Question correction method, electronic equipment and storage medium

Abstract

The invention discloses a question correcting method, electronic equipment and a storage medium, and belongs to the technical field of intelligent teaching, and the question correcting method disclosed by the embodiment of the invention comprises the following steps: receiving a correcting triggering operation; identifying a first question area in the collected first image; standard answers of the questions in the first question area are determined, question correction is performed on the to-be-corrected answers in the first question area based on the standard answers, and a correction result of the first question area is obtained; projecting and displaying a correction result of the first question area based on the position of the first question area in the first image; independently storing the questions in the first question area and the corresponding standard answers; performing pose correction on the second image, and identifying a second question area in the second image after pose correction; performing similarity judgment on the questions in the second question area and the independently stored questions; and correcting the questions in the second question area based on a similarity judgment result, and projecting and displaying a correction result of the second question area.

Description

Technical Field

[0001] The present application relates to the field of intelligent teaching technology, and in particular to a question grading method, an electronic device and a storage medium. Background Art

[0002] With the development of intelligent teaching technology, question grading equipment combined with micro-projection display technology has emerged. Users can place the question paper at the specified position of the device, and the device will capture the image of the question paper, use the built-in algorithm to correct the questions in the image, and superimpose the correction results on the real question paper with AR (Augmented Reality) effect through micro-projection display technology, so that users can intuitively view the correction results.

[0003] However, in the process of grading questions based on such devices, if the user moves the question paper midway, the device often needs to re-execute the entire grading process, which not only wastes time but also reduces the fluency of question grading and user experience. Summary of the invention

[0004] In a first aspect, the present application provides a method for correcting a question, comprising: Receive revision trigger action; identifying a first topic area in the captured first image; Determine standard answers to questions in the first question area, and correct the answers to be corrected in the first question area based on the standard answers to obtain correction results for the first question area; Based on the position of the first topic area in the first image, projecting and displaying the correction result of the first topic area; separately storing the questions and corresponding standard answers in the first question area; Performing posture correction on the second image, and identifying a second question area in the second image after the posture correction; Performing similarity judgment between the topics in the second topic area and the topics stored separately; The topics in the second topic area are corrected based on the result of the similarity judgment, and the correction results of the second topic area are projected and displayed.

[0005] In some embodiments, it also includes: After each projection displays the correction result, the overlapping images of the questions and the correction result in the question area are collected and retained; Acquire new images; If it is determined that the new image does not contain the topic area, the retained overlapping image is projected and displayed.

[0006] In some embodiments, grading the questions in the second question area based on the result of similarity judgment and projecting and displaying the grading result of the second question area includes: Determine that there are matching questions, and obtain the grading result of the matching questions in the first question area; Identify the positions of the matching questions, and project and display the operation frames and the grading results of the matching questions in the first question area at the positions of the matching questions.

[0007] In some embodiments, grading the questions in the second question area based on the result of similarity judgment and projecting and displaying the grading result of the second question area includes: Determine that there are matching questions, and determine the grading result of the matching questions in the second question area based on the standard answers stored separately for the matching questions; Identify the positions of the matching questions, and project and display the operation frames and the grading results of the matching questions in the second question area at the positions of the matching questions.

[0008] In some embodiments, determining the standard answers of the questions in the first question area includes: Search all historical questions in the historical question database and obtain the answers of the matching questions in the historical question database as the standard answers; For the questions that do not match successfully, obtain the corresponding standard answers based on the question-solving language model.

[0009] In some embodiments, the questions stored separately are stored in a separate area of the historical question database, and the query speed of the separate area is faster than the query speed of all historical questions in the historical question database.

[0010] In some embodiments, it further includes: Add the questions in the first question area to the historical questions in the historical question database, and replace the content stored in the separate area with the questions in the second question area and the corresponding standard answers.

[0011] In a second aspect, the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the question grading method as described in any one of the above is implemented.

[0012] In a third aspect, the present application further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the question grading method as described in any one of the above is implemented.

[0013] In a fourth aspect, the present application further provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned question correction methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the present application or related technologies, the following is a brief introduction to the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 It is one of the schematic diagrams of the effect of projecting the correction result on the question paper in the related technology.

[0016] Figure 2 This is one of the flow charts of the question correction method provided in the embodiment of the present application.

[0017] Figure 3 This is the second schematic diagram of the effect of projecting the correction results onto the question paper in the related art.

[0018] Figure 4 This is the third schematic diagram of the effect of projecting the correction result onto the question paper in the related art.

[0019] Figure 5 This is the second flow chart of the question correction method provided in the embodiment of the present application.

[0020] Figure 6 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first" and "second" are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally represents that the objects associated with each other are in an "or" relationship.

[0023] At present, the traditional way of marking and reviewing questions mainly relies on manual operation. Teaching assistants or students themselves need to review each question one by one and manually mark and annotate them. Although this method is accurate, it is inefficient. Especially when the marking task is heavy, the burden of manual marking increases significantly.

[0024] In order to reduce the burden of manual grading, automated question grading methods came into being.

