The Distinction Of Primitive Functions And The Control Of Pattern Recognition Conditions

The answers submitted by students and the correct answers should be compared when the homework is automatically corrected by computers. Since the primitive functions are not unique the students’answers and the correct answers have to be derived and then compared. Therefore the question of identification of primitive functions should be solved. The setting of graphics recognition criteria of the existing graphics recognition algorithms is fixed. It can only correctly recognize and correct the graphical content with doubtless topological relations and position relations. However, some mistakes may occur when it identifies and corrects the graphical content with uncertainty topological relations and position relations.In this paper, the above questions are studied based on the data format and graphics recognition algorithm set by the Intelligent Remote Operation System. Main achievements are as follows:1) Mathematical expressions are decomposed and the algorithm of automatic derivation for complex mathematical expression is designed using multi-level nest and recursion.2) The control flow charts of the derivation of sums, terms and factors are proposed. The program that realizes the automatic derivation of mathematical expression is developed in Delphi.3) Based on the definition of the topological relations, distance relations and direction relations between two graphical elements and the original image recognition algorithms, the control strategy of the graphics recognition criteria is studied. The fuzzy recognition and qualitative identification of the graphics can be better realized by such control strategy.4) The control module of pattern recognition conditions has been developed using Delphi. The module can be used to define and set up corresponding control conditions of pattern recognition according to the special demands of probblems.