Wavelet based Multimedia Data Compression Techniques

Abstract

The field of Multimedia computing has become major research into data compression. Multimedia computing is essentially the integration of audio, video (high or low definition), graphics, photography, text and software; has led to the development of various handling applications employing data compression. Frequently, a large amount of data created by these handlers has to be stored and often transferred from device/s to device/s over the Internet and various other transmission media. Accordingly, the often “bigger” data requires wider bandwidth, and longer processing and storage times. This research focused on multiple data compression techniques, depending on the expected quality of “de-compressed” data and in-image processing, in particular. A large amount of research involved development of image compression coding technologies and standards. The coding technologies applied image compression technologies employing: the Human Visual System (HVS) model including auditory characteristics, Continuous Wavelet Transform (CWT), Image Enhancement and Fractal theory. Data compression is most widely used in spreadsheet applications, backup utilities, graphics and database management system. Several reasons were investigated as to why multimedia signals require data to be compressed and as stated previously, the principal reasons include storage that needs to be managed by the quality of the data. Suitable use of Evolutionary Algorithm and effective use of the wavelets transform based Human Visual System in data compression is investigated in this thesis. Firstly, two approaches of data compression techniques are developed with the use of the aforementioned features. The first method aims to overcome the issues with the quality of the compressed images for cases when the original file cannot be recreated according to its uncompressed version while reducing bits required to transmit and store an image file. The second approach proposes a dedicated compression technique in which the input is obtained in its actual shape without loss of data, as after decoding the data can be restored in its original and exact form. These approaches were enhanced by developing the Quality Enhancement Techniques for effective imperceptibility measurement of compressed data. To assess the utility and viability of techniques proposed, they all have been empirically validated. This research considered data compression as a solution in the retrieval, store and transmission of data that ensures a balance between compression times, quality of compression and compression rate that utilizes the HVS system model.