Online learning Q&A method and related device

Abstract

An embodiment of the present application provides an online learning Q&A method and related device. The method includes: obtaining the learning posture information of a target learning user in online learning, and obtaining a first learning question input by the target learning user; determining the learning attention information of the target learning user according to the learning posture information; performing question correction processing on the first learning question according to the learning attention information to obtain a second learning question; performing Q&A analysis on the second learning question to obtain Q&A information corresponding to the second learning question; and displaying the Q&A information corresponding to the second learning question. It can automatically correct the learning questions of students according to the learning posture information of the learning user and determine the corresponding Q&A information, thereby reducing the occurrence of ineffective learning questions and improving the efficiency of answering learning questions.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular, to an online learning Q&A method and related devices. Background Art

[0002] With the continuous improvement of educational conditions, the online education method has been increasingly accepted by users, and online education has also developed rapidly. However, in existing online education, usually, teachers give online lectures and students study online. There is also a way that teachers play pre-recorded teaching videos and students watch the teaching videos online for online learning.

[0003] However, when using the above methods for online learning, when students ask learning questions, since teachers or administrators cannot understand the learning status of students, they cannot accurately determine whether the learning questions raised by students are valid learning questions, and some invalid learning questions are likely to appear. For example, learning questions that have been processed in the video, and variant questions of the learning questions, etc. As a result, invalid learning questions will also be answered during the answering process, which makes teachers spend a lot of time answering learning questions, resulting in low efficiency in answering learning questions by teachers. Summary of the Invention

[0004] Embodiments of this application provide a business product determination method and related devices, which can automatically correct students' learning questions according to the learning posture information of learning users, and determine corresponding Q&A information, thereby reducing the occurrence of invalid learning questions and improving the efficiency of answering learning questions.

[0005] In a first aspect of the embodiments of this application, an online learning Q&A method is provided, which is applied to an online learning robot. The method includes:

[0006] Obtain the learning posture information of a target learning user for online learning, and obtain a first learning question input by the target learning user;

[0007] Determine the learning attention information of the target learning user according to the learning posture information;

[0008] Perform question correction processing on the first learning question according to the learning attention information to obtain a second learning question;

[0009] Perform Q&A analysis on the second learning question to obtain Q&A information corresponding to the second learning question;

[0010] Display the Q&A information corresponding to the second learning question.

[0011] In this example, by obtaining the learning posture information of the target learning user in online learning and obtaining the first learning question input by the target learning user, determining the learning attention information of the target learning user according to the learning posture information, performing question correction processing on the first learning question according to the learning attention information to obtain a second learning question, performing question answering analysis on the second learning question to obtain question answering information corresponding to the second learning question, and displaying the question answering information corresponding to the second learning question. Therefore, it is possible to automatically correct the learning questions of students according to the learning posture information of the learning user and determine the corresponding question answering information, thereby reducing the occurrence of ineffective learning questions and improving the efficiency of answering learning questions.

[0012] In a possible implementation manner, the determining the learning attention information of the target learning user according to the learning posture information includes:

[0013] Extracting facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information from the learning posture information;

[0014] Performing attention analysis on the facial expression information to obtain first sub-learning attention information;

[0015] Performing attention analysis on the facial orientation information to obtain second sub-learning attention information;

[0016] Performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information;

[0017] Obtaining a set of attention influence factors corresponding to the learning posture information;

[0018] Determining the learning attention information of the target learning user according to the set of attention influence factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information.

[0019] In a possible implementation manner, the performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information includes:

[0020] Performing line-of-sight focusing analysis on the line-of-sight direction information to obtain a first line-of-sight focusing area;

[0021] Determining m first attention concentration areas in the display area of the to-be-learned video currently being displayed;

[0022] Obtaining first coverage information of the first line-of-sight focusing area on the m first attention concentration areas;

[0023] Extract the display information of each of the m first attention concentration regions to obtain m first display information;

[0024] Determine a core attention concentration region from the m first attention concentration regions according to the m first display information;

[0025] Obtain second coverage information of the first line-of-sight focus region on the core attention concentration region;

[0026] Determine reference learning attention information according to the first coverage information and the second coverage information;

[0027] Perform pose analysis according to the sitting pose information to obtain sitting pose attention correlation information;

[0028] Correct the reference learning attention information according to the sitting pose attention correlation information to obtain the third sub-learning attention information.

[0029] In a possible implementation, the problem correction process of the first learning problem according to the learning attention information to obtain a second learning problem includes:

[0030] Perform semantic analysis on the first learning problem to obtain first semantic information;

[0031] Perform implicit paraphrase analysis on the first learning problem to obtain implicit problem core information;

[0032] Obtain the degree of fit between the implicit problem core information and the first semantic information to obtain a first degree of fit;

[0033] If the first degree of fit is less than a preset degree of fit threshold, determine first learning problem correction information according to the learning attention information and the implicit problem core information;

[0034] Perform correction processing on the first learning problem according to the first learning problem correction information to obtain a second learning problem.

[0035] In a possible implementation, the performing implicit paraphrase analysis on the first learning problem to obtain implicit problem core information includes:

[0036] Perform keyword extraction on the first learning problem to obtain k first problem keywords;

[0037] Obtain the association relationship between the k first problem keywords to obtain a first association relationship graph;

[0038] Extract n target problem keywords from the k first problem keywords according to the first association relationship graph, where n is an integer less than or equal to k;

[0039] Determine the implicit problem logic information according to the n target problem keywords;

[0040] Determine the core information of the implicit problem according to the implicit problem logic information and the n target problem keywords.

