• 作者： 陳榮昌; 游寶達
• 作者服務機構： 中正大學資訊工程研究所
• 中文摘要： 模糊理論已經被利用於強化傳統的堆疊濾波器，使得該濾彼器具有柔性的計算能力。我們稱此種改良的模糊濾波器為模糊堆疊濾波器。因為模糊堆疊濾波器的基本架構還是因襲堆疊濾波器的分層分解架構，故其計算方式必須提出另外一種等價功能的計算方式，方能提升其濾波的效率。 在本文章中，我們將首先介紹模糊堆疊濾波器的設計理念，使讀者可以感受到模糊濾波器的柔性計算概念。然後，提出一種僅與視窗大小有關的快速演算法來加速模糊濾波器的計算速度，並用數學的方法證明其與原來的方法所得的結果是相同的。最後，以實驗的方式證明此快速演算法在從事影像處理應用問題時非常具有效率。
• 英文摘要： Fuzzy set theory has been successfully applied to extend the filtering capability of conventionalstack filters such that new fuzzy filters, called fuzzy stack filters, can possess soft computing capability.Since each fuzzy stack filter adopts the architecture of threshold decomposition to define its filteringoperation, it requires much computation time in computing the fuzzy value for each threshold level.Therefore, a fast algorithm for computing the filtering result of fuzzy stack filters is needed. In this paper, we first review the basic design concept of fuzzy stack filters and then propose afast filtering algorithm, whose computation complexity depends on the size of the window operator insteadof the size of threshold levels, to compute the filtering result of fuzzy stack filters. We will show thatthe fast filtering algorithm is equivalent to the original algorithm of threshold decomposition by math-ematical proof and computer simulation. We believe that this fast filtering algorithm can provide a basisfor employing fuzzy stack filters in several applications, for instance, image restoration, pattern recognition,etc.
• 中文關鍵字： stack filter; threshold decomposition; positive Boolean function; fuzzy set theory; filtering algorithm
• 英文關鍵字： --