Method and device for correcting phonetic notation of point-added character, electronic equipment and storage medium

Abstract

The invention relates to the technical field of computers, and discloses a correction method and device for adding character phonetic notation, electronic equipment and a storage medium, and the method provided by the invention comprises the following steps: obtaining a to-be-corrected picture; identifying the to-be-corrected picture to obtain each small question in the Chinese character added phonetic notation question, the added Chinese character of each small question and the answer result of the added Chinese character; identifying the added character of each small question and the answering result of the added character to obtain a target added character and a target answering phonetic notation; determining a target phonetic notation corresponding to the target point-added character based on a corresponding relationship between the character and the phonetic notation; and based on a comparison result of the target phonetic notation and the target answering phonetic notation, determining a correction result of the Chinese character adding phonetic notation question. According to the method, rules do not need to be used for question cutting, the operation efficiency is high, question positioning is more accurate, a question bank does not need to be prepared in advance, and the efficiency of adding character phonetic notation correction can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, electronic equipment and storage medium for correcting the pronunciation of underlined characters. Background Art

[0002] Pronunciation for underlined characters is a common question type. Specifically, the question presents a word with a character underlined, followed by two pronunciations. The question requires students to choose the correct pronunciation of the underlined character or give the correct pronunciation of the underlined character. Because these questions appear frequently and in large numbers, they are a burdensome task for grading. Therefore, a method for automatically grading these questions is needed. Summary of the Invention

[0003] In view of this, the present invention provides a method, device, electronic device and storage medium for correcting underlined characters and phonetic notation, so as to solve the problem of correcting underlined characters and phonetic notation.

[0004] In a first aspect, the present invention provides a method for correcting underlined characters and phonetic notation, the method comprising:

[0005] Get the pictures to be corrected;

[0006] Recognize the picture to be corrected, and obtain each sub-question under the underlined word phonetic notation question, the underlined word for each sub-question, and the answer result for the underlined word;

[0007] Recognize the underlined characters and the answer results of each question to obtain the target underlined characters and the target answer phonetic notation;

[0008] Determining the target pronunciation corresponding to the target underlined character based on the correspondence between the character and the pronunciation;

[0009] Based on the comparison result of the target phonetic notation and the target answer phonetic notation, the correction result of the underlined character phonetic notation question is determined.

[0010] The method for correcting underlined characters and phonetic notation provided by an embodiment of the present invention directly displays each sub-question under the underlined characters and phonetic notation question in the picture to be corrected, the underlined characters for each sub-question, and the recognition results of the underlined characters, so as to directly identify each sub-question without using rules to identify the question, resulting in high operating efficiency and more accurate question positioning. At the same time, after identifying the target underlined characters for each sub-question, the correspondence between the characters and phonetic notation is used to obtain the target phonetic notation corresponding to the target underlined characters, and then the correction result is obtained by comparing the target phonetic notation with the target answer phonetic notation. Since the correspondence between the characters and phonetic notation can be applied to all questions with underlined characters and phonetic notation, there is no need to prepare a question bank in advance, which can improve the efficiency of correcting underlined characters and phonetic notation.

[0011] In an optional embodiment, the answer result of the underlined characters includes the probability of the answer result, and the answer result of the underlined characters for each question is identified to obtain the target answer phonetic notation, including:

[0012] Get the preset answer probability;

[0013] If the probability of the answer result is less than the preset answer probability, the answer result of the underlined word is determined to be empty, and the correction result of the small question corresponding to the answer result is wrong.

[0014] The method for grading underlined characters and phonetic notation provided by the present invention first screens the answer results using a preset probability of answering. If the probability of an answer result is less than the preset probability, it indicates that the answer was not answered or that the answer does not meet the requirements of the question. Therefore, the corresponding question is directly graded as an error, eliminating the need for subsequent recognition, thereby improving the probability of correcting the assignment.

[0015] In an optional embodiment, the answer result of the underlined characters includes the position of the answer result, and the identifying the answer result of the underlined characters for each question to obtain the target answer phonetic notation further includes:

[0016] If the probability of the answer result is greater than or equal to the preset answer probability, obtaining a picture of the answer result based on the position of the answer result;

[0017] Performing text line recognition on the picture of the answer result to obtain a text line recognition result;

[0018] The text line recognition result is screened to determine the target answering phonetic notation, where the target answering phonetic notation includes a first preset number of answering phonetic notations.

[0019] In the method for correcting the phonetic notation of underlined characters provided in an embodiment of the present invention, when the probability of an answer result is greater than a preset answer probability, an image of the answer result is captured, text line recognition is performed on it, and the text line recognition results are obtained. The text line recognition results are then filtered to obtain a first preset number of phonetic notations of the answers. By retaining the first preset number of phonetic notations of the answers for subsequent comparison, the risk of misrecognition is reduced.