[0025] Among them, micro-projection display technology is a micro laser projection technology that combines two core technologies, micro-electromechanical technology and laser. Micro-projection display technology can effectively improve the projection quality while reducing the size of the projector. The development of micro-projection display technology has enabled the question grading equipment based on micro-projection display technology to be put into use.

[0026] For example, some learning machines that can realize automatic question correction combine micro-projection display and infrared control technologies. At the display level, such devices rely on micro-projection display technology to project the display screen, and at the interaction level, they rely on infrared operation technology to realize hand touch function and pen tip touch function.

[0027] When this type of device is used to correct questions, the implementation process is as follows: First, the user can place the question paper at the location specified by the device and trigger the device's camera correction function. Then, the device can capture the image of the question paper through the camera and use the built-in algorithm to correct the questions in the image to obtain the correction results. Then, the device can use micro-projection display technology to superimpose the correction results on the real question paper with AR effects. For example, Figure 1 It is one of the schematic diagrams of the effect of projecting the correction result on the question paper in the related art, such as Figure 1 As shown in the figure, the green frame and the correct mark, as well as the red frame and the wrong mark are the correction results projected on the question paper by the device through the micro-projection display technology. In this way, the user can intuitively view the correction results of each question on the question paper.

[0028] However, in the process of grading questions based on such devices, if the user moves the question paper midway, for example, taking the question paper out of the capture area of ​​the device and then putting it back into the capture area of ​​the device, or moving the question paper within the capture area of ​​the device, the device will often determine that it has captured an image of a new question paper that needs to be graded. At this time, the device will re-execute the complete grading process, which not only wastes time, but also reduces the fluency of question grading and user experience.

[0029] To this end, the present application provides a question grading method, which separately stores the questions and corresponding standard answers in the first question area in the first image, and makes a similarity judgment between the questions in the second question area in the second image and the separately stored questions, and corrects the questions in the second question area based on the result of the similarity judgment. As a result, during the question grading process, it is possible to quickly determine whether the questions in the second image are questions that have been corrected by making a similarity judgment with the separately stored questions, thereby avoiding the waste of time and computing resources caused by repeatedly grading the same questions. Even if the user moves the question paper in the middle, it can respond quickly following the movement of the question paper and project the correction results at the location of the question, thereby optimizing the fluency of question grading and user experience.

[0030] The question correction method provided in the embodiments of the present application can be applied to electronic devices, specifically to terminals with shooting and projection functions, for example, mobile phones, tablet computers, learning machines, dictionary pens, wearable devices, augmented reality / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA) and other terminals with shooting and projection functions. The embodiments of the present application do not impose any restrictions on the specific types of terminals.

[0031] For example, the terminal can be a station (STAION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA) device, a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a computer, a laptop computer, a handheld communication device, a handheld computing device, and / or other devices for communicating on a wireless system and a next-generation communication system, such as a mobile terminal in a 5G network, a mobile terminal in a future evolved Public Land Mobile Network (PLMN) or a mobile terminal in a future evolved Non-terrestrial Network (NTN), etc.

[0032] As an example but not limitation, when the terminal is a wearable device, the wearable device can also be a general term for wearable devices that are developed by applying wearable technology to intelligently design daily wearables, such as watches equipped with far-field communication modules and / or near-field communication modules, AR head-mounted display devices, VR head-mounted display devices, or MR (Mixed Reality) head-mounted display devices, etc.

[0033] Figure 2 This is one of the flow charts of the question correction method provided in the embodiment of the present application. Figure 2 The question correction method shown can be applied to the electronic device described above, and the method includes steps 210, 220, 230, 240, 250, 260, 270 and 280. The method flow steps are only a possible implementation of the present application.

[0034] Step 210: receiving a correction trigger operation.

[0035] Here, the correction trigger operation is used to trigger the electronic device to perform automatic question correction.

[0036] The correction trigger operation can be a trigger operation on the correction icon in the UI (User Interface) displayed by the electronic device, such as clicking the correction icon, or long pressing the correction icon, etc.; the correction trigger operation can also be a trigger operation on the physical buttons of the electronic device, such as tapping the shooting button, tapping the correction function button, etc., which is not specifically limited in the embodiments of the present application.

[0037] It is understandable that when the electronic device receives a correction trigger operation, the question correction process of the electronic device can be triggered.

[0038] Step 220: Identify a first topic area in the captured first image.

[0039] Specifically, after receiving the correction trigger operation, the electronic device can start the image acquisition function and continue to acquire images. Through image acquisition, the image of the question paper placed within the image acquisition range of the electronic device can be acquired, thereby capturing the question content in the question paper.

[0040] In the embodiment of the present application, the first image captured after receiving the correction trigger operation is recorded as the first image.

[0041] For the first image obtained by acquisition, the title area in the first image can be identified. The title area here refers to the area where the title contained in the image is located, and can also be understood as the effective range where the title contained in the image is located. It can be understood that each title corresponds to a title area, and the number of titles in the image is the same as the number of title areas. For example, Figure 1 In the effect diagram shown, the area framed by the green and red boxes can be regarded as the question area. Figure 1 The green box, red box, correct number and wrong number in the answer sheet are all the results of projection display. They do not exist on the question paper.

