color face databases, and attained results are thus contrasted with prior techniques which benefit in establishing its superiority.