01月08日上课资料-.ppt

* XIV. Applications of Wavelet Transforms Wavelet 所適用的 applications，通常有以下兩大特點： (1) 信號的頻率分佈，會隨著不同的時間(或地點)有較大變異 (2) Multiscale 的分析扮演重要的角色 Larger sampling interval ? ignoring the detail Smaller sampling interval ? requiring a lot of data Wavelet transforms compromise them. 目前，文獻上，80% 以上的應用和 image processing 有關 * (1) Image Compression (JPEG 2000) 傳統 JPEG 架構 8 × 8 4:2:0 DCT 量子化 量子化表 AC係數 Zigzag Scan DC係數 差分編碼 HuffmanCoding HuffmanCoding JPEGfile 檔頭 Image 問題：由於 8 ? 8 的切割，在高壓縮率時會造成 blocking effect RGB to YCbCr * JPEG 2000 架構 Image RGB to YCbCr 4:2:0 Discrete Wavelet Transform Quantization Bit Plane Conversion Fractional Bit-plane coding (Tier 1) Binary arithmetic coding Tier 2 Encoder JPEG 2000 file Quantization Table 檔頭 註：感謝 2010年修課的潘冠臣同學幫忙整理 Tier 1: zero coding, sign coding, magnitude refinement coding, run length coding Tier 2: 用以控制檔案大小 (例如只取比較重要的地方編碼) * Original image DCT-based image compression Wavelet-based image compression CR = 53.4333 CR = 51.3806 註：感謝 2006年修課的黃俊德同學 CR: compression ratio * bpp: bit per pixel (每一點平均需要多少個 bits) PSNR: peak signal to noise ratio (PSNR), see page 411 * 使用 JPEG 2000 做影像壓縮的優點： 所以，在高壓縮率之下，重建的影像仍有不錯的品質 Question: Why JPEG 2000 has not replace the status of JPEG now? C. Christopoulos, A. Skodras, and T. Ebrahimi, “The JPEG2000 still image coding system: An overview,” IEEE Trans. Consumer Electronics, vol. 46, no. 4, pp.1103-1127, Nov. 2000. 參考資料 (1) (2) (3) * B.J. Kim, Z. Xiong, and W.A. Pearlman. “Low bit-rate scalable video coding with 3-D set partitioning in hierarchical trees (3-D SPIHT),” IEEE Trans. Circuits Syst. Video Technol., vol. 10, pp. 1374-1387, 2000. Another Compression Algorithm: SPIHT Using the correlation among high frequency parts in different layers * (2) Edge and Corner Detection (3) Pattern recognition (4) 強調前景，壓縮背景 (a) Feature extraction (Using the wavelet features) (b) Computation Time 和縮小的 pattern 互相比較 (節省運算) * (5) Filter Design 如何不傷到 edge，又能夠將 noise 去除掉？ * x[n] g[n] h[n] x1,L[n