以小波轉換以及統一的向量量化架構進行失真到無失真的心電圖壓縮

Lossy-to-Lossless ECG Compression Using Wavelet Transform and a Unified Vector Quantization Framework

趙樹年(Shu-Nien Chao)

指導教授 ： 繆紹綱

中原大學/工學院/電子工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400396

摘要

之前我們提出以小波為基礎的向量量化法，應用於失真的心電圖訊號壓縮上，此種方法中的向量量化器結合失真限制的碼簿更新(DCCR)機制，以及一個著名的編碼技術—集合分割排序法(SPIHT)。
先前所提出的方法其效能是眾多最好的方法之一。然而，此方法在無失真壓縮上的表現不及在失真的情況，故在本論文中，我們研究並修正其在無失真壓縮上效率不佳的問題，並且延伸此方法，以便在單一架構上同時允許失真和無失真的壓縮。我們利用著名的9/7濾波器和5/3 整數濾波器，來分別實現失真和無失真壓縮的小波轉換，而DCCR機制的原始設計是用於失真壓縮，現在我們修改使得其亦可使用於無失真的壓縮。另外，我們提出一個新且符合成本效益的編碼方法，以提升SPIHT在較不重要的位元平面上的編碼效能。
在實驗中，我們利用MIT/BIH心律不整的心電圖資料庫以及European ST-T的資料庫當作測試的資料。對於無失真壓縮的編碼效能，根據實驗的結果顯示，本論文所提出的方法改善了直接用SPIHT的方法以及之前我們所提出的方法(VQ加DCCR)各約33％和26％的編碼效益。

關鍵字

失真和無失真 ； SPIHT ； 失真量限制的碼簿更新機制 ； 心電圖壓縮

並列摘要

In a prior work, a wavelet-based vector quantization (VQ) approach was proposed to perform lossy compression of electrocardiogram (ECG) signals. In that approach, a vector quantizer incorporates a distortion-constrained codebook replenishment (DCCR) mechanism and a well-known coding technique called the set partitioning in hierarchical trees (SPIHT).
The coding performance of that work is one of the best. However, it does not perform equally well in lossless compression. In this work, we investigate and fix its coding inefficiency problem in lossless compression and extend it to allow both lossy and lossless compression in a unified coding framework. The well-known 9/7 filters and 5/3 integer filters are used to implement the wavelet transform (WT) for lossy and lossless compression, respectively. The
DCCR mechanism, originally designed for lossy compression, is modified to allow lossless compression as well. In addition, a new and cost-effective coding strategy is proposed to enhance the coding efficiency of SPIHT at the less significant bit representation of a WT coefficient.
ECG records from the MIT/BIH Arrhythmia and European ST-T Databases are selected as test data. In terms of the coding efficiency for lossless compression, experimental results show that the proposed codec improves the direct SPIHT approach and the prior approach (VQ with DCCR) by about 33% and 26%, respectively.

並列關鍵字

ECG compression ； lossy and lossless ； distortion-constrained codebook replenishment ； SPIHT

參考文獻

[1] Z. Lu, D. Y. Kim, and W. A. Pearlman, “Wavelet compression of ECG signals by the set partitioning in hierarchical trees algorithm,” IEEE Trans. Biomed. Eng., vol. 47, no. 7, pp. 849-856, July 2000.

[2] J. Cardenas-Barrera and J. Lorenzo-Ginori, “Mean-shape vector quantizer for ECG signal compression,” IEEE Trans. Biomed. Eng., vol. 46, no. 1, pp. 62-70, Jan. 1999.

[3] S. G. Miaou, H. L. Yen, and C. L. Lin, “Wavelet-based ECG compression using dynamic vector quantization with tree codevectors in single codebook,” IEEE Trans. Biomed. Eng., vol. 49, no. 7, pp. 671-680, July 2002.

[4] R. S. H. Istepanian and A. A. Petrosian, “Optimal zonal wavelet-based ECG data compression for a mobile telecardiology system,” IEEE Trans. Inform. Technol. Biomed., vol. 4, no. 3, pp. 200-211, Sept. 2000.

[5] R. S. H. Istepanian, L. J. Hadjileontiadis, and S. M. Panas, “ECG data compression using wavelets and higher order statistics methods,” IEEE Trans. Inform. Technol. Biomed., vol. 5, no. 2, pp. 108-115, June 2001.