Automatic composition correcting and scoring method and terminal

Abstract

The embodiment of the invention relates to an automatic composition correcting and scoring method. The method comprises the following steps: receiving an uploaded handwritten composition image; extracting character features of the handwritten composition image through a neural network organization, and removing interference factors near the characters through connected domain analysis processing to obtain a processed handwritten composition image; performing character recognition on the processed handwritten composition image to obtain text content, and performing rhetortic manipulation use recognition, excellent sentence recognition and poem proverb reference recognition on the text content; and respectively scoring the text content in six dimensions of language, content, theme, materialselection, structure and rhetorme.

Description

technical field [0001] The invention relates to the field of education terminals, in particular to a method for automatic correction and scoring of compositions and a terminal. Background technique [0002] In the study of Chinese subjects, writing essays is an essential practical project. Unlike traditional question types such as filling in the blanks, the scoring and correction of Chinese essays requires a lot of energy from teachers or parents, and it is very prone to errors. Judgments and omissions. In the field of automatic composition correction and scoring, many attempts have been made in academia and industry in recent years. Traditional implementation solutions often rely on accurate feature extraction and training data labeling in the field of composition. Feature extraction is aimed at composition scenarios. Manual customization can Features that reflect the quality of composition writing, such as language model-related features, word usage distribution, average ...

AI Technical Summary

Problems solved by technology

The Xiaoyuan Soot app and Baikai.com provide relatively complete composition automatic correction and scoring functions, but they are limited to the field of English composition; IN Classroom provides related functions for Chinese composition correction, but only realizes the scoring function, and cannot Correct typos and typos written by students in the article

[0004] Existing composition correction and scoring related products also generally have the following two shortcomings: existing products are generally aimed at English composition in the K12 scene, and rarely involve correction of Chinese composition

IN classrooms can only give discrete scores for composition by dimension, but cannot give coherent comments using natural language in the way students are most familiar with

Furthermore, the IN classroom only provides the automatic grading function for composition, but cannot automatically identify and correct typos and grammar errors in the composition, so it can only play a limited role in improving students' writing level

To sum up, there is currently no commercial product that can take both composition scoring and error correction into account, and produce coherent comments in natural language

Images