[0042] Here, in order to realize the recognition of the question area, it can be realized by calling a pre-trained area recognition model. The area recognition model called here can be implemented by an OCR (Optical Character Recognition) tool, for example, it can be implemented by an open source OCR tool such as Tesseract and EasyOCR. The embodiment of the present application does not make any specific limitation on this.

[0043] For the convenience of explanation, the title area in the first image is referred to as the first title area below.

[0044] Step 230, determining standard answers to the questions in the first question area, and correcting the answers to be corrected in the first question area based on the standard answers to obtain correction results for the first question area.

[0045] Specifically, after obtaining the first question area of ​​the first image, the standard answer corresponding to the question can be obtained based on the question in the first question area. Here, for any first question area, the questions in the first question area can be matched with the questions in the pre-built question database, and the standard answers of the successfully matched questions in the question database are used as the standard answers to the questions in the first question area; or, the questions in the first question area can be input into a pre-trained test answer model, and the answers output by the test answer model for the questions are used as the standard answers to the questions. The embodiments of the present application do not specifically limit this.

[0046] After obtaining the standard answer, the question correction can be performed for the answer to be corrected in the first question area. Here, the question correction can be specifically to compare whether the standard answer and the answer to be corrected are consistent. If they are consistent, the correction result can be determined to be correct, and if they are inconsistent, the correction result can be determined to be wrong.

[0047] Step 240 : Based on the position of the first topic area in the first image, project and display the correction result of the first topic area.

[0048] Specifically, after obtaining the correction result of the first question area, the position of the first question area in the first image can be used as the projection position, and the correction result of the first question area can be used as the projection content and projected onto the question paper, thereby achieving an AR display effect of superimposing the correction result of each question on the question paper, so that the user can intuitively view the correction result of each question through the question paper.

[0049] Step 250, separately store the questions in the first question area and the corresponding standard answers.

[0050] Specifically, after the question correction for the first image is completed, the questions in the first question area used for correction and their corresponding standard answers can be stored separately. Here, storing separately means that the questions in the first question area and their standard answers are distinguished from other data and stored in an independent storage area. It can be understood that simply storing the questions and standard answers is more convenient and efficient when reading.

[0051] Step 260: perform posture correction on the second image, and identify a second question area in the second image after posture correction.

[0052] Specifically, after completing the question correction and the correction result projection for the first image, the electronic device can continue to capture images. In the embodiment of the present application, the image captured after the first image is recorded as the second image. It can be understood that the second image can be the second image captured after receiving the correction operation, or it can be the third, fourth or even later image captured after receiving the correction operation.

[0053] In the actual scenario of question correction, the user usually performs the correction trigger operation after placing the question paper properly. Therefore, when the first image is captured after receiving the correction trigger operation, the question paper is usually placed properly, and the question paper in the first image thus captured is usually a regular square. After the correction trigger operation, the user may move the placement of the question paper, which causes the placement of the question paper to no longer be correct. The question paper in the second image captured by the camera of the electronic device may appear as an irregular quadrilateral, such as a trapezoid.

[0054] In this case, in order to ensure the reliability of the question correction for the second image, the second image can be subjected to posture correction after the second image is acquired. The posture correction here is used to correct the posture of the question paper in the second image. In the second image after the posture correction, the posture of the question paper is correct, which makes it easier to identify the question area in the question paper. The posture correction can be implemented by the image warp algorithm provided by opencv, or it can be implemented by the posture correction algorithm based on geometric constraints, which is not specifically limited in the embodiments of the present application.

[0055] That is, in the embodiment of the present application, posture correction may not be performed on the first image, but only on the second image. This avoids the problem of question papers being placed improperly due to manual movement during the grading process, which affects the grading process. At the same time, it avoids redundant posture correction operations, thereby further improving the efficiency of question grading.

[0056] After completing the posture correction for the second image, the title area recognition can be performed for the second image after the posture correction. The method of performing the title area recognition for the second image is the same as the method of performing the title area recognition for the first image in step 220, and will not be repeated here. For the sake of distinction, in the embodiment of the present application, the title area in the second image obtained by performing the title area recognition for the second image after the posture correction is recorded as the second title area.

[0057] Step 270: Perform similarity determination between the topics in the second topic area and the separately stored topics.

[0058] Different from the correction method for the first topic area in step 230 after obtaining the first topic area, in the embodiment of the present application, after obtaining the second topic area, the similarity judgment can be directly performed on the topics in the second topic area and the separately stored topics.

[0059] Here, the separately stored questions are questions that have been corrected after the current correction trigger operation. The questions in the second question area are judged for similarity with the separately stored questions, that is, whether the questions in the second question area are questions that have been corrected after the current correction trigger operation. Among them, the similarity judgment is used to judge whether the questions in the second question area and the separately stored questions are the same questions from the perspective of question similarity. The question similarity here can be expressed as the similarity between the questions in the second question area and the separately stored questions, which can be specifically reflected in the semantic similarity between the questions in the second question area and the separately stored questions, or can be reflected in the consistency in character sequence between the questions in the second question area and the separately stored questions, etc. The embodiments of the present application do not make specific limitations on this.

