A Dongba structural single-element character recognition method, device, equipment and storage medium

Abstract

The application discloses a Dongba structure type single element character recognition method and device, equipment and storage medium, relates to the technical field of character recognition, and comprises the following steps: determining a to-be-recognized Dongba structure type single element character; extracting an initial feature curve of the to-be-recognized Dongba structure type single element character; obtaining a simplified feature curve through further simplification processing; extracting each local feature curve from the simplified feature curve, and calculating a curve weight based on the number of feature points in each local feature curve; based on the matching relationship between the preset weight of a template character and the curve weight of the local feature curve, filtering out a plurality of corresponding groups of template characters, and respectively calculating the similarity between the local feature curve and the template character to determine the target template character. In this way, the corresponding template character can be determined from the preset template character library according to the curve weight of the local feature curve in the Dongba structure type single element character, and the recognition efficiency and accuracy of the Dongba structure type single element character can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of character recognition, and in particular to a Dongba structured single-element character recognition method, device, equipment and storage medium. Background Art

[0002] Based on an analysis of the structural elements of Dongba characters, they can be divided into two categories: simple characters and complex characters. Simple characters refer to independent characters that can directly express sound and meaning, while complex characters are composed of two or more characters, generally using multiple characters to express sound and meaning. Furthermore, simple characters can be further divided into outline-type and structural-type simple characters. Contour-type simple characters express their meaning by imitating the external contours of objects, reflecting the essence of the characters through their external contours, and generally include complete character outlines. Structural-type characters, on the other hand, use simple character strokes to express meaning by depicting the structure of an object. The structure or skeleton of an object is prominent, but there are no obvious contour features. Therefore, the external contour of the character is generally used to describe the characteristics of outline-type simple characters, while the structure or skeleton of the character is used to describe the characteristics of structural-type simple characters.

[0003] Structural monograms are a type of writing that expresses meaning by describing the structural features of objects. These characters, primarily human-shaped, have prominent structural features and rarely depict the external characteristics of objects. In the study of Dongba scripts, the rapid recognition of Dongba structural monograms is a key challenge in this field. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method, device, equipment and storage medium for recognizing Dongba structural single-element characters, which can improve the efficiency of character recognition of Dongba structural single-element characters. The specific scheme is as follows:

[0005] In a first aspect, the present application provides a Dongba structural monomorphic character recognition method, comprising:

[0006] Determining a Dongba structural single element character to be identified; wherein the Dongba structural single element character to be identified includes several character structures;

[0007] Extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm;

[0008] Simplifying the initial characteristic curve based on a discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single-element character to be identified;

[0009] Extracting local feature curves corresponding to the character structures from the simplified feature curve, and calculating a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure;

[0010] Based on the matching relationship between the preset weights of the template characters and the curve weights of the local feature curves, several groups of template characters corresponding to the local feature curves are screened out from the preset template character library, and the similarity between each local feature curve and each template character in the corresponding group of template characters is calculated respectively, so as to determine the target template character corresponding to the Dongba structural single element character to be identified from the preset template character library based on the similarity.

[0011] Optionally, before extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm, the method further includes:

[0012] Preprocessing the Dongba character image containing the Dongba structural element character to obtain a processed character image of the Dongba structural element character;

[0013] Accordingly, the method for extracting the initial characteristic curve of Dongba structural single-element characters based on the connected domain priority marking algorithm includes:

[0014] The character feature curve in the processed character image is extracted based on the connected component labeling algorithm to obtain the initial feature curve.

[0015] Optionally, the preprocessing of the Dongba character image containing the Dongba structural element character to obtain a processed character image of the Dongba structural element character comprises:

[0016] The resolution of the Dongba character image is adjusted based on a preset image resolution to obtain the processed character image of the Dongba structural element character.

[0017] Optionally, the method of extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm includes:

[0018] Determine the starting point of the character corresponding to the Dongba structural single element character according to the connected domain priority marking algorithm, and use the starting point as the current known point;

[0019] Calculating priority weights of several vertices adjacent to the current known point based on the connected domain priority marking algorithm;

[0020] The next known point is determined according to the priority weight, and the next known point is determined as the new current known point, and then the step of calculating the priority weights of several vertices adjacent to the current known point based on the connected domain priority marking algorithm is jumped to obtain the initial characteristic curve of the Dongba structural single element character.

[0021] Optionally, the calculating the curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure includes:

[0022] Counting the number of all feature points in the simplified feature curve to obtain the total number of feature points;

[0023] The curve weight corresponding to each local feature curve is calculated according to the ratio between the number of feature points in each local feature curve in the simplified feature curve and the total number of feature points.

[0024] Optionally, the step of selecting a plurality of groups of template characters corresponding to the local feature curves from a preset template character library based on a matching relationship between preset weights of the template characters and curve weights of the local feature curves includes:

[0025] The similarity between the local characteristic curve with the largest curve weight in the Dongba structural single element character and several local characteristic curves with the largest curve weight in the template character library is calculated to screen out several groups of template characters whose similarity is greater than a first preset similarity threshold.