[0020] In an optional embodiment, the text line recognition result includes a probability that a character at each position in the answer result belongs to a preset character, and screening the text line recognition result to determine the target answer phonetic notation includes:

[0021] For each position, the maximum probability that the query character belongs to the preset characters;

[0022] Obtaining a first answer result based on the preset character corresponding to the maximum probability of each position;

[0023] For each position, querying the second largest probability that the character belongs to the preset character and calculating the difference between the maximum probability and the second largest probability;

[0024] Obtaining a target position corresponding to a minimum difference value based on the difference value of each position;

[0025] The character at the target position in the first answer result is replaced with a preset character corresponding to the second most probable character at the target position to obtain a second answer result, wherein the target answer phonetic notation includes the first answer result and the second answer result.

[0026] In the method for correcting underlined characters with phonetic notation provided by the present invention, a close probability difference indicates the presence of similar characters; a larger probability difference indicates a more certain recognition result. Based on this, a second answer is obtained by comparing the probability differences, thereby avoiding missed or misidentified characters due to similar characters and improving the recognition accuracy of the answer.

[0027] In an optional embodiment, the identifying the underlined characters for each question to obtain the target underlined characters includes:

[0028] Obtaining an underlined word image from the corresponding question based on the position of the underlined word;

[0029] Character classification and recognition are performed on the underlined character picture to obtain a second preset number of underlined characters, and the target underlined characters include the second preset number of underlined characters.

[0030] The method for correcting the pronunciation of underlined characters provided by the embodiments of the present invention employs a character classification recognition method to identify images of underlined characters, thereby improving the accuracy of the resulting underlined characters. Furthermore, the second preset number of underlined characters obtained after character classification recognition can provide a certain degree of recognition redundancy for the recognition results, thereby ensuring the reliability of subsequent correction results.

[0031] In an optional embodiment, if the target underlined characters include a first preset number of underlined characters, determining the target pronunciation corresponding to the target underlined characters based on the correspondence between characters and pronunciation includes:

[0032] For each of the underlined characters, the correspondence between the character and the pronunciation is queried to obtain at least one pronunciation corresponding to the underlined character. If the underlined character is a polyphonic character, the underlined character corresponds to at least two of the pronunciations.

[0033] The method for correcting the pronunciation of underlined characters provided by the embodiment of the present invention obtains at least two pronunciations for underlined characters of polyphonetic characters, both of which are retained for subsequent comparison processing, thereby further improving the reliability of the correction results.

[0034] In an optional embodiment, determining the correction result of the underlined character phonetic notation question based on the comparison result of the target phonetic notation and the target answer phonetic notation includes:

[0035] If any of the target phonetic notations matches any of the target answer phonetic notations, the correction result of the question with matching phonetic notations is determined to be correct.

[0036] The method for correcting the phonetic notation of underlined characters provided in an embodiment of the present invention matches any phonetic notation with any phonetic notation of an answer due to errors in character recognition or text recognition. If any one of them matches, it indicates that the correction result is that the answer is correct, which can effectively reduce the risk of errors in recognizing similar characters.

[0037] In a second aspect, the present invention provides a device for correcting underlined characters and phonetic notation, the device comprising:

[0038] Image acquisition module, used to obtain images to be corrected;

[0039] An image recognition module is used to recognize the image to be corrected, and obtain each sub-question under the underlined character phonetic notation question, the underlined character for each sub-question, and the answer result for the underlined character;

[0040] An answer recognition module is used to recognize the underlined characters of each question and the answer results of the underlined characters to obtain the target underlined characters and the target answer phonetic notation;

[0041] A phonetic notation determination module, configured to determine a target phonetic notation corresponding to the target underlined character based on a correspondence between the character and the phonetic notation;

[0042] The correction result determination module is used to determine the correction result of the underlined character phonetic notation question based on the comparison result of the target phonetic notation and the target answer phonetic notation.

[0043] In a third aspect, the present invention provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to execute the method for correcting the underlined characters and phonetic notation of the above-mentioned first aspect or any corresponding embodiment thereof.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for correcting the underlined characters and phonetic notation according to the first aspect or any corresponding embodiment thereof.