[0041] The second aspect of the embodiments of the present application provides an online learning Q&A device, which is applied to an online learning robot. The device includes:

[0042] An acquisition unit, configured to acquire the learning posture information of a target learning user in online learning, and acquire a first learning problem input by the target learning user;

[0043] A determination unit, configured to determine the learning attention information of the target learning user according to the learning posture information;

[0044] A correction unit, configured to perform problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem;

[0045] An analysis unit, configured to perform Q&A analysis on the second learning problem to obtain Q&A information corresponding to the second learning problem;

[0046] A display unit, configured to display the Q&A information corresponding to the second learning problem.

[0047] In a possible implementation manner, the determination unit is configured to:

[0048] Extract the facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information in the learning posture information;

[0049] Perform attention analysis on the facial expression information to obtain first sub-learning attention information;

[0050] Perform attention analysis on the facial orientation information to obtain second sub-learning attention information;

[0051] Perform attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information;

[0052] Obtain a set of attention influence factors corresponding to the learning posture information;

[0053] Determine the learning attention information of the target learning user according to the set of attention influence factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information.

[0054] In a possible implementation manner, in terms of performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain the third sub-learning attention information, the determining unit is configured to:

[0055] Perform line-of-sight focusing analysis on the line-of-sight direction information to obtain a first line-of-sight focusing area;

[0056] Determine m first attention concentration areas in the display area of the to-be-learned video currently being displayed;

[0057] Obtain first coverage information of the first line-of-sight focusing area on the m first attention concentration areas;

[0058] Extract the display information of each of the m first attention concentration areas in the m first attention concentration areas to obtain m first display information;

[0059] Determine a core attention concentration area from the m first attention concentration areas according to the m first display information;

[0060] Obtain second coverage information of the first line-of-sight focusing area on the core attention concentration area;

[0061] Determine reference learning attention information according to the first coverage information and the second coverage information;

[0062] Perform posture analysis according to the sitting posture information to obtain sitting posture attention correlation information;

[0063] Correct the reference learning attention information according to the sitting posture attention correlation information to obtain the third sub-learning attention information.

[0064] In a possible implementation manner, the correcting unit is configured to:

[0065] Perform semantic analysis on the first learning problem to obtain first semantic information;

[0066] Perform implicit paraphrase analysis on the first learning problem to obtain core information of the implicit problem;

[0067] Obtain the degree of fit between the core information of the implicit problem and the first semantic information to obtain a first degree of fit;

[0068] If the first degree of fit is less than a preset degree-of-fit threshold, determine first learning problem correction information according to the learning attention information and the core information of the implicit problem;

[0069] Perform correction processing on the first learning problem according to the first learning problem correction information to obtain a second learning problem.

[0070] In a possible implementation, in terms of performing paraphrasing analysis on the first learning problem to obtain the core information of the implicit problem, the correction unit is configured to:

[0071] Extract keywords from the first learning problem to obtain k first problem keywords;

[0072] Obtain the association relationship between the k first problem keywords to obtain the first association relationship graph;

[0073] According to the first association relationship graph, extract n target problem keywords from the k first problem keywords, where n is an integer less than or equal to k;

[0074] Determine the logical information of the implicit problem according to the n target problem keywords;

[0075] Determine the core information of the implicit problem according to the logical information of the implicit problem and the n target problem keywords.

[0076] A third aspect of the embodiments of the present application provides a terminal, including a processor, an input device, an output device, and a memory, where the processor, the input device, the output device, and the memory are interconnected. Among them, the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the step instructions in the first aspect of the embodiments of the present application.

[0077] A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute some or all of the steps described in the first aspect of the embodiments of the present application.

[0078] A fifth aspect of the embodiments of the present application provides a computer program product, where the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute some or all of the steps described in the first aspect of the embodiments of the present application. The computer program product can be a software installation package. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0080] Figure 1The figure shows a flowchart of an online learning Q&A method provided by an embodiment of the present application;

[0081] Figure 2 The figure shows a schematic structural diagram of a terminal provided by an embodiment of the present application;

[0082] Figure 3 The figure shows a schematic structural diagram of an online learning Q&A device provided by an embodiment of the present application. Detailed implementation manners

[0083] Next, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0084] The terms "first", "second", etc. in the specification and claims of the present application and the above accompanying drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0085] Referring to "embodiment" in the present application means that a specific feature, structure or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described in the present application can be combined with other embodiments.

[0086] To better understand an online learning Q&A method provided by an embodiment of the present application, the scenarios where the online learning Q&A method is applied will be briefly introduced below. When a target learning user conducts online learning, they usually do so through devices such as computers and mobile phones. Among them, the target learning user can be any user who needs to conduct online learning, such as students, teachers in need of training, trainees, etc. The online learning robot can be an application installed on a computer or mobile phone device. The target learning user can input learning questions into a computer, etc. during the Q&A stage of online learning. The online learning robot can receive the learning questions input by the target learning user through hardware devices such as computers, and collect and obtain the learning posture information of the target learning user. The online learning robot can judge the learning attention of the target learning user based on the learning posture information, and use this learning attention to correct the learning questions. Specifically, for example, when the learning question proposed by the target learning user is an invalid learning question, it can be adjusted for this invalid learning question, and a question related to the invalid learning question can be obtained, and the obtained related question can be answered, and the answer information of the answer can be displayed. In addition, when displaying the answer information, the corresponding answer position of the invalid learning question input by the user in online learning can also be displayed, so that the target learning user can more deeply understand and learn the answer information related to the learning question proposed, and can also improve the efficiency and accuracy when answering the learning question.

