Image and speech signal enhancement in time-frequency domain via adaptive lifting structures

This thesis addresses the problem of image and speech enhancement for various noise environments using adaptive lifting schemes. A new space adaptive lifting scheme algorithm is proposed for 1-D (speech) and 2-D (image) signals. The space adaptive lifting schemes provide better signal representation and better enhancement results. The proposed speech enhancement method aims to remove the noise in order to improve the quality and the intelligibility of the enhanced speech signal. In order to improve the quality of the enhanced speech signal, an auditory model (Critical Bands) is integrated with the proposed speech enhancement method. The single channel estimators are employed for subbband speech enhancement since they are practical. The proposed image enhancement method is based on space adaptive 2-D lifting scheme. The aim of proposed image enhancement method is to remove the noise while retaining significant features of the image.The gray-level noisy images are decomposed into subbands using the proposed space adaptive 2-D lifting scheme algorithm. Spatial domain estimators and wavelet thresholding-based estimators are used for subband image enhancement. The experimental and objective evaluation results show the performance of proposed speech and image enhancement methods.Keywords: speech enhancement, space-adaptive lifting, wavelet, critical band analysis, single channel estimators, image enhancement.