[0045] The device for correcting underlined characters with phonetic notation, the electronic device and the computer-readable storage medium provided in the embodiments of the present invention correspond to the above-mentioned method for correcting underlined characters with phonetic notation. Based on this, for the beneficial effects of the device for correcting underlined characters with phonetic notation, the electronic device and the computer-readable storage medium, please refer to the description of the corresponding beneficial effects of the method for correcting underlined characters with phonetic notation above, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 1 is a flow chart of a method for correcting the pronunciation of underlined characters according to an embodiment of the present invention;

[0048] Figure 2 is a flow chart of another method for correcting the pronunciation of underlined characters according to an embodiment of the present invention;

[0049] Figure 3 is a schematic diagram of text line recognition result screening according to an embodiment of the present invention;

[0050] Figure 4 1 is a structural block diagram of a device for correcting underlined Chinese characters with phonetic notation according to an embodiment of the present invention;

[0051] Figure 5 FIG. 4 is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0053] In related technologies, the method for correcting the pronunciation of underlined characters is to detect the text lines of the entire document, use rules to determine the title description and content of each question, then use keywords to determine whether the question is a question type that requires selecting the correct pronunciation of underlined characters, detect all checked positions in the question, use rules to determine the checked pinyin, search for the answer to the question from a pre-prepared question bank, and use the rules to determine whether the selected pinyin is correct. However, this solution requires identifying the text lines of the entire document, resulting in low operating efficiency; and using rules for question segmentation and question type judgment is prone to misjudgment due to the limited scope of the rules. At the same time, this solution requires the pre-preparation of a question bank, resulting in high costs.

[0054] Based on this, an embodiment of the present invention provides a method for correcting the pronunciation of underlined characters. By identifying each question, the underlined characters in each question, and the answer, each of these is recognized separately to determine the pinyin of the underlined characters in each question and the recognition result of the answer. The two are then compared to determine the correction result for each question. This method does not require the use of rules to segment the questions, but directly locates each question and the characters in each question. It does not require the recognition of text lines on the entire page, and its operation efficiency is relatively high.

[0055] According to an embodiment of the present invention, an embodiment of a method for correcting underlined characters with phonetic notation is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0056] In this embodiment, a method for correcting underlined characters with phonetic notation is provided, which can be used in electronic devices such as mobile phones, tablet computers, etc. Figure 1 Flowchart of the method for correcting the pronunciation of underlined characters according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0057] Step S101: Obtain the picture to be corrected.

[0058] The image to be corrected can be acquired by an image acquisition device of the electronic device, or can be acquired by a third-party device and sent to the electronic device. Of course, other methods can also be used to obtain it, and there is no limitation on them here. For example, the answer sheet can be scanned by a scanning device to obtain the image to be corrected corresponding to the answer sheet.

[0059] The images to be corrected include questions with underlined characters and phonetic notation. These questions can include choosing the correct pronunciation of the underlined characters, deleting incorrect pronunciations of the underlined characters, or writing out the pinyin of the underlined characters. For example, a question may include an underlined character and two pinyin characters. The question requires choosing the correct pronunciation from the two pinyin characters by circling or ticking the box. Alternatively, the question requires deleting the incorrect pronunciation from the two pinyin characters by drawing a slash.

[0060] Step S102: Recognize the image to be corrected, and obtain each sub-question under the question with underlined characters and phonetic notation, the underlined characters for each sub-question, and the answer result for the underlined characters.

[0061] The recognition of images for correction is based on an image detection model. The model takes an image as input and outputs each sub-question within the "underlined characters" phonetic notation question, the underlined characters for each sub-question, and the answer to the underlined characters. Specifically, the output includes the coordinates and probability of the question box for each sub-question, the coordinates of the underlined characters, and the coordinates and probability of the answer box for the answer.

[0062] The image detection model is trained using sample images. The labels for the sample images include the coordinates of each sub-question in the underlined character phonetic notation test, the coordinates of the underlined characters in each sub-question, and the answer to the underlined characters. The sample images and their labels are used to train the initial image detection model, iteratively updating its parameters. If the iterative update cutoff condition is met, training is complete, the model parameters are fixed, and the image detection model is obtained.

[0063] Step S103: Identify the underlined characters and the answer results of each question to obtain the target underlined characters and the target answer phonetic notation.

[0064] The processing in step S102 above obtains the position information of the underlined characters and the answer results for each question. Furthermore, based on this position information, it is necessary to identify the specific content of the underlined characters and the pinyin corresponding to the answer results. The identification of the underlined characters for each question can be based on a text line recognition model or a character recognition model. The input of the text line recognition model or the character recognition model is an image of the underlined characters, and the output is the characters in the image, i.e., the target underlined characters.

[0065] The recognition of answers for underlined characters can be performed based on a text line recognition model. Because answers for underlined characters are represented in phonetic notation, which includes multiple characters in a sequential order, using a text line recognition model to recognize answers for underlined characters can yield a highly accurate phonetic representation of the target answer.