[0087] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of an online learning Q&A method provided by an embodiment of the present application. As Figure 1 shown, this method is applied to an online learning robot, and this method includes:

[0088] 101. Obtain the learning posture information of the target learning user in online learning, and obtain the first learning question input by the target learning user.

[0089] Among them, the target learning user can be any user who needs to conduct online learning, such as students, teachers in need of training, trainees, etc.

[0090] The learning posture information may include facial expression information, facial orientation information, line-of-sight direction information, sitting posture information, etc. The method for obtaining the learning posture information may be to obtain the online learning image of the target learning user through a camera, and extract the learning posture information from the online learning image. Specifically, for example, multiple consecutive online learning images may be obtained, and the learning posture information may be extracted from each of the multiple consecutive online learning images. Then, clustering processing may be performed on the line-of-sight direction information in the learning posture information to obtain the clustered line-of-sight direction information, etc. For other learning posture information, classification may be performed according to the number of occurrences, and the one with the largest number of occurrences may be used as the corresponding learning posture information. For example, in multiple consecutive images, two different facial expression information appear, and the number of images corresponding to each facial expression information is different. Then, the facial expression information with a higher number of images may be determined as the facial expression information in the learning posture information, and the facial orientation information may also be obtained in the same way. When specifically determining the facial expression, facial orientation information, line-of-sight direction information, and sitting posture information of the target learning user, general image processing methods may be used to obtain the facial expression, facial orientation information, line-of-sight direction information, and sitting posture information of the target learning user, etc.

[0091] The first learning problem of the target learning user may be obtained through the terminal used by the target learning user for online learning. The target learning user may input the first learning problem through the input device of the terminal. The input device may be, for example, a graphics tablet, a keyboard, etc. The first learning problem may be, for example, a problem raised by the target learning user during learning. Specifically, for example, "What is the principle of the xx process?" or "The grammar and usage of the word xx", etc.

[0092] 102. Determine the learning attention information of the target learning user according to the learning posture information.

[0093] The corresponding sub-attention information may be obtained respectively according to the facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information in the learning posture information, and the attention influence factor set corresponding to the learning posture information may be obtained. Finally, the learning attention information may be determined according to the attention influence factor set corresponding to the learning posture information and the sub-attention information.

[0094] The learning attention information can be understood as the attention of the target learning user during online learning. For example, whether the user is concentrating on online learning, the degree of concentration during online learning, etc. The degree of concentration can be understood as the degree of attention and interest of the target learning user in the content shown during online learning. The higher the degree of attention, the higher the degree of concentration; the lower the degree of attention, the lower the degree of concentration; the higher the degree of interest, the higher the degree of concentration; the lower the degree of interest, the lower the degree of concentration. When the degree of interest of the target learning user is low, when the target learning user asks questions, it is easy to ask some questions that have nothing to do with the teaching content, or some questions that have been explained during the online learning process.

[0095] 103. Perform problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem.

[0096] Among them, when performing correction processing on the first learning problem according to the learning attention information, it can be to analyze and process the first learning problem to obtain the corresponding semantic information and the core information of the implicit problem, and when the fit degree between the two is less than the threshold, correct the first learning problem according to the learning attention information and the core information of the implicit problem to obtain the second learning problem.

[0097] 104. Perform question answering analysis on the second learning problem to obtain question answering information corresponding to the second learning problem.

[0098] Among them, for example, it can be to retrieve in the database to obtain the question answering information corresponding to the second learning problem, or to determine the corresponding question answering information according to the question keywords of the second learning problem, etc. Of course, a general method for performing question answering analysis on learning problems can also be used to obtain the question answering information corresponding to the second learning problem.

[0099] 105. Display the question answering information corresponding to the second learning problem.

[0100] The question answering information can be displayed at a specific position on the display screen of the terminal. For example, it can be to display the question answering information in the lower left corner area, etc.

[0101] In this example, by obtaining the learning posture information of the target learning user in online learning and obtaining the first learning question input by the target learning user, determining the learning attention information of the target learning user according to the learning posture information, performing question correction processing on the first learning question according to the learning attention information to obtain a second learning question, performing question answering analysis on the second learning question to obtain question answering information corresponding to the second learning question, and displaying the question answering information corresponding to the second learning question. Therefore, it is possible to automatically correct the learning questions of students according to the learning posture information of the learning user and determine the corresponding question answering information, thereby reducing the occurrence of ineffective learning questions and improving the efficiency of answering learning questions.

[0102] In a possible implementation manner, a method for determining the learning attention information of the target learning user according to the learning posture information includes:

[0103] A1. Extract the facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information from the learning posture information;

[0104] A2. Perform attention analysis on the facial expression information to obtain first sub-learning attention information;

[0105] A3. Perform attention analysis on the facial orientation information to obtain second sub-learning attention information;

[0106] A4. Perform attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information;

[0107] A5. Obtain the set of attention influence factors corresponding to the learning posture information;

[0108] A6. Determine the learning attention information of the target learning user according to the set of attention influence factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information.

[0109] Among them, the learning posture information includes facial expression information, facial orientation information, line-of-sight direction information, sitting posture information, etc. Therefore, the facial expression information, facial orientation information, line-of-sight direction information, sitting posture information, etc. can be extracted from the learning posture information according to the corresponding identification information, etc. of the facial expression information, facial orientation information, line-of-sight direction information, sitting posture information, etc.

