Hassan Rashidi - Computer Engineering Allameh tabatabae University Tehran, Iran
Amin Akbari - Computer Engineering Islamic Azad University Qazvin, Iran

In recent years, different methods have been proposed for automatic classifying texture images in various application domains. In common classification algorithms, each pattern have been assigned only to a category and patterns have been classified into non-connected categories [1]. Thus, they have acceptable performance in compressed and quite distinctclasses. But, in most cases, especially in classifying the texture images such as seabed images and distribution of their in to two set, have overlapping in space of combination and performance of algorithm is not appropriate [2].Due to absence of definite boundaries between different sets of texture images, problem of edge correction has been considered in texture images [3].Problem of precession is one of the challenges in determination of the boundaries in texture images and classification of their.The method proposed in this paper by using fuzzy theory is eliminated problem ofprecession by using fuzzy theory in edge correction part. Rest of paper is organized into four sections. The related work is presented in Section II. Section III is introduced the proposed method. Experimental results are represented in Section IV. At last, the conclusion is in Section V.