Automatic Recognition of Fuzzy Characters in the Transport Task Tables of Scanned Handwritten Student Papers

Students’ handwritten work should be checked during distance learning. Students scan their work and submit it for review using the distance learning system or by email. The challenge is therefore to automate the checking of student work, requiring the recognition of handwritten papers. Recognition automation is especially important for checking large handwritten tables, such as transport tables and simplex tables used in logistics and linear programming problems. The current study presents an analysis of the process of recognising individual handwritten characters found in tables or formulas of scanned student work. The most important is the recognition of digital tabular data simplex-tables and tables used in solving logistical problems. It is necessary to recognize large-volume tables used in solving the transportation problem using the methods of North-western corner, least element, improvement of the transportation plan by Vogel method, improvement of the transportation plan by the method of potentials. A skeletal method of character recognition is considered. Recognized handwritten tables designed for automatic and semi-automatic verification. Methods of skeletonization of selected areas based on the wave approach are analyzed. Alternatively, the paper proposes a method in which skeleton graphs are constructed multiple times for fuzzy characters in the decision tree, using different filtering and skeletonisation options. The software implementation of the proposed method is also proposed.