[0110] The method of performing attention analysis on facial expression information to obtain the first sub-learning attention information may be: different facial expression information has its corresponding facial expression, and the first sub-learning attention information may be obtained according to the facial expression. Specifically, for example, if the facial expression is a focused expression, the first sub-learning attention information may be highly concentrated attention, which indicates that the target learning user may be seriously learning the online learning content; if the facial expression is a lazy expression, the first sub-learning attention information may be distracted, which indicates that the target learning user may not be seriously learning the online learning content; if the facial expression is a puzzled expression, the first sub-learning attention information may be highly concentrated attention, which indicates that the target learning user is seriously learning and thinking about the online learning content, but has certain doubts or research on the learning content; if the facial expression is an excited expression, the first sub-learning attention information may be moderately concentrated attention, because the facial expression of the target learning user is an excited expression, it indicates that the target learning user is interested in the learning content and may have learned the content. If the facial expression is a sleepy expression, the first sub-learning attention information may be low-concentrated attention, which indicates that the target learning user is not very interested in the learning content, etc.

[0111] Attention analysis can be performed based on the facial orientation information to obtain the second sub-learning attention information, which can be specifically as follows: Here, taking a computer as an example for explanation, it is possible to determine whether there is an intersection between the facial orientation and the display plane of the computer's monitor. Whether there is an intersection between the facial orientation and the display plane can be understood as whether the facial orientation has an overlapping part with the display plane, and the display plane can be understood as the display interface of the monitor. For example, if the facial orientation does not overlap with the display interface (the face and the display interface are perpendicular to each other), then the second sub-learning attention information can be determined to be low-level concentration; if the facial orientation partially overlaps with the display interface (the angle of intersection between the face and the display interface is greater than zero degrees and less than ninety degrees), then the second sub-learning attention information can be determined to be moderate-level concentration; if the facial orientation completely overlaps with the display interface (the face is parallel to the display interface), then the second sub-learning attention information can be determined to be highly-level concentration.

[0112] Attention analysis can also be performed on the line of sight direction information and the sitting posture information to obtain third sub-learning attention information, specifically: focusing area analysis is performed according to the line of sight direction information to obtain the focusing area, and reference learning attention information is determined according to the focusing area and the attention concentration area, and sitting posture attention association information is obtained according to the sitting posture information, and the reference learning attention information is corrected using the sitting posture attention association information to obtain the third sub-learning attention information, so that the third sub-learning attention information can be accurately acquired.

[0113] The set of attention influence factors may include attention influence factors corresponding to the first sub - learning attention information, the second sub - learning attention information, and the third sub - learning attention information respectively, and the attention influence factors can be set by empirical values or historical data.

[0114] The attention influence factors corresponding to the first sub - learning attention information, the second sub - learning attention information, and the third sub - learning attention information respectively can be subjected to a weight operation with the first sub - learning attention information, the second sub - learning attention information, and the third sub - learning attention information, and finally the learning attention information of the target learning user is obtained. When performing the weight operation, the first sub - learning attention information, the second sub - learning attention information, and the third sub - learning attention information can be converted into corresponding attention values, so as to achieve the weight operation.

[0115] In this example, by performing attention analysis on the facial expression information, the first sub - learning attention information is obtained; by performing attention analysis on the facial orientation information, the second sub - learning attention information is obtained; by performing attention analysis on the line - of - sight direction information and the sitting posture information, the third sub - learning attention information is obtained. Therefore, attention information can be obtained in multiple dimensions, and finally, based on the set of attention influence factors, a weight operation is performed to obtain the learning attention information of the target learning user, improving the accuracy when obtaining the learning attention information of the target learning user.

[0116] In a possible implementation manner, a method for performing attention analysis on the line - of - sight direction information and the sitting posture information to obtain the third sub - learning attention information includes:

[0117] B1. Perform line - of - sight focusing analysis on the line - of - sight direction information to obtain the first line - of - sight focusing area;

[0118] B2. Determine m first attention - concentrated areas in the display area of the video to be learned currently being displayed;

[0119] B3. Obtain the first coverage information of the first line - of - sight focusing area on the m first attention - concentrated areas;

[0120] B4. Extract the display information of each of the m first attention - concentrated areas in the m first attention - concentrated areas to obtain m first display information;

[0121] B5. Determine the core attention - concentrated area from the m first attention - concentrated areas according to the m first display information;

[0122] B6. Obtain the second coverage information of the first line - of - sight focusing area on the core attention - concentrated area;

[0123] B7. Determine the reference learning attention information based on the first coverage information and the second coverage information;

[0124] B8. Perform posture analysis based on the sitting posture information to obtain the sitting posture attention correlation information;

[0125] B9. Correct the reference learning attention information according to the sitting posture attention correlation information to obtain the third sub-learning attention information.

[0126] Among them, the line-of-sight direction indicated by the implementation direction information can be obtained, and the line-of-sight intersection between the line where the line-of-sight direction is located and the display interface can be marked. The preset area range where the line-of-sight intersection is located can be determined as the first line-of-sight focus area. The preset range can be a circular area or a rectangular area centered on the line-of-sight intersection. If it is a rectangular area, the length or width of the rectangular area is parallel to the length or width of the display interface. The diameter of the circular area can be set by empirical values or historical data.