[0060] It is understandable that the separately stored questions reflect the questions that have been corrected after the current correction trigger operation, and the number of questions stored separately is greatly reduced compared to all correctable questions. Therefore, the amount of calculation required for similarity judgment between the questions in the second question area and the separately stored questions is very small and the calculation time is very short.

[0061] Step 280: correct the questions in the second question area based on the result of the similarity judgment and project and display the correction results of the second question area.

[0062] Specifically, after obtaining the similarity judgment result, it can be determined whether the title in the second title area matches the separately stored title, that is, it can be determined whether the title in the second title area is a title that has been corrected after the current correction trigger operation.

[0063] In the case where the questions in the second question area match the separately stored questions, it can be determined that the user has not changed the question paper. For the second image, there is no need to re-correct it according to the correction process of the first image. The standard answers of the separately stored questions that match the questions in the second question area can be directly obtained to determine the correction results of the second question area, or the correction results of the separately stored questions that match the questions in the second question area can be directly obtained as the correction results of the second question area.

[0064] If the questions in the second question area do not match the separately stored questions, the user may have changed the question paper during the question grading process. At this time, the second image needs to be re-graded according to the grading process of the first image to determine the grading result of the second question area.

[0065] After obtaining the correction result of the second question area, the position of the second question area in the second image can be used as the projection position, and the correction result of the second question area can be used as the projection content to be projected on the question paper, thereby achieving an AR display effect of superimposing the correction result of each question on the question paper, so that the user can view the correction result of each question intuitively and quickly through the question paper.

[0066] In an embodiment of the present invention, the questions and corresponding standard answers in the first question area in the first image are stored separately, the questions in the second question area in the second image are judged for similarity with the separately stored questions, and the questions in the second question area are corrected based on the result of the similarity judgment. As a result, during the question correction process, it is possible to quickly determine whether the questions in the second image are questions that have been corrected by making a similarity judgment with the separately stored questions, thereby avoiding the waste of time and computing resources caused by repeatedly correcting the same questions. Even if the user moves the question paper in the middle, it can respond quickly following the movement of the question paper and project the correction results at the location of the question, thereby optimizing the smoothness of question correction and user experience.

[0067] During the process of grading questions, the user may remove the question paper from the capture area of ​​the electronic device. At this time, the grading results projected by the device will still be superimposed on the original place and will not disappear. The effect presented is that only the correct and wrong symbols obtained by the grading are visible, but the question itself is not visible. Figure 3 This is the second schematic diagram of the effect of projecting the correction result on the question paper in the related art, such as Figure 3 As shown, after the question paper is taken away, only a residual image of the correction result is shown on the desktop, which affects the user experience.

[0068] In view of the above problems, in some embodiments, the question correction method further includes: After each projection displays the correction result, the overlapping images of the questions and the correction result in the question area are collected and retained; Acquire new images; If it is determined that the new image does not contain the topic area, the retained overlapping image is projected and displayed.

[0069] Specifically, after each projection of the correction result, the electronic device can capture the projected image superimposed with the correction result through the camera, which is referred to as the overlapped image. It can be understood that the overlapped image includes the questions on the question paper and the correction result superimposed on the question, that is, the overlapped image can be understood as the effect image of the correction result projected on the question paper, for example, it can be Figure 1 form.

[0070] After the overlapping image acquisition is completed, a new image can be acquired for the question paper. After the new image acquisition is completed, the question area recognition can be performed on the new image, that is, it can be determined whether there is a question area in the new image. If there is a question area in the new image, it means that there is a question paper in the capture area of ​​the electronic device, and the new image can be used as the second image newly acquired to continue the question correction.

[0071] If the question area does not exist in the new image, it means that there is no question paper in the capture area of ​​the electronic device, that is, the user has taken the question paper away from the capture area of ​​the electronic device. In this case, if the correction results continue to be projected, a residual image of the correction results will appear on the desktop, affecting the user's perception. Therefore, in the embodiment of the present application, in the case where there is no question area in the new image, the above-mentioned overlapping image can be projected and displayed, so that the content projected and displayed on the desktop includes the question and the correction result of the question, and the user can view the question and its correction result in full through the projected content, thereby improving the user experience.

[0072] In addition, during the question grading process, the user may move the question paper within the capture area of ​​the electronic device, that is, the question paper is still within the capture area, but the position of the question paper within the capture area has changed. Since the projection position of the grading result does not change with the position change of the question paper, the question on the question paper and the projected grading result will be misaligned. For example, Figure 4 This is a third schematic diagram of the effect of projecting the correction result on the question paper in the related art, such as Figure 4 As shown, due to the displacement of the question paper, the correction results displayed on the projection are not aligned with the questions on the question paper. The serious misalignment makes it impossible for the user to obtain effective correction results through the projected content, affecting the user experience.

