傅楸善臺灣大學:資訊工程學研究所沈顥祐Shen, Hou-YoHou-YoShen2010-05-172018-07-052010-05-172018-07-052009U0001-2306200916102200http://ntur.lib.ntu.edu.tw//handle/246246/183361　　我們提出新的階層式去雜訊方法。　王彥鈞的方法在保留細節的處理上表現的不錯，但在執行速度上需耗費可觀的時間。在這次的版本改進了一些演算法及判定機制，使得保留細節以及計算速度有著明顯的提升。　　In this thesis, we propose a new hierarchical noise reduction.　Y. C. Wang’s method is very powerful for noise reduction, but very slow in computation and may be difficult to use. Therefore, we modify some mechanisms to keep more detail and speed up.誌 謝 ii 要 iiibstract iv 錄 v 錄 vihapter 1 Introduction 1.1 Motivation 1.2 Introduction to Noise Types 1.2.1 Digital Image and Color Space Transformation 2.2.2 Noise Characteristics [6] 3hapter 2 Analysis of Noise 5.1 Noise Level Measurement 5.2 Noise Models 5.3 Related Works 7.3.1 Gaussian filter 7.3.2 Bilateral Filtering [10] 7.3.3 Nonlocal Mean Filter 8hapter 3 Original Hierarchical Method 10.1 Introduction to Hierarchical Method 10.1.1 Background of Hierarchical Method 10.1.2 Detailed Description of Hierarchical Method [8] 10.1.3 Different Mask Sizes 18.1.4 Pixel distance 18hapter 4 Our Proposed Method 20.1 Disable Clusters 20.2 Keep Ambiguous Pixels 21.3 Reduced Edge Pixel Directions 21hapter 5 Experiments and Results 23hapter 6 Conclusion and Future Work 36.1 Conclusion 36.2 Future Work 36eference 37application/pdf1768337 bytesapplication/pdfen-US雜訊Denoisethesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/183361/1/ntu-98-R96922085-1.pdf