[0127] The method for determining m first attention concentration areas can be: extract the display information in the display area of the video to be learned, and extract the learning outline information (or learning catalog information) of the video to be learned. Obtain the outline information corresponding to the display information from the learning outline information, and extract the keywords in the display information. Divide the display area according to the keywords to obtain p key areas; extract the correlation degree between the keywords and the outline information, and determine the m key areas with a correlation degree higher than the preset correlation degree as the first attention concentration areas. The method for extracting the keywords in the display information can adopt a general keyword extraction method to extract the keywords in the display information. A rectangular area where the keyword is located can be determined as the key area. Specifically, the method for determining the line-of-sight focus area described above can be referred to to determine the key area, etc.

[0128] Of course, it can also be to obtain the attention habit information of the target learning user, and obtain m first attention concentration areas according to the attention habit information. Specifically, the attention habit information can include information about the attention concentration areas of the target learning user during learning. For example, when the target learning user is learning, the attention concentration area is likely to be concentrated near the center area of the display interface. Then, the preset circular area including the center area can be determined as the total attention concentration area, and then the total attention concentration area is evenly divided into m first attention concentration areas. Specifically, it can be centered on the center of the total attention concentration area and evenly divide the total attention concentration area to obtain m first attention concentration areas. Of course, m first attention concentration areas can also be obtained through other means.

[0129] m first display information can be extracted from the display image of the video to be learned. Specifically, m first display information can be obtained through general image recognition methods, which will not be elaborated here.

[0130] The first coverage information of the first line-of-sight focus area on the first attention concentration area can be understood as the coverage degree information of the first line-of-sight focus area on the first attention concentration area. Specifically, it can be characterized by the coverage percentage. The larger the coverage percentage, the more the coverage; the smaller the coverage percentage, the less the coverage.

[0131] The core display information can be determined from the m first display information, and the first attention concentration area corresponding to the core display information is determined as the core attention concentration area. The method for determining the core display information from the m first display information can be: the association information between each of the m first display information can be obtained; an association relationship graph can be constructed according to the association information; and the core display information can be determined from the m first display information according to the association relationship graph.

[0132] The association information between the first display information can be understood as whether there is an association between every two first display information. For example, if one first display information is "the core business data of project a is b", and another first display information is "the business data of project a includes b, c, d, etc.", then it can be determined that there is an association between the above two first display information at this time. An association relationship graph can be constructed through the association relationship. There are multiple nodes in the association relationship graph, and each node can represent a first display information. The connection lines between the nodes can represent the existence of an association relationship between the nodes, and the length of the connection lines between the nodes can represent the tightness of the association relationship between the nodes. The shorter the connection line length, the higher the tightness; the longer the connection line length, the lower the tightness. The tightness can be understood as the degree of association between the nodes. The higher the degree of association, the higher the tightness; the lower the degree of association, the lower the tightness. For example, the degree of association of the inclusion relationship is high, and the degree of association of the divergent association relationship is lower than that of the inclusion relationship. The node with the most connection lines in the association relationship graph can be determined as the core node, and the first display information corresponding to the core node is determined as the core display information. When there are multiple nodes with the same and the most connection lines, the node with the shortest total connection line distance is determined as the core node, and the first display information corresponding to the core node is determined as the core display information.

[0133] The method for obtaining the second coverage information of the first line-of-sight focus area on the core attention concentration area can refer to the method for obtaining the first coverage information of the first line-of-sight focus area on the first attention concentration area in the foregoing embodiments, or it can be directly extracting the second coverage information corresponding to the core attention concentration area.

[0134] The global attention information of the target learning user can be determined according to the first coverage information, the core attention information can be determined according to the second coverage information, and finally, the global attention information and the core attention information are subjected to a weight operation to obtain the reference learning attention information. Among them, the method for determining the global attention information of the target learning user based on the first coverage information can be: the larger the coverage rate in the first coverage information, the higher the attention corresponding to the global attention information; the smaller the coverage rate in the first coverage information, the lower the attention corresponding to the global attention information. The method for determining the core attention information according to the second coverage information can be: the larger the second coverage rate, the higher the attention corresponding to the core attention information; the smaller the coverage rate in the second coverage information, the lower the attention corresponding to the core attention information. The weights corresponding to the global attention information and the core attention information can be determined through empirical values or historical data, so as to perform subsequent weight operations to obtain the reference learning attention information.

[0135] The method for performing pose analysis based on the sitting posture information to obtain the sitting posture attention correlation information can be: different sitting postures correspond to different sitting posture attention correlation information. Specifically, it can be: the positive correction strength of the sitting posture attention correlation information corresponding to sitting upright is relatively high, the positive correction strength of the sitting posture attention correlation information corresponding to sitting obliquely is relatively small, the negative correction strength of the sitting posture attention correlation information corresponding to lying on the back is relatively high, and the higher the angle of lying on the back, the higher the negative correction strength; the lower the angle of lying on the back, the lower the negative correction strength.

[0136] The negative correction strength can be understood as reducing the value of the attention parameter in the reference learning attention information during correction. The greater the negative correction strength, the greater the reduced value; the smaller the negative correction strength, the smaller the reduced value.

[0137] Therefore, the sitting posture attention correlation information can be used to correct the reference attention information to obtain the third sub-learning attention information.

[0138] In this example, the reference learning attention information can be jointly determined by the line-of-sight direction to determine the line-of-sight focus area and the attention concentration area, then the sitting posture attention correlation information can be determined according to the sitting posture information, and finally, the sitting posture attention correlation information is used for correction to obtain the third sub-learning attention information, which improves the accuracy in determining the third sub-learning attention information.

