The influence of the selection of distribution of grapheme image dots on symbol recognition using the shape context method

The paper is concerned with the shape and image recognition method of the shorthand reports’ images. The method is studied with the aim of optimal algorithm determination for the symbol image dots’ distribution. This work contains a description of three grapheme image dots selection methods and results of their influence on the effectiveness of recognition. The research is based on the collection of shorthand reports of the XIX th century. All collected images were segmented and transferred into a binary form. Every method was tested by the control image set. The size of the control image set contains 300 symbols. The obtained results assisted in determination of the most suitable dot selection algorithm for the shape and image recognition method. This method is a selection method based on structural features.