New Algorithms for Perceptual-based Image Compression and Anaglyph Stereo Video Compression

Autor: Adhatus Solichah Ahmadiyah
Rok vydání: 2013
Druh dokumentu: 學位論文 ; thesis
Popis: 101
Compression is an important issue in digital applications as it can reduce the storage size and minimize the bandwith usage for file transfer over networks. In this document, two topics about compression is discussed. The first topic presents a novel block-based image coding algorithm which integrates tree-structured multitree dictionary and perceptual-based rate distortion optimization scheme. While multitree dictionary is employed to support a very large number of different tilings, perceptual-based rate distortion optimization utilizes the SSIM metric, instead of popular MSE metric, to allocate bit rate according to human visual system. Experimental results show that our proposed method outperforms many existing techniques in both subjective and objective image quality measures. The second topic presents two novel end-to-end stereo video compression pipelines consisting of single-sensor digital camera pairs, the legacy consumer-grade video decoders, and anaglyph displays. As 3D videos contain a large amount of data, efficient compression methods to distribute streams over the current communication infrastructure are highly required. We proposed two methods to transmit a single encoded stream containing only required data to create anaglyph video from single-sensor camera pairs. The experimental results demonstrate the superior performance of our proposed methods over the traditional one by achieving up to 4.66 dB improvement in terms of Composite PSNR.
Databáze: Networked Digital Library of Theses & Dissertations
načítá se...