[0139] In a possible implementation manner, a method for performing problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem includes:

[0140] C1. Perform semantic analysis on the first learning problem to obtain first semantic information;

[0141] C2. Perform implicit paraphrase analysis on the first learning problem to obtain the core information of the implicit problem;

[0142] C3. Obtain the degree of fit between the core information of the implicit problem and the first semantic information to get the first degree of fit;

[0143] C4. If the first degree of fit is less than the preset degree-of-fit threshold, determine the first learning problem correction information according to the learning attention information and the core information of the implicit problem;

[0144] C5. Perform correction processing on the first learning problem according to the first learning problem correction information to obtain the second learning problem.

[0145] A general semantic analysis method can be used to perform semantic analysis on the first learning problem to obtain the first semantic information. The method for performing implicit paraphrase analysis on the first learning problem to obtain the core information of the implicit problem can be as follows: Multiple keywords in the first learning problem can be extracted, the association relationship between the multiple keywords can be obtained, and the core keyword can be extracted from the multiple keywords according to the association relationship; Extract the semantic conversion information between the core keyword and other keywords in the multiple keywords, and determine the core information of the implicit problem according to the semantic conversion information between the core keyword and other keywords in the multiple keywords. Among them, the method for extracting multiple keywords from the first learning problem can be to use a general keyword extraction method for extraction to obtain multiple keywords. The core keyword can be obtained according to the method for obtaining the core display information in the foregoing embodiments, which will not be elaborated here. After obtaining the core keyword, the semantic conversion information between the core keyword and other keywords in the multiple keywords can be determined according to the corresponding association relationship graph. For example, the semantic conversion information can be determined through the association relationship in the association relationship graph. Specifically, for example, if the association relationship is an inclusion relationship, the semantic conversion information can be to convert the semantics of the core keyword into the semantic information corresponding to the inclusion relationship. If the association relationship is a divergent association relationship, the semantic conversion information can be to convert the semantic information of other keywords into the semantic information of the core keyword. A specific example can be:

[0146] The first problem is "how to avoid extracting the content of the core business data b from the business data 'a, b, c'". The core keyword is "business data includes b, c, d, etc.", the first keyword with an associated relationship is "the core business data is the content of b", and the second keyword is "how to avoid extracting business data". Then the final semantic conversion information can be: The semantic conversion information between the core keyword and the first keyword is: business data is converted to core business data b; the semantic conversion information between the core keyword and the second keyword is: avoiding extracting business data is converted to business data. Therefore, the final core information of the implicit problem is "extracting the core business data b". At this time, the first semantic information is how to avoid extracting the content of the core business b. The fit between the core information of the implicit problem and the first semantic is relatively low. After the fit is lower than the preset fit threshold, the first learning problem correction information can be determined according to the learning attention information and the core information of the implicit problem. Specifically, if the learning attention indicated by the learning attention information at this time is relatively high, it indicates that the target learning user may have some remaining vulnerability problems in the non-normal processing of teaching information (the core business data b cannot be directly extracted). Then the core information of the implicit problem can be avoided, and the first learning problem correction information can be obtained. When performing the avoidance process, the core information of the implicit problem can be hidden. For example, the problem can be hidden and processed as: how to extract general business data using a general method. When correcting the first learning problem with the first learning problem correction information, the first learning problem correction information can be used to replace the first learning problem, thereby obtaining the second learning problem.

[0147] In this example, by performing semantic analysis on the first learning problem to obtain the first semantic information, and performing implicit paraphrase analysis on the first learning problem to obtain the core information of the implicit problem, and finally performing correction processing when the fit between the two is relatively low to obtain the second learning problem, the accuracy of subsequent teaching can be improved.

[0148] In a possible implementation manner, another method for performing implicit paraphrase analysis on the first learning problem to obtain the core information of the implicit problem includes:

[0149] D1. Extract k first problem keywords from the first learning problem;

[0150] D2. Obtain the association relationship between the k first problem keywords to obtain the first association relationship graph;

[0151] D3. According to the first association relationship graph, extract n target problem keywords from the k first problem keywords, where n is an integer less than or equal to k;

[0152] D4. Determine the implicit problem logic information according to the n target problem keywords;

[0153] D5. Determine the core information of the implicit problem according to the implicit problem logic information and the n target problem keywords.

[0154] Among them, the method for keyword extraction of the first learning problem can adopt a general keyword extraction method to obtain k first problem keywords.

[0155] The method for obtaining the association relationship between the k first problem keywords and the method for constructing the first association relationship graph can refer to the method for constructing the association relationship graph in the foregoing embodiments, which will not be elaborated here.

[0156] The first problem keywords with the number of association relationship connection lines exceeding a preset value in the first association relationship graph can be determined as target problem keywords, and the preset value is set through empirical values or historical data.

[0157] The method for determining the implicit problem logic information according to the n target problem keywords can be: the logical relationship between the n target keywords can be obtained, and the implicit problem logic information can be determined according to this logical relationship. Specifically, it can be: the logical relationship between the n target keywords can be understood as the logical transmission relationship between the target keywords, specifically including inclusion relationship, parallel relationship, derivation relationship, etc. Thus, the core information of the implicit problem can be determined according to the implicit problem logic information and the target problem keywords. Specifically, it can be: the target problem keywords are logically combined according to the implicit problem logic information to obtain the core information of the implicit problem.

[0158] In this example, n target problem keywords are determined through the first association relationship graph, and logical combination is performed according to the implicit problem logic information determined by the n target problem keywords, so as to obtain the core information of the implicit problem, improving the accuracy in determining the core information of the implicit problem.