[0073] In view of the above problem, in some embodiments, in step 280, correcting the questions in the second question area based on the result of the similarity judgment and projecting and displaying the corrected results of the second question area includes: Determine whether there is a matching question, and obtain the correction result of the matching question in the first question area; The position of the matching topic is identified, and an assignment frame and a correction result of the matching topic in the first topic area are projected and displayed at the position of the matching topic.

[0074] Specifically, by performing similarity determination on the questions in the second question area and the separately stored questions, a result of similarity determination can be obtained. Here, the result of similarity determination is that there is a matching question, or there is no matching question.

[0075] The existence of matching questions means that there are questions matching the questions in the second question area among the separately stored questions. In other words, the questions in the second question area are questions that have been corrected after the current correction trigger operation. In this case, the second image and the first image captured before the second image may be images captured for the same question paper. Considering that the question paper may be displaced during the question correction process, the second image and the first image may reflect different positions of the same question paper at different times.

[0076] In view of this situation, when it is determined that there is a matching question, the correction result of the matching question in the first question area can be obtained. Here, the correction result of the matching question in the first question area is, that is, the correction result of the matching question after it was corrected before. In addition, the position of the matching question can be re-identified, that is, the position of the matching question in the most recently collected second image can be identified, thereby reflecting the latest position of the matching question. After obtaining the position of the matching question and the correction result of the matching question in the first question area, the position of the matching question can be used as the projection position, and the assignment frame and the correction result of the matching question in the first question area can be used as the projection content for projection display. The assignment frame here is Figure 1 As for the green box and the red box in the figure, it can be understood that different colored assignment boxes may correspond to different grading results, and the assignment box may be used to select the question, and the assignment box may be the border of the question area.

[0077] In an embodiment of the present application, by making a similarity judgment between the questions in the second question area and the separately stored questions, and obtaining the correction results of the matching questions in the first question area when there are matching questions, the correction results of the questions in the second question area can be quickly obtained without correcting the questions. The correction results are projected at the position of the matching questions in the second image, so that the projection position of the correction results can change with the position of the question paper in the latest collected second image. Even if the question paper is displaced, the projected correction results can be aligned with the questions in the question paper, and the user experience can be optimized.

[0078] In some embodiments, in step 280, correcting the questions in the second question area based on the result of the similarity judgment and projecting and displaying the corrected results of the second question area includes: Determining that there is a matching question, and determining a correction result of the matching question in the second question area based on a standard answer stored separately for the matching question; The position of the matching topic is identified, and an assignment frame and a correction result of the matching topic in the second topic area are projected and displayed at the position of the matching topic.

[0079] Specifically, by performing similarity determination on the questions in the second question area and the separately stored questions, a result of similarity determination can be obtained. Here, the result of similarity determination is that there is a matching question, or there is no matching question.

[0080] The existence of matching questions means that there are questions matching the questions in the second question area among the separately stored questions. In other words, the questions in the second question area are questions that have been corrected after the current correction trigger operation.

[0081] In this case, the second image and the first image captured before the second image may be images captured for question papers with the same question but different answers to be corrected. In order to ensure the accuracy of question correction, the standard answer that is stored separately for the matching question can be obtained. It can be understood that before this, the question and its standard answer in the first image are stored separately. In the case of matching the separately stored question, the separately stored standard answer can be directly obtained without the need for additional matching or problem solving operations.

[0082] After obtaining the standard answer to the matching question, the correction result of the second question area can be determined by comparing the standard answer to the matching question with the answer to be corrected in the second question area.

[0083] In addition, the position of the matching question can be re-identified, that is, the position of the matching question in the most recently acquired second image can be identified, thereby reflecting the latest position of the matching question. After obtaining the position of the matching question and the correction result of the matching question in the second question area, the position of the matching question can be used as the projection position, and the homework frame and the correction result of the matching question in the second question area can be used as the projection content for projection display.

[0084] In an embodiment of the present application, by making a similarity judgment between the questions in the second question area and the separately stored questions, and obtaining the correction results of the matching questions in the first question area when there are matching questions, the separately stored standard answers can be directly obtained without correcting the questions to quickly determine the correction results of the questions in the second question area. This can cope with the situation where question papers with the same questions but different answers to be corrected are replaced during the question correction process, effectively improving the efficiency and accuracy of question correction.

[0085] In some embodiments, in step 230, determining the standard answers to the questions in the first question area includes: Search all historical questions in the historical question database and obtain answers to matching questions in the historical question database as standard answers; For questions that are not successfully matched, the corresponding standard answers are obtained based on the test question answering language model.

[0086] Specifically, in order to obtain the standard answers corresponding to the questions, a historical question database can be pre-built. The historical question database here can store all the questions that have been answered in the past and their standard answers. The questions that have been answered in the past referred to here are recorded as historical questions in the embodiment of the present application. The historical questions include all the questions answered after receiving the correction trigger operation this time, and also include all the questions answered after receiving the correction trigger operation before receiving the correction trigger operation this time.