[0066] It should be noted that due to the error in the recognition results, the target highlighted characters and target answer phonetic notation can be the top N most likely highlighted characters and answer phonetic notation to avoid the impact of recognition error on the correction results. Specifically, the target highlighted characters include multiple highlighted characters, and the target answer phonetic notation includes multiple answer phonetic notation.

[0067] Step S104: determining the target pronunciation corresponding to the target underlined character based on the correspondence between the characters and the pronunciation.

[0068] Each character has a corresponding pronunciation. Of course, due to the existence of polyphonetic characters, some characters may have two or more pronunciations. Therefore, the relationship between a character and pronunciation can be a one-to-many relationship. After the processing of the above step S103, the target dotted character is obtained. Based on the corresponding relationship between the target dotted character query character and the pronunciation, the target pronunciation corresponding to the target dotted character is obtained. Since there may be multiple dotted characters in the target dotted character, and these multiple dotted characters may contain polyphonetic characters, accordingly, each dotted character corresponds to at least one target pronunciation.

[0069] Step S105: determining the correction result of the underlined character phonetic notation question based on the comparison result of the target phonetic notation and the target answer phonetic notation.

[0070] Since the target pronunciation corresponding to the target underlined character is determined and the target answer pronunciation is obtained in step S103 above, the two are compared to determine the correction result of the underlined character pronunciation question. If the question requires selecting the correct pronunciation, then the target pronunciation corresponding to the target underlined character and the target answer pronunciation are consistent, indicating that the correction result is correct. If the question requires deleting the incorrect pronunciation, then the target pronunciation corresponding to the target underlined character and the target answer pronunciation are inconsistent, indicating that the correction result is correct.

[0071] The method for correcting underlined characters and phonetic notation provided by an embodiment of the present invention directly displays each sub-question under the underlined characters and phonetic notation question in the picture to be corrected, the underlined characters for each sub-question, and the recognition results of the underlined characters, so as to directly identify each sub-question without using rules to identify the question, resulting in high operating efficiency and more accurate question positioning. At the same time, after identifying the target underlined characters for each sub-question, the correspondence between the characters and phonetic notation is used to obtain the target phonetic notation corresponding to the target underlined characters, and then the correction result is obtained by comparing the target phonetic notation with the target answer phonetic notation. Since the correspondence between the characters and phonetic notation can be applied to all questions with underlined characters and phonetic notation, there is no need to prepare a question bank in advance, which can improve the efficiency of correcting underlined characters and phonetic notation.

[0072] In this embodiment, a method for correcting underlined characters with phonetic notation is provided, which can be used in electronic devices such as mobile phones, tablet computers, etc. Figure 2 Flowchart of the method for correcting the pronunciation of underlined characters according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0073] Step S201: Get the image to be corrected. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0074] Step S202: Recognize the image to be corrected, and obtain each sub-question under the question with underlined characters and phonetic notation, the underlined characters for each sub-question, and the answer result for the underlined characters.

[0075] The image to be corrected is recognized. The recognition results include the coordinates and probability of each sub-question under the underlined phonetic notation question, the coordinates of the underlined characters for each sub-question, and the coordinates and probability of the answer to the underlined characters. The probability of each sub-question is used to screen each sub-question. For example, a sub-question recognition probability is set and the probability of each sub-question is compared with the sub-question recognition probability. If the probability is greater than the sub-question recognition probability, the sub-question is retained; if it is less than or equal to the sub-question recognition probability, the sub-question is misidentified and is screened out.

[0076] The answers in bold include the probability of the answer, which is used to filter the answers to identify those that were not answered or did not follow the instructions. The probability of an unanswered answer is lower than the preset probability. Similarly, if the answer was not followed, the probability of the answer was also lower than the preset probability. If the question requires a check mark to select the correct answer, and the respondent selects the correct answer by circling it, this answer is considered to be a failure to follow the instructions.

[0077] Step S203: Identify the underlined characters and the answer results of each question to obtain the target underlined characters and the target answer phonetic notation.

[0078] Specifically, the above step S203 includes:

[0079] Step S2031: Obtain an image of the underlined characters from the corresponding question based on the position of the underlined characters.

[0080] The position of the highlighted characters is used to locate the highlighted characters in the image to be corrected, and then the image with the highlighted characters is cut out from the image to be corrected to obtain the highlighted character image. The position of the highlighted characters is generally represented by a detection box, specifically by giving the coordinates of the upper left corner of the detection box and the width and height of the detection box; it can also be represented by giving the coordinates of the upper left corner and the lower right corner of the detection box.

[0081] Step S2032: performing character classification and recognition on the underlined character image to obtain a second preset number of underlined characters.

[0082] The target highlighted characters include a second preset number of highlighted characters.

