Development of algorithm for segmentation of handwritten text using masks

Abstract

The work highlights the problem of extracting handwritten text on the image for further recognition. Describes common methods of segmentation, allowing to solve this problem: method of color image analysis the binary image segmentation method of fixing level, as well as, a method for the separation and analysis of circuits. The proposed algorithm for text segmentation based on the use of the mask of the original image to simplify the segmentation of text in images with a non-monotonic background. Also the comparative analysis methods is done.