A Region Growing Based Segmentation for Recognition System Method Implement with Coin based Application

To analyze coin image has to be segmented into two regions once of the coin and the area belonging to the background. We focus on the segmentation task as a preprocessing step for any automated text localization and feature extraction system. Firstly, we present a simple and flexible method for coin segmentation, based on double seed of region growing of coin on Gaussian distributions that allow segmenting various style of coin such as holed coins, triangle coins. Secondly, in the second stage, an active model based segmentation approach extracts precisely the coin from the image with features extraction. Thus, the coin is identified to a monetary class represented by a template coin. The similarity score of two coins is computed from feature constructed by feature point’s results with an identification accuracy of 94.4% on 2238 coin images of 120 classes. Keyword: Coin identification, segmentation, feature extraction, pattern recognition, region growing, shadow detection coins. Nevertheless, both of approach is not suitable on coins which likely show no perfect circularity and holed Coins  Thresholding base approach Basically, Thresholding methods define a specify range of brightness values in the original image and select the pixels within this range as belonging to the foreground, any else it pertain into the background. [3] Developed a coin sorting system call Dagobert. This system is threshold based segment on the assumption that the coins itself are brighter than its background. However, it works well on perfect condition. This approach is no guarantee that the pixels identified by the threshold process are contiguous such as different lighting or image, which is assumed to provide higher responses at coin