Automatic processing method for test paper answer sheets based on machine learning

Abstract

The invention discloses an automatic processing method for test paper answer sheets based on machine learning. The method comprises the steps that firstly a scribbling target detector is trained; thena test paper format is determined, and test papers are generated and printed; each examinee and reviewer fill in corresponding areas, the examinee scribbles an objective question answering region anda subjective question answering region, and the reviewer reviews subjective questions and scribbles a corresponding scoring region; then each completed test paper is scanned into an image, and imageprocessing and regional cutting are performed; target judgment is performed by using the trained scribbling target detector, and then score calculation is carried out by combining correct answers; andfinally statistics and analysis are performed on the test papers, wherein the step further comprises the sub-steps of calculating an average score of the whole batch of test papers, calculating an average score of each test question, calculating distribution conditions of the whole batch of test papers and each test question, performing trend analysis and calling an answer part from a database for teacher or individuals to consult. The automatic processing method is low in cost, high in accuracy and good in flexibility, and can perform efficient statistics and analysis on answering conditionsof the test papers.

Description

technical field [0001] The invention relates to the technical fields of computer vision and machine learning, in particular to a method for automatically processing test paper answer sheets based on machine learning. Background technique [0002] Computer vision has been a research hotspot and difficulty in recent years, and object detection, as the basis of high-level vision tasks, has become a crucial research problem in the field of computer vision. Computer vision is to imitate the visual system of the human eye through machine vision. Research in cognitive psychology and neurobiology has found that humans have a strong ability to perceive objects before recognizing what a specific object is. For a complex image, the human visual system will only pay attention to some parts of it in the initial reaction time, and ignore the rest of the insignificant parts of the image. This further shows that before recognizing a specific object, there is a simple visual attention mecha...

AI Technical Summary

Problems solved by technology

[0003] There are many restrictions on the use of traditional answer sheets that use a cursor reader to identify answers, such as the need for specific recognition hardware, high cost of answer sheets, poor flexibility in the format of test papers, and the need for specific pencil scribbles and cumbersome revisions after scribbling errors.

In addition, the statistics and analysis of the test situation by the teaching staff are inefficient, and the electronic marking is difficult to be practically used in small tests

Images