Persian Handwritten Words Detection Based on Features Extraction and Fuzzy Algorithm

In this paper, extracted word features have been used toward offline detection of Persian handwritten words. This has been done using fourteen features in Persian word structure. In this method after required pre-processing, numbers and location of dots versus base line are extracted then the word becomes thin and the feature extraction stage starts. In this stage the words are detected by using extracted features and fuzzy algorithm .In this paper, a practical case has been investigated, in which we use a group of words including names of Iran provinces. The proposed method is faster in comparison to the common algorithms due to the fact that there is no learning procedure in the proposed method. According to the collected database, the presented method is able to detect %82.3 of handwritten Persian words.