以临界值控制区域函式分水岭演算法为基础之彩色影像分割Color.PDF

以臨界值控制區域函式分水嶺演算法為基礎之 彩色影像分割 Color Image Segmentation based on Watershed Algorithm using Threshold-Controlled Segmentation Function 孫振東 莊書瑋 中國文化大學資訊科學系 中國文化大學資訊科學系 jdsun@.tw Edward.ch@ transform may produce a severe over-segmentation of the image, i.e., many small regions are produced due 摘要 to many local minima in the input image. In this paper we propose an algorithm for eliminating irrelevant minima in the resulting gradient images. 彩色影像分割是一項快速發展的數位影像處 We use gradient-threshold-controlled segmentation 理技術，影像分割目的是將影像切割為數個不相重 function for watershed transformation. The global 疊而各具有相同特性的區域或物件，典型的應用有 gradient threshold is obtained from the first 物件抽取、物件辨識、以物件為基礎的影像壓縮、 derivative of the region-function with respect to 以及以內容為基礎的影像擷取等，近來有許多影像 threshold. The experimental results show that the 分割的相關技術與演算法被提出，其中分水嶺轉換 proposed algorithm can effectively improve segmentation accuracy for different image natures, (Watershed transformation)能有效地用在灰階與彩 such as presence or not of non-homogeneous 色影像分割，然而傳統的分水嶺轉換可能會產生嚴 illumination, texture, contours, and shadows. 重的過度分割，也就是因原影像有許多局部最小點 (Local minima)經轉換後產生許多小區域，本論文 Key words ：Gradient