Design And Research Of Prototype System For Auxiliary Scoring Of Mathematical Problem Solving

In recent years,the market scale of K12 extracurricular training has expanded rapidly,and a large amount of capital has poured in,which has added a burden to almost every family.However,it is inevitably closely related to every student.Educational software has become a very important link in the software industry chain in China,with various categories.However,when teachers are correcting papers,multiple choice questions have realized the automatic marking of card painting.However,filling in the blank and answering questions still need to be manually graded,which takes up a lot of manpower and material resources,and it is difficult to avoid the influence of teachers' personal preferences.There are often differences in the grading of a set of papers for different teachers.Therefore,the production of an automatic scoring auxiliary system has a very strong practical significance,this paper carries on the preliminary exploration and implementation of automatic scoring.However,due to the limitations of current technical means,the recognition accuracy of offline handwritten Chinese characters is not high,so we only do auxiliary automatic scoring for high school math solutions and blank-filling problems with more symbols.At present,most of the mathematical solutions and blanks are scanned into the computer first.Based on this scanning data,we input the key steps,values and scoring standards by the grading professionals,and then develop the scoring algorithm to initially realize the automatic scoring function.OCR(Optical Character Recognition)is the main recognition method of Chinese characters and symbols.OCR technology itself has emerged for a long time and its development is relatively mature.This scheme is adopted in this paper to avoid the risks caused by the uncertainty of technology.The system stipulates that the key step of scoring is the characteristic step,the key result of scoring is the key value,and the system only gives points at the characteristic step and the key value.Then the realization of the system mainly adopts the algorithm of feature step and key value location,symbol standardization detection,feature step and key value matching and score setting.In order to further ensure the accuracy of automatic grading,we added the link of manual rechecking,marking those special cases such as unclear handwriting and scribbled handwriting,and then manually grading,and finally achieved a relatively satisfactory result.