查詢結果分析
相關文獻
- 基於RRO-NRDPWT心電圖資料壓縮法中之線性PRD-CR規劃
- 心電圖(ECG)資料壓縮使用基於SAA架構之非遞迴式小波轉換
- 一個適用於近似週期信號的新自適性向量量化法及其在心電圖資料壓縮上的應用
- 利用小波與類神經網路進行心電圖特徵擷取與病症分類
- 以小波轉換與自適性向量量化作近乎無失真心電圖資料壓縮
- Data Compression of Voltage Flicker Waveforms Using Discrete Wavelet Transform
- 應用離散小波轉換壓縮電力品質資料波形
- 新型衛星影像壓縮法則與實現
- A High-Speed /Ultra Low-Power Architecture for Two-Dimensional Discrete Wavelet Transform
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
頁籤選單縮合
題 名 | 心電圖(ECG)資料壓縮使用基於SAA架構之非遞迴式小波轉換=ECG Data Compression Using SAA-Based Non-Recursive Discrete Periodized Wavelet Transform |
---|---|
作 者 | 辜正同; | 書刊名 | 慈惠學術專刊 |
卷 期 | 1 民94.10 |
頁 次 | 頁102-114 |
分類號 | 410.1644 |
關鍵詞 | 心電圖; 資料壓縮; 小波轉換; 非遞迴式小波轉換; 可回復性四捨五入線性轉換; ECG; Data compression; Wavelet transform; NRDPWT; RROLT; |
語 文 | 中文(Chinese) |
中文摘要 | 傳統遞過式離散小波轉換 (recursive discrete wavelet transform; RDWT) 已成功應用於心電圖 (Electrocardiogram; ECG) 資訊的壓縮 [1-8],然RDWT是基於遞迴式金字塔演算法 (recursive pyramid algorithm) 的運算方式,在實現時會面臨需要冗長的時問延遲 (long latency) 及位元長度成長效應 (word-length growth effect) 等兩困難,此會降低需高階分解時 的硬體實現效益。而ECG的資料壓縮則需要高階分解才可得到高壓縮比[1][4][8]。 本論文提出使用區段累加演算 (segment accumulation algorithm: SAA) 架構之非遞迴式離散週期性小波轉換法 (non-recursive discrete periodized wavelet transform. NRDPWT) [9]於心電圖資料壓縮之應用。SAA-based NRDPWT是一種適用可回復性四捨五入線性轉換理論 (reversible round-off linear transformation theorem; RROLT)[10] 之一新的小波轉換演算法,原被開發於圓型辨識之應用,可有效抑制高階分解中的位元長度成長效應且適於VLSI實現。本論文擴展SAA-based NRDPWT於全頻道之分解,再結合RROLT理論尋找 一組最佳量化參數。以此一新演算法取代傳統wavelet-based ECG壓縮系統中之RDWT 運算並提出一使用全高頻資料之ECG即時壓縮技術。 實驗證明此一全高頻壓縮技術可得比傳統wavelet-based方法更高的壓縮比 (Compression Ratio, CR) 及更低的失真度 (Percentage Root mean square Difference, PRD)。在不改變後級不失真壓縮法的情況下,對MIT- BIH record: 117之ECG訊號,其壓縮率可達 CR=23.47及PRD=2.82%之高壓縮效能。 |
英文摘要 | Traditional recursive discrete wavelet transform (RDWT) had been successfully applied in electrocardiogram (ECG) data compression [1-8]. For high compression ratio, six stage decompositions are generally required in wavelet-based ECG compression system [1,4,8]. However the RDWT based on the recursive pyramid algorithm is inhered in two drawbacks, the word-length growth effect and long time latency, which will degrade the hardware efficiency of high octave decomposition. This paper proposes a new ECG compression algorithm based on the non-recursive discrete periodized wavelet transform (NRDPWT) and the segment accumulation algorithm (SAA) architecture [9]. SAA-based NRDPWT is a new wavelet transform suitable for the application of the reversible round-off linear transformation (RROLT) theorem [10]. It is used for pattern recognition at the first purpose. It will constrain to the word-length growth effect of the high octave decomposition and fits for VLSI implement. This paper extends SAA-based NRDPWT in full channel decomposing, and then added to the RROLT theorem to look for the combination of the optimistic quantization parameter. This new algorithm will be substituted for the RDWT operation unit on traditional wavelet-based ECG compression system, and proposed the ECG real-time compression technique of using the all high frequency data. The experimental results show that this all high frequency compression technique has higher compression ratio (CR) and lower percentage root mean square difference (PRD) than the previously proposed ECG compression algorithms. The ECG test signal MIT-BIH: 117 gets CR=23.47 and PDR=2.82% in high compression efficiency, it doesn't change the last stage of lossless coding on traditional wavelet-based ECG compression system. |
本系統中英文摘要資訊取自各篇刊載內容。