[0083] Character classification and recognition are obtained using a character classification and recognition model. The input of the character classification and recognition model is an image, and the output is the underlined characters in the image. When retaining the output results of the character classification and recognition model, the top N underlined characters with the highest probability are retained. N is the second preset number, which is set based on actual needs, including but not limited to 2 or 3.

[0084] For each inscribed character image, the first N inscribed characters output by the character classification and recognition model are retained to obtain a second preset number of inscribed characters. That is, for each inscribed character image, the second preset number of inscribed characters is obtained.

[0085] Step S2033: Obtain a preset answer probability.

[0086] The preset answer probability is used to filter out the answer results that are not answered or not answered as required. The specific value is set according to actual needs.

[0087] Step S2034: If the probability of the answer result is less than the preset answer probability, the answer result of the underlined word is determined to be empty.

[0088] Among them, the correction result of the question corresponding to the answer result is wrong.

[0089] If the probability of the answer result is less than the preset answer probability, it means that the answer to the question is invalid, that is, no answer is given or the answer is not answered according to the requirements, and the answer result with the underlined words is determined to be empty. Accordingly, the correction result of the question corresponding to the answer result is wrong.

[0090] It should be noted that step S2033 and step S2034 can be executed before step S2031, that is, the answer results that are not answered or not answered as required are first screened according to the preset answer probability. If the probability of the answer result is less than the preset answer probability, the above steps S2031-S2032 do not need to be executed for this question.

[0091] Step S2035: If the probability of the answer result is greater than or equal to the preset answer probability, a picture of the answer result is obtained based on the position of the answer result.

[0092] If the probability of the answer result is greater than or equal to the preset answer probability, it means that the current answer is a valid answer and needs further analysis. In the above step S202, the position of each answer result is obtained, and the image to be corrected is captured using the position to obtain the image of the answer result.

[0093] Step S2036: Perform text line recognition on the picture of the answer result to obtain a text line recognition result.

[0094] The image of the answer result is input into the text line recognition model to obtain the text line recognition result. The text line recognition result includes the probability that each character in the answer result belongs to the preset characters. For example, the preset characters include all characters in the pinyin table and the tones of the finals. When the text line recognition result is output, the probability of each character belonging to the preset characters is output.

[0095] Step S2037: Filter the text line recognition results to determine the target answer phonetic notation.

[0096] The target answering phonetic notation includes a first preset number of answering phonetic notations.

[0097] Since the text line recognition result includes the probability that a character belongs to each preset character, then during screening, only the top M preset characters with the highest probability can be retained for each character, or the text line recognition result can be screened in other ways, which are not limited here. After screening the text line recognition results, the target answer phonetic notation is obtained. Since multiple answer phonetic notations are included, the target answer phonetic notation includes a first preset number of answer phonetic notations. For example, for each answer result, 3 answer phonetic notations are included, and these 3 answer phonetic notations constitute the target answer phonetic notation for the answer result.

[0098] In some optional implementations, the text line recognition result includes the probability that the character at each position in the answer result belongs to the preset character. Based on this, the above step S2037 includes:

[0099] Step a1: For each position, the maximum probability that the query character belongs to the preset characters.

[0100] Step a2: Obtain a first answer result based on the preset character corresponding to the maximum probability of each position.

[0101] Step a3: for each position, the second highest probability that the query character belongs to the preset characters is determined and the difference between the maximum probability and the second highest probability is calculated.

[0102] Step a4: Obtain the target position corresponding to the minimum difference based on the difference of each position.

[0103] Step a5: Replace the character at the target position in the first answer result with a preset character corresponding to the second most probable target position to obtain a second answer result.

[0104] The target answer phonetic notation includes the first answer result and the second answer result.

[0105] The text line recognition results can be used Figure 3The first column of the table represents the preset characters. Starting from the second column, each column represents the probability that the character at that position belongs to each preset character. Taking the second column as an example, the second column represents the probability that the first character in the answer belongs to each preset character.

[0106] Each position represents the position in the answer result according to the order from front to back, corresponding to Figure 3 Each column starting from the second column. Starting from the second column, query the preset character corresponding to the maximum probability in each column in turn, and combine the preset characters found according to the order of position to obtain the first answer result. For example, the preset character corresponding to the maximum probability in the second column is y, and the preset character corresponding to the maximum probability in the third column is ā, then the first answer result is yā. Starting from the second column again, query the preset character corresponding to the second highest probability in each column, and calculate the difference between the maximum probability and the second highest probability, and then compare the sizes of the differences to determine the column where the minimum difference is located. For example, Figure 3 The difference between the second highest probability and the highest probability in column 2 is 0.9, and the difference between the second highest probability and the highest probability in column 3 is 0.3. Therefore, the column with the smallest difference is column 3. Based on this, the characters in the corresponding positions in the first answer are replaced with the preset characters corresponding to the second highest probability in column 3 to obtain the second answer. That is, replace ā with á, and the second answer is yā.