[0087] For the first image captured after receiving the correction trigger operation, i.e., the first image, in order to obtain the standard answer to the question in the first question area in the first image, the historical question database can be searched first, thereby matching the question in the first question area with the historical question in the historical question database, where the matching method can be similarity judgment. The matching result between the question in the first question area and the historical question in the historical question database can be obtained, which is specifically reflected in whether there is a matching question.

[0088] In the case of obtaining matching questions in the historical question database, that is, the questions in the first question area have matching historical questions in the historical question database, the answers to the matching questions in the historical question database can be directly obtained as the standard answers corresponding to the questions in the first question area.

[0089] In the case where no matching questions are obtained from the historical question database, that is, the questions in the first question area do not have matching historical questions in the historical question database, the questions in the first question area are not matched successfully, and the standard answers corresponding to the questions in the first question area cannot be obtained from the historical question database. In this case, the corresponding standard answers can be obtained based on the test question answering language model.

[0090] Here, the test question answering language model is a large language model with the function of answering test questions, such as the iFlytek Spark large model, the Tongyi Qianwen large model, the Zhipu Qingyan large model, etc. The questions in the first question area can be input into the test question answering language model, and the test question answering language model answers the questions in the first question area, thereby obtaining the answers output by the test question answering language model as the standard answers to the questions in the first question area.

[0091] In an embodiment of the present application, in order to obtain the standard answers to the questions in the first question area, the historical question database is searched first, and if no match is successful, the test question answer language model is called. This effectively reduces the frequency of calling the test question answer language model to obtain the standard answers while ensuring that the standard answers can be obtained, thereby avoiding the loss of computing resources and time caused by frequent calls to the test question answer language model, thereby improving the overall question grading efficiency.

[0092] In some embodiments, the separately stored topic is separately stored in a separate area of ​​the historical topic database, and the query speed of the separate area is faster than the query speed of all historical topics in the historical topic database.

[0093] Specifically, the above-mentioned separately stored questions also belong to the historical questions in the historical question database. Considering the questions in the second question area of ​​the second image collected later, there is a high probability that they can match the separately stored questions. In order to improve the efficiency of similarity judgment between the questions in the second question area and the separately stored questions to determine the correction results of the second question area, a separate area can be set when storing the questions in the historical question database.

[0094] Here, the separate area is an area set up for storing questions that match the questions in the second question area. The separate area is isolated from the storage areas of other questions in the historical question database, and the query speed of the separate area is faster than that of the storage areas of other questions in the historical question database. Therefore, when making similarity judgments on the questions in the second question area of ​​the second image and then determining the correction results, the questions used for similarity judgments can be read more quickly from the separate area, thereby achieving rapid determination of the correction results for the second image and improving the efficiency of question correction.

[0095] Here, the separate area may be a separate memory area in the historical question database. Since the data access speed of the memory area is much faster than that of the disk, reading questions from the separate memory area for similarity judgment can speed up the efficiency of question correction for the second image.

[0096] In some embodiments, the method further comprises: The questions in the first question area are added to the historical questions in the historical question database, and the content stored in the separate area is replaced with the questions in the second question area and the corresponding standard answers.

[0097] Specifically, after completing the correction of the second question area in the second image, the questions in the first question area that were originally stored separately can be added to the historical question database as historical questions, and the questions in the second question area and their corresponding standard answers can be stored in a separate area to replace the content originally stored in the separate area.

[0098] It is understandable that, after each subsequent correction of the newly acquired second image and projection of the correction results, the questions originally stored in the separate area can be added to the historical question database, and the questions in the second question area of ​​the new second image and their corresponding standard answers are stored in the separate area to replace the previously stored content. Thus, in the process of each time a new second image is acquired for correction, the questions in the second question area of ​​the new second image are judged similar to the questions in the last acquired image stored in the separate area, that is, the content stored in the separate area is continuously updated as the image acquisition continues, and the content stored in the separate area can be maintained as the questions in the most recently acquired image and their standard answers, which provides convenience for correcting the newly acquired image.

[0099] In some embodiments, Figure 5 This is the second flow chart of the question correction method provided in the embodiment of the present application. Figure 5 As shown, the question correction method may include the following steps: S1. Start the question correction process.

[0100] Specifically, the user can click on the question correction icon on the electronic device interface to open the question correction program, and accordingly, the electronic device starts the question correction process. Then, step S2 is executed.

[0101] S2. Determine whether a correction event has been triggered.

[0102] Here, the basis for judging whether the correction event has been triggered is whether the electronic device has received the correction trigger operation. If the correction trigger operation has not been received, the process proceeds to step S3. If the correction trigger operation has been received, the process proceeds to step S5. The correction trigger operation here is used to trigger the photo correction of the question paper.

[0103] S3. Trigger correction.

[0104] Here, triggering correction means that after the electronic device receives the correction trigger operation, it performs photo correction under the trigger of the correction trigger operation. Step S3 may specifically include steps S301-S305: S301, first topic area identification: After the photo correction is started, the first image taken can be used as the first image, and the title area recognition can be performed on the first image to obtain the first title area in the first image. Then, step S302 can be executed.

