JISE

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

Journal of Information Science and Engineering, Vol. 31 No. 2, pp. 475-489

A Novel Image Compression Algorithm Based on Multitree Dictionary and Perceptual-based Rate-Distortion Optimization

KAI-LUNG HUA⁺, ADHATUS SOLICHAH AHMADIYAH AND YENI ANISTYASARI
Department of Computer Science and Information Engineering
National Taiwan University of Science and Technology
Taipei, 106 Taiwan
E-mail: {hua; m9915809; d1015801}@mail.ntust.edu.tw

This paper presents a novel block-based image coding algorithm which applies a tree-structured multitree dictionary and a 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.

Keywords: image coding, multitree dictionary, perceptual-based rate-distortion, optimization, SSIM metric

Retrieve PDF document (JISE_201502_07.pdf)