Automatic scoring method and system to output a test score based on a comparison of a current test paper image and a reference test paper image when determining a script color is a designated script color

Patent metadata

Field	Value
Publication number	US-12406513-B2
Application number	US-202318136309-A
Country	US
Kind code	B2
Filing date	Apr 18, 2023
Priority date	Apr 20, 2022
Publication date	Sep 2, 2025
Grant date	Sep 2, 2025

Abstract

An automatic scoring method, performed by a processing device, includes: obtaining a current test paper image, wherein a hand-writing script on the current test paper image has a script color; determining whether the script color is a designated color; determining whether a database stores a reference text string if the script color is not the designated color; aligning the current test paper image with a reference test paper image corresponding to the reference text string and using reference coordinates to compare the current test paper image with the current test paper image to output a test score when the database stores the reference text string; and identifying the hand-writing script to store an identification result as the reference text string and storing position coordinates of the hand-writing script on the current test paper image as the reference coordinates if the script color is the designated color.

First claim

What is claimed is:

1. An automatic scoring method, performed by a processing device, comprising:
obtaining a current test paper image, wherein a hand-writing script on the current test paper image has a script color;
determining whether the script color is a designated color;
determining whether a database stores a reference text string if the script color is not the designated color;
aligning the current test paper image with a reference test paper image corresponding to the reference text string and using a plurality of reference coordinates to compare the current test paper image with the reference test paper image to output a test score when the database stores the reference text string; and
identifying the hand-writing script to store an identification result as the reference text string and storing a plurality of position coordinates of the hand-writing script on the current test paper image as the plurality of reference coordinates if the script color is the designated color.

2. The automatic scoring method according to claim 1 , wherein using the plurality of reference coordinates to compare the current test paper image with the reference test paper image to output the test score comprises:
capturing a plurality of current blocks from the current test paper image according to the plurality of reference coordinates;
identifying the plurality of current blocks to correspondingly obtain a plurality of identifying results and grouping the plurality of identifying results into a current text string; and
calculating the test score according to a matching degree between the current text string and the reference text string.

3. The automatic scoring method according to claim 1 , further comprising extracting a plurality of reference feature points from the reference test paper image, wherein aligning the current test paper image with the reference test paper image corresponding to the reference text string comprises:
extracting a plurality of current feature points from the current test paper image;
obtaining a homography matrix according to the plurality of current feature points and the plurality of reference feature points; and
performing a mapping transformation on the current test paper image according to the homography matrix for the current test paper image to be aligned with the reference test paper image.

4. The automatic scoring method according to claim 1 , further comprising, when the database does not store the reference text string, performing:
storing the current test paper image as a history test paper image;
obtaining a second test paper image as the current test paper image after storing the history test paper image;
determining whether the script color is the designated color again; and
aligning the history test paper image with the reference test paper image corresponding to the reference text string and using the plurality of reference coordinates to compare the history test paper image with the reference test paper image to output another test score after storing the reference text string and the plurality of reference coordinates.

5. An automatic scoring system, comprising:
a database;
a photographing device configured to capture a current test paper image and outputting the current test paper image, wherein a hand-writing script on the current test paper image has a script color; and
a processing device connected to the database and the photographing device, and configured to obtain the current test paper image,
determine whether the script color is a designated color,
determine whether the database stores a reference text string if the script color is not the designated color,
align the current test paper image with a reference test paper image corresponding to the reference text string and use a plurality of reference coordinates to compare the current test paper image with the reference test paper image to output a test score when the database stores the reference text string, and
identify the hand-writing script to store an identification result as the reference text string and store a plurality of position coordinates of the hand-writing script on the current test paper image as the plurality of reference coordinates if the script color is the designated color.

6. The automatic scoring system according to claim 5 , wherein the processing device performing using the plurality of reference coordinates to compare the current test paper image with the reference test paper image to output the test score comprises:
capturing a plurality of current blocks from the current test paper image according to the plurality of reference coordinates;
identifying the plurality of current blocks to correspondingly obtain a plurality of identifying results and grouping the plurality of identifying results into a current text string; and
calculating the test score according to a matching degree between the current text string and the reference text string.

7. The automatic scoring system according to claim 5 , wherein the processing device is further configured to extract a plurality of reference feature points from the reference test paper image, and the processing device performing aligning the current test paper image with the reference test paper image corresponding to the reference text string comprises:
extracting a plurality of current feature points from the current test paper image;
obtaining a homography matrix according to the plurality of current feature points and the plurality of reference feature points; and
performing a mapping transformation on the current test paper image according to the homography matrix for the current test paper image to be aligned with the reference test paper image.

8. The automatic scoring system according to claim 5 , wherein the processing device is further configured to perform:
storing the current test paper image as a history test paper image;
obtaining a second test paper image as the current test paper image after storing the history test paper image;
performing determining whether the script color is the designated color again; and
aligning the history test paper image with the reference test paper image corresponding to the reference text string and using the plurality of reference coordinates to compare the history test paper image with the reference test paper image to output another test score after storing the reference text string and the plurality of reference coordinates.

Assignees

• Aver Information Inc

Inventors

• Lin Chen
• Sie Yun-Long

Classifications

• G06V30/418

  Document matching, e.g. of document images · CPC title

• G06Q50/20 Primary

  Education · CPC title

• G06V30/153 Primary

  using recognition of characters or words · CPC title