[0107] By analogy, the difference between the maximum probability and the third-highest probability can be calculated, and the position corresponding to the minimum difference can be obtained by comparison. Then, the characters of the first answer result can be replaced using the above method to obtain the third answer result. It should be noted that the number of answer results finally obtained is set according to actual needs and is not limited here.

[0108] Since the probability difference is close, it indicates that similar characters may exist; and the larger the probability difference, the more certain the recognition result. Based on this, the second answer is obtained by comparing the probability differences, which avoids missed recognition or misrecognition caused by similar characters, thereby improving the recognition accuracy of the answer.

[0109] Step S204: determining the target pronunciation corresponding to the target underlined character based on the correspondence between the character and the pronunciation.

[0110] Please refer to the above text for the correspondence between characters and phonetic symbols. Figure 1 The detailed description of step S104 in the illustrated embodiment is omitted here.

[0111] In some embodiments, the above step S204 includes: for each underlined character, querying the correspondence between the character and the pronunciation, obtaining at least one pronunciation corresponding to the underlined character; if the underlined character is a polyphonetic character, the underlined character corresponds to at least two pronunciations.

[0112] The correspondence between characters and phonetic notation is based on the character itself. Due to the existence of polyphonic characters, the number of phonetic notations corresponding to a character may be one, two, or more. That is, for polyphonic characters, the dotted characters correspond to at least two phonetic notations.

[0113] For each underlined character, at least one phonetic notation corresponding to each underlined character is obtained by searching for the correspondence between the character and the phonetic notation.

[0114] For polyphonic characters, at least two pronunciations are obtained, both of which are retained for subsequent comparison processing, further improving the reliability of the correction results.

[0115] Step S205: determining the correction result of the underlined character phonetic notation question based on the comparison result between the target phonetic notation and the target answer phonetic notation.

[0116] For a comparison of the target phonetic notation and the target answer results, please see above. Figure 1 The detailed description of step S105 in the illustrated embodiment is omitted here.

[0117] In some optional implementations, the above step S205 includes: if any of the target phonetic notations matches any of the target answer phonetic notations, determining that the correction result of the question with the matching phonetic notations is correct.

[0118] For each underlined character, its target phonetic notation corresponds to at least one phonetic notation; for the target answer phonetic notation of the answer result including at least one answer phonetic notation, at least one phonetic notation in the target phonetic notation is matched with at least one answer phonetic notation in the target answer phonetic notation. If they match, the correction result of the question with matched phonetic notation is determined to be correct.

[0119] The matching method includes but is not limited to calculating phonetic similarity, etc., and is specifically set according to actual needs and is not limited here.

[0120] Due to errors in character recognition or text recognition, any phonetic notation is matched with any phonetic notation of the answer. If any one of them matches, it means that the correction result is that the answer is correct, which can effectively reduce the risk of errors in recognizing similar characters.

[0121] The method for correcting the pronunciation of underlined characters provided by the embodiment of the present invention adopts a character classification recognition method to recognize the underlined character image, which can improve the accuracy of the obtained underlined characters. At the same time, after character classification recognition, a second preset number of underlined characters are obtained, which can provide a certain recognition redundancy for the recognition result to ensure the reliability of the subsequent correction results. The answer result is first screened using the preset answer probability. When the probability of the answer result is less than the preset answer probability, it means that there is no answer or the answer result does not meet the requirements of the question. Therefore, the correction result of the small question corresponding to the answer result is directly set as an error without the need for subsequent recognition, which improves the probability of homework correction. When the probability of the answer result is greater than the preset answer probability, the picture of the answer result is cut out, and text line recognition is performed on it to obtain the text line recognition result, and then it is screened to obtain the first preset number of answer pronunciations. By retaining the first preset number of answer pronunciations for subsequent comparison, the risk of misrecognition is reduced.

[0122] As a specific application example of the present invention, a correction application is installed in a mobile phone. When it is necessary to perform pronunciation recognition of underlined characters, the underlined character pronunciation recognition function in the correction application is activated. Accordingly, the camera of the mobile phone is activated to capture an image of the homework page to obtain the image to be corrected. The image to be corrected is recognized to obtain each sub-question in the underlined character pronunciation question, the underlined characters in each sub-question, and the answer result of the underlined characters. First, the probability in the answer result of the underlined characters is compared with the preset answer probability. If the probability is less than the preset answer probability, it means that the sub-question corresponding to the probability is an invalid answer, and accordingly, the correction result of the sub-question is wrong. For the sub-questions whose probability in the answer result of the underlined characters is greater than or equal to the preset probability, it means that the sub-question corresponding to the probability is a valid answer. Accordingly, the sub-question is further recognized to obtain the underlined characters and the answer pronunciation corresponding to the answer result; then, the correspondence between the characters and the pronunciation is used to obtain the target pronunciation corresponding to each underlined character. For each highlighted character, the phonetic notation in the target phonetic notation is matched with each answer phonetic notation in the target answer phonetic notation. If any of them match, the question is graded correctly. Finally, the grading result for the highlighted character phonetic notation question is displayed on the corresponding page on the phone, for example, the incorrectly answered question is displayed in the image to be graded.