[0105] It should be noted that the title area recognition here only refers to the projection area in the image. If there is a title outside the projection area of ​​the image, it is not within the scope of the title area recognition here.

[0106] S302. Judgment on similarity with all historical questions: For each topic in the first topic area, a similarity judgment can be made with each historical topic in the historical topic database to determine whether there is a matching topic in the first topic area, that is, whether there is a historical topic in the historical topic database that is the same as the topic in the first topic area.

[0107] In the case where the same questions exist, the standard answers to the history questions can be obtained as the standard answers to the questions in the first question area, and step S304 is executed. In the case where the same questions do not exist, step S303 can be executed.

[0108] S303, call the model to solve the problem: If the questions in the first question area are not successfully matched with the historical questions, the test question answering language model can be called to obtain the corresponding standard answer. Then, step S304 is executed.

[0109] In addition, after calling the test question answering language model to obtain the standard answers to the questions in the first question area, step S4 may also be executed.

[0110] S304, Question Correction: After obtaining the standard answers to the questions in the first question area, the standard answers may be compared with the answers to be corrected in the first question area, thereby obtaining the correction results of the first question area, and then executing step S305.

[0111] S305, projecting and displaying the correction results: After obtaining the correction result, the correction result of the first topic area can be projected at the position of the first topic area in the first image. In addition, the operation box corresponding to the first topic area can be projected at the position of the first topic area in the first image. After the projection is completed, step S6 can be executed.

[0112] S4. Information storage: After step S303 is completed, the questions and their standard answers in the first question area can be stored in the historical question database to achieve information storage. In addition, a separate area is set inside the historical question database to store the questions and their standard answers in the most recently collected image. After the first image is collected, the questions and their standard answers in the first image can be stored separately in the separate area. And, each subsequent time a second image is collected, the content in the separate area can be replaced based on the questions and their standard answers in the second image.

[0113] S5. After the revision has been triggered: In the case where the correction has been triggered, that is, the electronic device has received a correction trigger operation, the image can be continuously collected and corrected. Step S5 can specifically include steps S501-S505: S501, image posture correction: After the photo correction is started, an image captured after the first image may be used as the second image.

[0114] In order to prevent the user from moving the question paper during the question marking process, causing the question paper to be placed improperly and affecting the marking effect, after the second image is obtained, the second image can be corrected in position and posture. After the position and posture correction, step S502 can be executed.

[0115] S502, second topic area identification: The topic area recognition may be performed on the second image, thereby obtaining the second topic area in the second image. Subsequently, step S503 may be performed.

[0116] S503: Determine similarity with the separately stored questions: For each question in the second question area, similarity judgment can be performed with the questions separately stored in the historical question database, thereby determining whether there are matching questions in the second question area, that is, whether there are historical questions in the separately stored questions that are the same as the questions in the second question area. It can be understood that the separately stored questions here are the questions contained in the image collected last time.

[0117] In the case where the same questions exist, the standard answers to the questions stored separately may be obtained as the standard answers to the questions in the second question area, and step S504 is executed.

[0118] S504, Question Correction: After obtaining the standard answers to the questions in the second question area, the standard answers may be compared with the answers to be corrected in the second question area, thereby obtaining the correction results of the second question area, and then executing step S505.

[0119] S505, projecting and displaying the correction results: After obtaining the correction result, the correction result of the second topic area can be projected at the position of the second topic area in the second image. In addition, the work box corresponding to the second topic area can be projected at the position of the second topic area in the second image. After the projection is completed, step S6 can be executed.

[0120] S6. Determine whether the question paper has been removed: The camera of the electronic device can capture images at a preset frequency. For example, the preset frequency can be 1 frame / s. After capturing a new image, the title area can be identified for the image to determine whether the title area exists in the new image. It can be understood that if the title area exists in the new image, it means that the title paper has not been removed, and step S8 can be executed; if the title area does not exist in the new image, it means that the title paper has been removed and there is no title paper in the capture area of ​​the camera, and step S7 is executed.

[0121] S7, projecting and displaying overlapping images: Here, the overlapping image includes the questions in the question paper and the correction results superimposed on the questions, and the overlapping image can be acquired when step S305 or step S505 is executed. By projecting and displaying the overlapping image, the user can still view the questions and their correction results completely through the projected content when the question paper is removed.

[0122] S8. Projection display of the correction results: The assignment frame and correction result corresponding to the question area can be projected at the position of the question area. Here, in order to avoid displacement of the question paper during the question correction process, the position of the question area can be re-identified before projection, thereby ensuring that the projected assignment frame and correction result can be aligned with the question in the question paper.

[0123] In an embodiment of the present invention, the questions and corresponding standard answers collected in the most recent image are stored separately, a similarity judgment is made between the questions in the second question area in the second image and the separately stored questions, and the questions in the second question area are corrected based on the result of the similarity judgment. As a result, during the question correction process, it is possible to quickly determine whether the questions in the second image are questions that have been corrected by making a similarity judgment with the separately stored questions, thereby avoiding repeated correction of the same questions and causing a waste of time and computing resources.