[0159] Consistent with the above embodiments, please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a terminal provided by an embodiment of the present application. As shown in the figure, it includes a processor, an input device, an output device, and a memory. The processor, input device, output device, and memory are interconnected. Among them, the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions. The above program includes instructions for performing the following steps;

[0160] Obtain the learning posture information of the target learning user for online learning, and obtain the first learning problem input by the target learning user;

[0161] Determine the learning attention information of the target learning user according to the learning posture information;

[0162] Perform problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem;

[0163] Perform question answering analysis on the second learning problem to obtain question answering information corresponding to the second learning problem;

[0164] Display the question answering information corresponding to the second learning problem.

[0165] The above mainly introduces the solution of the embodiment of the present application from the perspective of the execution process on the method side. It can be understood that in order for the terminal to implement the above functions, it includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, combining the units and algorithm steps of each example described in the embodiments provided in this article, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0166] The embodiment of the present application can divide the functions of the terminal according to the above method examples. For example, each function unit can be divided corresponding to each function, or two or more functions can be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or in the form of a software function unit. It should be noted that the division of units in the embodiment of the present application is illustrative, only a logical function division, and there may be other division methods in actual implementation.

[0167] Consistent with the above, please refer to Figure 3 , Figure 3 which is a schematic structural diagram of an online learning question answering device provided by an embodiment of the present application. As Figure 3 shown, applied to an online learning robot, the device includes:

[0168] An acquisition unit 301, configured to acquire the learning posture information of a target learning user for online learning, and acquire a first learning problem input by the target learning user;

[0169] A determination unit 302, configured to determine the learning attention information of the target learning user according to the learning posture information;

[0170] A correction unit 303, configured to perform problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem;

[0171] An analysis unit 304, configured to perform answer analysis on the second learning problem to obtain answer information corresponding to the second learning problem;

[0172] A display unit 305, configured to display the answer information corresponding to the second learning problem.

[0173] In a possible implementation manner, the determining unit 302 is configured to:

[0174] Extract facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information from the learning posture information;

[0175] Perform attention analysis on the facial expression information to obtain first sub-learning attention information;

[0176] Perform attention analysis on the facial orientation information to obtain second sub-learning attention information;

[0177] Perform attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information;

[0178] Obtain a set of attention influence factors corresponding to the learning posture information;

[0179] Determine the learning attention information of the target learning user according to the set of attention influence factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information.

[0180] In a possible implementation manner, in terms of performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information, the determining unit 302 is configured to:

[0181] Perform line-of-sight focus analysis on the line-of-sight direction information to obtain a first line-of-sight focus area;

[0182] Determine m first attention concentration areas in the display area of the to-be-learned video currently being displayed;

[0183] Obtain first coverage information of the first line-of-sight focus area on the m first attention concentration areas;

[0184] Extract the display information of each of the m first attention concentration areas in the m first attention concentration areas to obtain m first display information;

[0185] Determine a core attention concentration area from the m first attention concentration areas according to the m first display information.

[0186] Obtain the second coverage information of the first line-of-sight focus area on the core attention concentration area;

[0187] Determine the reference learning attention information according to the first coverage information and the second coverage information;

[0188] Perform posture analysis according to the sitting posture information to obtain sitting posture attention correlation information;

[0189] Correct the reference learning attention information according to the sitting posture attention correlation information to obtain the third sub-learning attention information.

[0190] In a possible implementation manner, the correction unit 303 is configured to:

[0191] Perform semantic analysis on the first learning problem to obtain first semantic information;

[0192] Perform implicit paraphrase analysis on the first learning problem to obtain implicit problem core information;

[0193] Obtain the degree of fit between the implicit problem core information and the first semantic information to obtain a first degree of fit;

[0194] If the first degree of fit is less than a preset degree of fit threshold, determine first learning problem correction information according to the learning attention information and the implicit problem core information;

[0195] Perform correction processing on the first learning problem according to the first learning problem correction information to obtain a second learning problem.

[0196] In a possible implementation manner, in terms of performing implicit paraphrase analysis on the first learning problem to obtain implicit problem core information, the correction unit 303 is configured to:

[0197] Extract keywords from the first learning problem to obtain k first problem keywords;

[0198] Obtain the association relationship between the k first problem keywords to obtain a first association relationship graph;

[0199] According to the first association relationship graph, extract n target problem keywords from the k first problem keywords, where n is an integer less than or equal to k;

[0200] Determine implicit problem logic information according to the n target problem keywords;

[0201] Determine the implicit problem core information according to the implicit problem logic information and the n target problem keywords..

[0202] An embodiment of the present application also provides a computer storage medium. The computer storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute some or all of the steps of any one of the online learning and Q&A methods described in the foregoing method embodiments.

[0203] An embodiment of the present application also provides a computer program product. The computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program enables a computer to execute some or all of the steps of any one of the online learning and Q&A methods described in the foregoing method embodiments.

[0204] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0205] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0206] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0207] 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 can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0208] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software program modules.

[0209] When the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned memory includes various media that can store program codes, such as USB flash drives, read-only memories (ROM), random access memories (RAM), mobile hard disks, magnetic disks, or optical discs.

[0210] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. This program can be stored in a computer-readable memory, and the memory can include: flash drives, read-only memories, random access memories, magnetic disks, or optical discs, etc.

