基於全域離散餘弦轉換及利用視覺特性之彩色影像編碼最佳化之研究

A PERCEPTUALLY OPTIMIZED CODER FOR COLOR IMAGES BASED ON FULL-BAND DISCRETE COSINE TRANFORM

王耀德(Yao-Te Wang)

指導教授 ： 周俊賢

大同大學/電機資訊學院/通訊工程研究所/碩士(2009年)

摘要

Joint Photographic Experts Group (JPEG) 是目前常用的一個靜態影像壓縮標準，JPEG對於靜態的全彩和灰階影像是一種很有效的壓縮方法。但經JPEG壓縮後，因它將影像分割成8*8大小的區塊去做離散餘弦轉換(DCT)，所以每個區塊影像的最高頻率資訊僅能達到4Hz，未能完全充分利用人類視覺系統對於頻域靈敏度的特性，也因此JPEG在編碼的過程中可能造成編碼的浪費及無法真正反應符合人類視覺特性的視覺品質。
在本篇論文中，提出了二個方法，第一個方法是全域離散餘弦轉換，達到能充分利用人類視覺系統對於頻域靈敏度的特性。第二是利用人類視覺特性的方法去做量化，能更精確的移除人類視覺系統對於影像訊號的多餘量。增進編碼效率及真正反應出符合人類視覺系統特性的視覺品質。
本篇論文的目標是希望，在同樣的壓縮率情況下，能比JPEG更能真正反映出符合人類視覺系統特性的視覺品質。在模擬結果的顯示下，本篇論文提供了一個經由人類視覺特性的彩色影像編碼最佳化的研究。

關鍵字

影像壓縮 ； 全域離散餘弦轉換 ； 人類視覺特性 ； 頻域靈敏度

並列摘要

JPEG is an international standard for color image compression. It provides high coding efficiency with 8×8 discrete cosine transform (DCT). Even though the characteristics of human visual system has been utilized by JPEG standard, the sensitivity of human visual perception is not fully exploited in the DCT domain.
In this paper, as an extension of the JPEG image coding standard, a perceptually optimized coder and full-band discrete cosine transform for color images is proposed. First, we used full-band discrete cosine transform to take advantage of the sensitivity of human visual perception in the DCT domain. Second, we used a visual model for measuring the perceptual redundancy inherent in color images as just noticeable distortion (JND), by which the coding efficiency is increased while maintaining the visual quality of the compressed color images.
Simulation results show the bit rate of the proposed perceptually optimized coder is lower than that of the JPEG coder on the condition of the same visual quality.

並列關鍵字

JPEG ； discrete cosine transform ； human visual system ； just noticeable distortion

參考文獻

[2] N. Jayant, J. Johnston, and R. Safranek, “Signal compression based on models of human perception , ” Proc. IEEE, 81, 1993, pp.1385-1422.

[3] J. Mannos and D. Sakrison, “The effects of a visual fidelity criterion of the encoding of images, ” IEEE Trans. Inform. Theory, 20, 1974, pp.525-536.

[4] D. L. McLaren and D. T. Nguyen, “Removal of subjective redundancy from DCT-coded images, ” IEE Proc.- I, 138, 1991, pp.345-350.

[5] B. Tao, “On adaptive quantization and rate control for MPEG video coding environment ”(Sarnoff Corporation, Princeton, NJ, Tech. Rep., Aug. 1996).

[6] J. Jiang, “A low-cost content-adaptive and rate-controllable near-lossless image codec in DPCM domain, IEEE Trans. Image Processing, ” 9, 2000, pp.543-554.