[0124] Moreover, the position of the question area can be re-identified before projection, so even if the user moves the question paper midway, it can respond quickly following the movement of the question paper and project the correction result at the location of the question, thereby optimizing the smoothness of question correction and user experience.

[0125] Furthermore, when it is determined that the question paper has been removed, the projection displays an overlapping image to avoid only presenting a residual image of the correction result on the desktop, thereby ensuring that the user can still fully view the question and its correction result through the projected content after removing the question paper.

[0126] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the question correction method, which includes: receiving a correction trigger operation; identifying a first question area in the first image collected; determining the standard answer of the question in the first question area, and correcting the answer to be corrected in the first question area based on the standard answer to obtain the correction result of the first question area; projecting and displaying the correction result of the first question area based on the position of the first question area in the first image; separately storing the question in the first question area and the corresponding standard answer; performing posture correction on the second image, and identifying the second question area in the second image after posture correction; performing similarity judgment on the question in the second question area and the separately stored question; correcting the question in the second question area based on the result of the similarity judgment and projecting and displaying the correction result of the second question area.

[0127] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the relevant technology. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0128] On the other hand, the present application also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the question correction method provided by the above methods, which method includes: receiving a correction trigger operation; identifying a first question area in a captured first image; determining a standard answer to a question in the first question area, and correcting the answer to be corrected in the first question area based on the standard answer to obtain a correction result of the first question area; projecting and displaying the correction result of the first question area based on the position of the first question area in the first image; separately storing the questions in the first question area and the corresponding standard answers; performing posture correction on the second image, and identifying a second question area in the second image after posture correction; making a similarity judgment between the questions in the second question area and the separately stored questions; correcting the questions in the second question area based on the result of the similarity judgment and projecting and displaying the correction result of the second question area.

[0129] On the other hand, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the question grading method provided by the above-mentioned methods, the method comprising: receiving a grading trigger operation; identifying a first question area in a captured first image; determining a standard answer to a question in the first question area, and grading the answer to be graded in the first question area based on the standard answer to obtain a grading result for the first question area; projecting and displaying the grading result for the first question area based on the position of the first question area in the first image; separately storing the questions in the first question area and the corresponding standard answers; performing posture correction on the second image, and identifying a second question area in the second image after posture correction; making a similarity judgment between the questions in the second question area and the separately stored questions; grading the questions in the second question area based on the result of the similarity judgment and projecting and displaying the grading result for the second question area.

[0130] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0131] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for correcting a question, comprising: Receive revision trigger action; identifying a first topic area in the captured first image; Determine standard answers to questions in the first question area, and correct the answers to be corrected in the first question area based on the standard answers to obtain correction results for the first question area; Based on the position of the first topic area in the first image, projecting and displaying the correction result of the first topic area; separately storing the questions and corresponding standard answers in the first question area; Performing posture correction on the second image, and identifying a second question area in the second image after the posture correction; Performing similarity judgment between the topics in the second topic area and the topics stored separately; The topics in the second topic area are corrected based on the result of the similarity judgment, and the correction results of the second topic area are projected and displayed.

2. The method for correcting a question according to claim 1, further comprising: After each projection displays the correction result, the overlapping images of the questions and the correction result in the question area are collected and retained; Acquire new images; If it is determined that the new image does not contain the topic area, the retained overlapping image is projected and displayed.

3. The method for correcting questions according to claim 1, wherein: Correcting the topic in the second topic area based on the result of the similarity judgment and projecting and displaying the correction result of the second topic area includes: Determine whether there is a matching question, and obtain the correction result of the matching question in the first question area; The position of the matching topic is identified, and an assignment frame and a correction result of the matching topic in the first topic area are projected and displayed at the position of the matching topic.

4. The method for correcting questions according to claim 1, wherein: Correcting the topic in the second topic area based on the result of the similarity judgment and projecting and displaying the correction result of the second topic area includes: Determining that there is a matching question, and determining a correction result of the matching question in the second question area based on a standard answer stored separately for the matching question; The position of the matching topic is identified, and an assignment frame and a correction result of the matching topic in the second topic area are projected and displayed at the position of the matching topic.

5. The method for correcting questions according to claim 1, wherein: Determine the standard answers for the questions in the first question area, including: Search all historical questions in the historical question database and obtain answers to matching questions in the historical question database as standard answers; For questions that are not successfully matched, the corresponding standard answers are obtained based on the test question answering language model.

6. The method for correcting questions according to claim 5, wherein: The separately stored topic is separately stored in a separate area of ​​the historical topic database, and the query speed of the separate area is faster than the query speed of all historical topics in the historical topic database.

7. The question correction method according to claim 6 further comprises: The questions in the first question area are added to the historical questions in the historical question database, and the content stored in the separate area is replaced with the questions in the second question area and the corresponding standard answers.

8. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the question grading method according to any one of claims 1 to 7 when executing the computer program.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the question marking method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for correcting questions according to any one of claims 1 to 7 is implemented.

Images