الفهرس | Only 14 pages are availabe for public view |
Abstract Video streaming over packet-loss networks faces the challenge that the networks are error-prone,which makes transmission of compressed multimedia video over unreliable network a rich topic of research.As compression is inversely proportional to data redundancy,highly compressed data are more susceptible to errors.Moreover,transmission bandwidth is limited and fluctuating and due to these factors the user device capabilities are varied,and networks are heterogeneous.These challenges made the need for smart adaptation of the pre-coded video a must This study investigates two of the main problems of compressed multimedia video in real-time applications streaming: the first one is the loss of some information during transmission and the second is the noise addition to the transmitted video during transmission. These problems could happen on a bandwidth limited wired networks or wireless networks with error-prone channels.The focus of the thesis is to improve the reconstructed video quality by using suitable techniques that aim at minimizing errors and noise in delivered video frames.Firstly,we consider the problem of suppressing additive noise in video data.To tackle this problem we propose an Intelligent Denoising System for Spatial Video Denoising that is suitable for real-time applications, which requires its parameters to be adjusted according to the frame noise level and making it an important quantity to be estimated. This proposed system has the advantages of consuming less time than the other techniques which is suitable for battery operated devices. Secondly, we present a set of algorithms for error minimization in the received video for both spatial and temporal types of error.Finally,we combined the two approaches for enhancing the video (denoising and error minimization) into a received-video enhancement framework that minimizes both errors and noise in the received video stream.The algorithms used in the proposed framework are characterized by their simplicity and speed which make them suitable for limited capability devices and for real-time applications. To support the novelty of our work, we introduce a comparison between some available researches in the literature with our proposed algorithms,it is obviously clear that our proposed algorithms enjoy the merits of simplicity and speed without scarifying the quality of the video produced.The proposed algorithms yield better visual results with higher Peak-Signal-To-Noise Ratio(PSNR)values gained at less time consumed.This is mainly due to the low complexity of the proposed algorithms. |