Image enhancement in cell phone cameras using wavelet transform

In this research work a novel color correction algorithm for enhancing the color of digital images captured from low resolution cell phone cameras is proposed here. The procedure involves capturing images using various cell phones like Samsung, Nokia and Lenova having resolution of 2 mega pixels, 3 mega pixels and 5 mega pixels respectively. The scope of the work includes image acquisition using cell phones of various resolutions. The captured images are subjected to histogram analysis as an indication of preprocessing. The histogram equalization and clipping is done to increase the global contrast and reduce the dynamic range. Then Transportation Map Registration (TMR) filter is used for removing any artifacts present in the extracted images. Feed Forward Neural Network (FNN) trained with Back Propagation Algorithm (BPA) is used for efficient retrieval of the images based on its content. The wavelet transform is used to estimate the Gaussian mixture model (GMM), Resolution Synthesis Color Correction (RSCC) and Generalized Gaussian mixture model (GGMM). The output of this enhancement algorithm is used to get an optimized result. The performance of the proposed method to other existing color correction algorithms on cell phone camera (various resolution) images obtained from different sources are compared. The subjective and objective quality analysis is carried out to substantiate the improved quality of the proposed algorithm over the existing methods.