[0123] In this embodiment, a device for correcting underlined characters and adding phonetic notation is also provided. The device is used to implement the above-mentioned embodiments and preferred embodiments, and the details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0124] This embodiment provides a device for correcting underlined characters with phonetic notation. Figure 4 Shown, including:

[0125] The picture acquisition module 401 is used to acquire the picture to be corrected.

[0126] The image recognition module 402 is used to recognize the image to be corrected, and obtain each sub-question under the question with underlined characters and phonetic notation, the underlined characters for each sub-question, and the answer result of the underlined characters.

[0127] The answer recognition module 403 is used to recognize the underlined characters and the answer results of each question to obtain the target underlined characters and the target answer phonetic notation.

[0128] The phonetic notation determination module 404 is configured to determine the target phonetic notation corresponding to the target underlined character based on the correspondence between the character and the phonetic notation.

[0129] The grading result determination module 405 is used to determine the grading result of the underlined character phonetic notation question based on the comparison result of the target phonetic notation and the target answer phonetic notation.

[0130] In some optional implementations, the answer result of the underlined word includes the probability of the answer result, and the answer recognition module 403 includes:

[0131] The preset answer probability obtaining unit is used to obtain the preset answer probability.

[0132] The first answer result determination unit is used to determine that the answer result of the underlined word is empty if the probability of the answer result is less than the preset answer probability, and the correction result of the small question corresponding to the answer result is wrong.

[0133] In some optional implementations, the answer result of the underlined word includes the position of the answer result, and the answer recognition module 403 further includes:

[0134] The picture acquisition unit is used to acquire a picture of the answer result based on the position of the answer result if the probability of the answer result is greater than or equal to the preset answer probability.

[0135] The text line recognition unit is used to perform text line recognition on the picture of the answer result to obtain a text line recognition result.

[0136] The recognition result screening unit is used to screen the text line recognition results and determine target answering phonetic notations, where the target answering phonetic notations include a first preset number of answering phonetic notations.

[0137] In some optional implementations, the text line recognition result includes a probability that the character at each position in the answer result belongs to a preset character, and the recognition result screening unit includes:

[0138] The query subunit is used to query the maximum probability that the character belongs to the preset character at each position.

[0139] The first answer result determination subunit is used to obtain a first answer result based on the preset character corresponding to the maximum probability of each position.

[0140] The probability calculation subunit is used to, for each position, query the second largest probability that the character belongs to the preset character and calculate the difference between the maximum probability and the second largest probability.

[0141] The target position determination subunit is used to obtain the target position corresponding to the minimum difference value based on the difference value of each position.

[0142] The second answer result determination subunit is used to replace the character at the target position in the first answer result with the preset character corresponding to the second most probable target position to obtain the second answer result. The target answer phonetic notation includes the first answer result and the second answer result.

[0143] In some optional implementations, the answer recognition module 403 includes:

[0144] The underlined word image acquisition unit is used to acquire the underlined word image from the corresponding question based on the position of the underlined word.

[0145] The character classification and recognition unit is used to perform character classification and recognition on the underlined character picture to obtain a second preset number of underlined characters, and the target underlined characters include the second preset number of underlined characters.

[0146] In some optional implementations, if the target underlined characters include a first preset number of underlined characters, the phonetic notation determination module 404 includes:

[0147] The phonetic notation query unit is used to query the correspondence between each underlined character and the phonetic notation, and obtain at least one phonetic notation corresponding to the underlined character. If the underlined character is a polyphonic character, the underlined character corresponds to at least two phonetic notations.

[0148] In some optional implementations, the review result determination module 405 includes:

[0149] The phonetic notation matching unit is used to determine that the correction result of the question with the matching phonetic notation is correct if any phonetic notation in the target phonetic notation matches any phonetic notation in the target answer phonetic notation.

[0150] The device for correcting the underlined characters and phonetic notation in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0151] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0152] An embodiment of the present invention further provides an electronic device having the above Figure 4 The device for correcting the pronunciation of underlined characters shown.

