US 11,721,229 B2
Question correction method, device, electronic equipment and storage medium for oral calculation questions
Fan Shi, Zhejiang (CN); Tao He, Zhejiang (CN); Huan Luo, Zhejiang (CN); and Mingquan Chen, Zhejiang (CN)
Assigned to Hangzhou Dana Technology Inc., Zhejiang (CN)
Appl. No. 16/756,468
Filed by Hangzhou Dana Technology Inc., Zhejiang (CN)
PCT Filed Sep. 11, 2019, PCT No. PCT/CN2019/105321
§ 371(c)(1), (2) Date Apr. 15, 2020,
PCT Pub. No. WO2020/063347, PCT Pub. Date Apr. 2, 2020.
Claims priority of application No. 201811125657.5 (CN), filed on Sep. 26, 2018; and application No. 201811125659.4 (CN), filed on Sep. 26, 2018.
Prior Publication US 2021/0192965 A1, Jun. 24, 2021
Int. Cl. G06N 3/08 (2023.01); G06F 16/245 (2019.01); G09B 7/02 (2006.01); G06F 40/284 (2020.01); G06N 3/04 (2023.01); G09B 19/02 (2006.01); G06V 30/40 (2022.01); G06F 18/21 (2023.01); G06V 30/14 (2022.01); G06V 30/148 (2022.01); G06V 30/19 (2022.01); G06V 10/82 (2022.01); G06V 30/413 (2022.01)
CPC G09B 7/02 (2013.01) [G06F 16/245 (2019.01); G06F 18/21 (2023.01); G06F 40/284 (2020.01); G06N 3/04 (2013.01); G06N 3/08 (2013.01); G06V 10/82 (2022.01); G06V 30/1444 (2022.01); G06V 30/153 (2022.01); G06V 30/19173 (2022.01); G06V 30/40 (2022.01); G06V 30/413 (2022.01); G09B 19/025 (2013.01)] 22 Claims
OG exemplary drawing
 
1. A question correction method for oral calculation questions using an electronic equipment, wherein the electronic equipment is configured to perform an image recognition to implement the question correction method, and comprises a processor and a memory, the memory is configured to store a plurality of trained neural network-based modules, the method comprises:
executing the trained neural network-based modules to implement following steps, wherein the trained neural network-based modules comprises a detection and identification module, a question searching module, a test paper determining module, an oral calculation question determining module and an oral calculation question correcting module, the steps comprises:
step S11: by the detection and identification module, detecting an image of a test paper to be searched, detecting an area of each question to be searched on the image, determining a type of said each question to be searched, and identifying the content of token in a stem in the area of said each question to be searched;
step S12: by the question searching module, obtaining a feature vector of the question to be searched on the image according to the content of token in the stem of said each question to be searched, and conducting search in a question bank according to the feature vector of the question to be searched to find a question most similar to the question to be searched;
step S13: by the test paper determining module, summarizing all the test papers with the question most similar to the question to be searched, and comparing the summarized test papers with a preset condition, if a target test paper satisfying the preset condition is found, the test paper satisfying the preset condition is determined to be a target test paper matching the test paper to be searched;
step S14: by the oral calculation question determining module, in the case that the test paper to be searched comprises a question to be searched in the form of oral calculation question, for said each question to be searched in the form of oral calculation question, the feature vector of the question to be searched and a feature vector of said each question in the target test paper are subjected to shortest editing distance matching to determine the target question that matches the question to be searched in the target test paper, if the type of question of the target question is an oral calculation question, the question to be searched is determined as an oral calculation question to be corrected; and
step S15: by the oral calculation question correcting module, for said each oral calculation question to be corrected, a preset oral calculation engine is used to calculate the oral calculation question to be corrected, and a calculation result of the oral calculation engine is output as an answer to the oral calculation question to be corrected, thereby completing correction of oral calculation questions to be corrected on the test paper.