Improved Water Algorithm Based on K - Means Clustering Character Segmentation Method
Haibo Lin, Haoran Liu, Shuaiyi Li, Shanli Tao, Shuang Song · 2020
When the adhesion of the adjacent two characters happen., if using traditional dripping algorithm., can make the water droplets in the vertical drop down two conglutination., character split phenomenon. If the character split character for character recognition operation., can't get the correct recognition as a result., when recognition accuracy is greatly reduced. Therefore., in order to solve this problem., based on the traditional algorithm on the basis of drop water., through the K - Means clustering method to determine the initial point of water., put forward the character segmentation method based on improved drop algorithm., reduce the error of the characters segmentation and segmentation.