利用動態樹狀結構及區塊預測技術的動量絕對值區塊截短編碼研究

一般而言，以空間域爲主的影像壓縮技術，相鄰的影像區塊間通常具有非常高的空間相似性，所以區塊預測技術便可利用來降低影像壓縮的位元率。此外，有許多以區塊爲主的壓縮技術根據運用場合的不同，藉著互相結合運用以達到更好的效能。因此本篇論文提出了一個植基於區塊截短編碼的影像壓縮技術，此方法結合了原本運用在無失真向量量化索引值壓縮上的動態樹狀結構編碼法觀念及一種利用量化表技術的改良式區塊截短編碼壓縮法，不但能降低影像壓縮位元率，且有不錯的重建影像品質。根據實驗結果顯示，我們的方法所需要的位元率非常低且重建品質相當不錯，以灰階512×512像素的Lena影像而言，可以達到PSNR值接近32dB、壓縮位元率0.6 bits/pixel左右的效果。特別的是，相較於其他方法，我們提出的方法是以更簡單的方法達到非常高的效率，計算複雜度相當的小。

關鍵字

區塊預測技術；動量絕對值區塊截短編碼；向量量化；動態樹狀結構編碼法

並列摘要

In spatial domain compression schemes, there is usually a high correlation between the neighboring blocks. Therefore, the Block Prediction Technique can be employed to further reduce the bit rate of a compressed image. Moreover, many block-dominated compression schemes have been combined to achieve good performance according to various applications. Thus, we proposed a method based on Absolute Moment Block Truncation Coding (AMBTC), which combines the concepts of the Dynamic Tree Coding Scheme of Vector Quantization and Block Code Modified-BTC. The experimental results show that the proposed algorithm achieves a significant reduction in the bit rate and yields a reconstructed image of very acceptable quality. For a 512×512 pixel image, ”lean” with 256 gray levels, the quality of the resulting reconstructed image is about 32 dB at 0.6 bits/pixel. More importantly, our proposed method achieves compression efficiency while maintaining a low computational complexity.

並列關鍵字

block prediction technique ； absolute moment block truncation coding ； vector quantization ； dynamic tree coding scheme

國際替代計量

利用動態樹狀結構及區塊預測技術的動量絕對值區塊截短編碼研究

全文下載

主題瀏覽