[0153] See also Figure 5 , Figure 5 is a structural diagram of an electronic device provided by an optional embodiment of the present invention, such as Figure 5 As shown, the electronic device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0154] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0155] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0156] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0157] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0158] The electronic device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 5 The bus connection is taken as an example.

[0159] The input device 30 can receive input digital or character information and generate key signal input related to user settings and function control of the electronic device, such as a touch screen, a keypad, a mouse, a trackpad, a touch pad, an indicator stick, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 can include a display device, an auxiliary lighting device (e.g., an LED), and a tactile feedback device (e.g., a vibration motor). The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display, and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0160] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0161] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for correcting underlined characters with phonetic notation, characterized in that: The method comprises: Get the pictures to be corrected; Recognize the picture to be corrected, and obtain each sub-question under the underlined word phonetic notation question, the underlined word for each sub-question, and the answer result for the underlined word; Recognize the underlined characters and the answer results of each question to obtain the target underlined characters and the target answer phonetic notation; Determining the target pronunciation corresponding to the target underlined character based on the correspondence between the character and the pronunciation; Based on the comparison result of the target phonetic notation and the target answer phonetic notation, the correction result of the underlined character phonetic notation question is determined.

2. The method according to claim 1, characterized in that The answer result of the underlined word includes the probability of the answer result, and the answer result of the underlined word of each question is recognized to obtain the target answer phonetic notation, including: Get the preset answer probability; If the probability of the answer result is less than the preset answer probability, the answer result of the underlined word is determined to be empty, and the correction result of the small question corresponding to the answer result is wrong.

3. The method according to claim 2, characterized in that The answer result of the underlined characters includes the position of the answer result, and the answer result of the underlined characters of each question is recognized to obtain the target answer phonetic notation, further comprising: If the probability of the answer result is greater than or equal to the preset answer probability, obtaining a picture of the answer result based on the position of the answer result; Performing text line recognition on the picture of the answer result to obtain a text line recognition result; The text line recognition result is screened to determine the target answering phonetic notation, where the target answering phonetic notation includes a first preset number of answering phonetic notations.

4. The method according to claim 3, characterized in that The text line recognition result includes a probability that a character at each position in the answer result belongs to a preset character. The screening of the text line recognition result to determine the target answer phonetic notation includes: For each position, the maximum probability that the query character belongs to the preset characters; Obtaining a first answer result based on the preset character corresponding to the maximum probability of each position; For each position, querying the second largest probability that the character belongs to the preset character and calculating the difference between the maximum probability and the second largest probability; Obtaining a target position corresponding to a minimum difference value based on the difference value of each position; The character at the target position in the first answer result is replaced with a preset character corresponding to the second most probable character at the target position to obtain a second answer result, wherein the target answer phonetic notation includes the first answer result and the second answer result.

5. The method according to claim 1, wherein The step of identifying the underlined characters for each question to obtain the target underlined characters includes: Obtaining an underlined word image from the corresponding question based on the position of the underlined word; Character classification and recognition are performed on the underlined character picture to obtain a second preset number of underlined characters, and the target underlined characters include the second preset number of underlined characters.

6. The method according to claim 1, characterized in that If the target underlined characters include a first preset number of underlined characters, determining target pronunciation corresponding to the target underlined characters based on the correspondence between characters and pronunciation includes: For each of the underlined characters, the correspondence between the character and the pronunciation is queried to obtain at least one pronunciation corresponding to the underlined character. If the underlined character is a polyphonic character, the underlined character corresponds to at least two of the pronunciations.

7. The method according to any one of claims 1 to 6, characterized in that The step of determining the correction result of the underlined character phonetic notation question based on the comparison result between the target phonetic notation and the target answer phonetic notation includes: If any of the target phonetic notations matches any of the target answer phonetic notations, the correction result of the question with matching phonetic notations is determined to be correct.

8. A device for correcting underlined characters with phonetic notation, characterized in that: The device comprises: Image acquisition module, used to obtain images to be corrected; An image recognition module is used to recognize the image to be corrected, and obtain each sub-question under the underlined character phonetic notation question, the underlined character for each sub-question, and the answer result for the underlined character; An answer recognition module is used to recognize the underlined characters of each question and the answer results of the underlined characters to obtain the target underlined characters and the target answer phonetic notation; A phonetic notation determination module, configured to determine a target phonetic notation corresponding to the target underlined character based on a correspondence between the character and the phonetic notation; The correction result determination module is used to determine the correction result of the underlined character phonetic notation question based on the comparison result of the target phonetic notation and the target answer phonetic notation.

9. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for correcting the underlined Chinese characters and phonetic notation according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for correcting the underlined Chinese characters and phonetic notation according to any one of claims 1 to 7.

Images