[0211] The above has introduced the embodiments of this application in detail. Specific examples are used in this article to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. An online learning Q&A method, characterized in that, Applied to an online learning robot, the method includes: Obtaining the learning posture information of a target learning user for online learning, and obtaining a first learning question input by the target learning user; Determining the learning attention information of the target learning user according to the learning posture information; Performing question correction processing on the first learning question according to the learning attention information to obtain a second learning question; Performing answer analysis on the second learning question to obtain answer information corresponding to the second learning question; Displaying the answer information corresponding to the second learning question; The determining the learning attention information of the target learning user according to the learning posture information includes: Extracting facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information from the learning posture information; Performing attention analysis on the facial expression information to obtain first sub-learning attention information; Performing attention analysis on the facial orientation information to obtain second sub-learning attention information; Performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information; Obtaining a set of attention influencing factors corresponding to the learning posture information; Determining the learning attention information of the target learning user according to the set of attention influencing factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information; The performing question correction processing on the first learning question according to the learning attention information to obtain a second learning question includes: Performing semantic analysis on the first learning question to obtain first semantic information; Performing implicit paraphrase analysis on the first learning question to obtain core information of the implicit question; Obtaining the degree of fit between the core information of the implicit question and the first semantic information to obtain a first degree of fit; If the first degree of fit is less than a preset degree-of-fit threshold, determining first learning question correction information according to the learning attention information and the core information of the implicit question; Performing correction processing on the first learning question according to the first learning question correction information to obtain a second learning question.

2. The online learning Q&A method according to claim 1, wherein The performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information includes: Performing line-of-sight focus analysis on the line-of-sight direction information to obtain a first line-of-sight focus area; Determining m first attention concentration areas in the display area of the to-be-learned video currently being displayed; Obtaining first coverage information of the first line-of-sight focus area on the m first attention concentration areas; Extracting the display information of each of the m first attention concentration areas in the m first attention concentration areas to obtain m first display information; Determining a core attention concentration area from the m first attention concentration areas according to the m first display information; Obtaining second coverage information of the first line-of-sight focus area on the core attention concentration area; Determining reference learning attention information according to the first coverage information and the second coverage information; Performing posture analysis according to the sitting posture information to obtain sitting posture attention correlation information; Correct the reference learning attention information according to the sitting posture attention correlation information to obtain the third sub-learning attention information.

3. The online learning Q&A method according to claim 2, characterized in that, Performing implicit paraphrase analysis on the first learning problem to obtain core information of the implicit problem, including: Extracting keywords from the first learning problem to obtain k first problem keywords; Obtaining the association relationship between the k first problem keywords to obtain a first association relationship graph; Extracting n target problem keywords from the k first problem keywords according to the first association relationship graph, where n is an integer less than or equal to k; Determining implicit problem logic information according to the n target problem keywords; Determining the core information of the implicit problem according to the implicit problem logic information and the n target problem keywords.

4. An online learning Q&A device, characterized in that, Applied to an online learning robot, the device includes: An acquisition unit for acquiring the learning posture information of the target learning user in online learning and acquiring the first learning problem input by the target learning user; A determination unit for determining the learning attention information of the target learning user according to the learning posture information; A correction unit for performing problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem; An analysis unit for performing question answering analysis on the second learning problem to obtain question answering information corresponding to the second learning problem; A display unit for displaying the question answering information corresponding to the second learning problem; The determination unit is used for: Extracting facial expression information, facial orientation information, line-of-sight direction information, and sitting posture information from the learning posture information; Performing attention analysis on the facial expression information to obtain first sub-learning attention information; Performing attention analysis on the facial orientation information to obtain second sub-learning attention information; Performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information; Obtaining a set of attention influence factors corresponding to the learning posture information; Determining the learning attention information of the target learning user according to the set of attention influence factors, the first sub-learning attention information, the second sub-learning attention information, and the third sub-learning attention information; In the aspect of performing problem correction processing on the first learning problem according to the learning attention information to obtain a second learning problem, the determination unit is used for: Performing semantic analysis on the first learning problem to obtain first semantic information; Performing implicit paraphrase analysis on the first learning problem to obtain core information of the implicit problem; Obtaining the degree of fit between the core information of the implicit problem and the first semantic information to obtain a first degree of fit; If the first degree of fit is less than a preset degree-of-fit threshold, determining first learning problem correction information according to the learning attention information and the core information of the implicit problem; Performing correction processing on the first learning problem according to the first learning problem correction information to obtain a second learning problem.

5. The online learning Q&A device according to claim 4, characterized in that In the aspect of performing attention analysis on the line-of-sight direction information and the sitting posture information to obtain third sub-learning attention information, the determination unit is used for: Perform line-of-sight focusing analysis on the line-of-sight direction information to obtain a first line-of-sight focusing area; Determine m first attention concentration areas in the display area of the to-be-learned video currently being displayed; Obtain first coverage information of the first line-of-sight focusing area on the m first attention concentration areas; Extract the display information of each of the m first attention concentration areas in the m first attention concentration areas to obtain m first display information; Determine a core attention concentration area from the m first attention concentration areas according to the m first display information; Obtain second coverage information of the first line-of-sight focusing area on the core attention concentration area; Determine reference learning attention information according to the first coverage information and the second coverage information; Perform posture analysis according to the sitting posture information to obtain sitting posture attention correlation information; Correct the reference learning attention information according to the sitting posture attention correlation information to obtain the third sub-learning attention information.

6. A terminal, characterized in that, Comprising a processor, an input device, an output device and a memory, the processor, the input device, the output device and the memory are interconnected, wherein the memory is used for storing a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method according to any one of claims 1-3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program includes program instructions, and the program instructions, when executed by a processor, cause the processor to execute the method according to any one of claims 1-3.

Images