[0026] Optionally, respectively calculating the similarity between each of the local feature curves and each template character in a corresponding set of template characters, so as to determine a target template character corresponding to the Dongba structural element character to be identified from the preset template character library based on the similarity, comprises:

[0027] Calculating similarities between all local feature curves in the Dongba structural single element character and all local feature curves in the plurality of template character groups based on the curve weights, and calculating an average of the similarities between the local feature curves to obtain an average similarity;

[0028] If the average similarity is greater than a second preset similarity threshold, it is determined that the corresponding template character corresponds to the Dongba structural type single element character.

[0029] In a second aspect, the present application provides a Dongba structured single-element character recognition device, comprising:

[0030] A character determination module is used to determine a Dongba structural element character to be identified; the Dongba structural element character to be identified includes several character structures;

[0031] A feature curve extraction module is used to extract the initial feature curve of the Dongba structural single element character to be identified based on a connected domain priority marking algorithm;

[0032] a characteristic curve simplification module, configured to simplify the initial characteristic curve based on a discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single element character to be identified;

[0033] a weight calculation module, configured to extract local feature curves corresponding to the character structures from the simplified feature curve, and calculate a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure;

[0034] The template character determination module is used to screen out a plurality of groups of template characters corresponding to each of the local feature curves from a preset template character library based on a matching relationship between the preset weights of the template characters and the curve weights of the local feature curves, and calculate the similarity between each of the local feature curves and each template character in the corresponding group of template characters, so as to determine a target template character corresponding to the Dongba structural single element character to be identified from the preset template character library based on the similarity.

[0035] In a third aspect, the present application provides an electronic device, comprising:

[0036] Memory, used to store computer programs;

[0037] A processor is used to execute the computer program to implement the above-mentioned Dongba structural single-element character recognition method.

[0038] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the Dongba structural monomorphic character recognition method as described above.

[0039] It can be seen that the application first determines a to-be-recognized Dongba structure type single-element character, the to-be-recognized Dongba structure type single-element character contains a plurality of character structures, then extracts an initial feature curve of the to-be-recognized Dongba structure type single-element character based on a connected domain priority marking algorithm, then simplifies the initial feature curve based on a discrete curve evolution algorithm to obtain a simplified feature curve of the to-be-recognized Dongba structure type single-element character, then extracts a local feature curve corresponding to each character structure from the simplified feature curve, and calculates a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure, then filters a plurality of groups of template characters respectively corresponding to each local feature curve from a preset template character library based on a matching relationship between a preset weight of a template character and the curve weight of the local feature curve, and respectively calculates a similarity between each local feature curve and each template character in a corresponding group of template characters, so as to determine a target template character corresponding to the to-be-recognized Dongba structure type single-element character from the preset template character library based on the similarity. In this way, the application can calculate the curve weight of the local feature curve in the corresponding simplified feature curve according to the number of feature points in the Dongba structure type single-element character, then calculate the similarity of the local feature curve between the Dongba structure type single-element character and the preset template character based on the curve weight, and determine the template character corresponding to the Dongba structure type single-element character from the preset template character library according to the similarity. In this way, the similarity between the local feature curves of the Dongba structure type single-element character and the preset template character is respectively calculated according to the different curve weights, which can improve the matching degree between the local feature curves, and further improve the recognition efficiency of the Dongba structure type single-element character. BRIEF DESCRIPTION OF DRAWINGS

[0040] 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 needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on the provided drawings.

[0041] Figure 1 A Dongba structure type single-element character recognition method flowchart disclosed by the present application;

[0042] Figure 2 A Dongba structure type single-element character schematic diagram under different resolutions disclosed by the present application;

[0043] Figure 3 A character feature curve extraction result schematic diagram disclosed by the present application;

[0044] Figure 4A character feature curve simplification result schematic view disclosed by the present application;

[0045] Figure 5 A specific Dongba structure type single element character recognition method flow chart disclosed by the present application;

[0046] Figure 6 A character similarity calculation result schematic view disclosed by the present application;

[0047] Figure 7 A Dongba structure type single element character recognition device structure schematic view disclosed by the present application;

[0048] Figure 8 An electronic device structure view disclosed by the present application. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only 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 skilled in the art without creative labor fall within the scope of protection of the present application.

[0050] The technical solutions of the present application can recognize Dongba structure type single element characters. Specifically, the feature curve extraction of Dongba structure type single element characters can be realized by using a connected domain priority marking algorithm based on chain code, to obtain an initial feature curve. It should be noted that the initial feature curve of the Dongba structure type single element character has been segmented into several local feature curves while being simplified and extracted by using the connected domain priority marking algorithm. Therefore, in the present application, the priority of each local feature curve can be determined according to the curve weight of each local feature curve in the Dongba structure type single element character relative to the entire Dongba structure type single element character. Then, the similarity between each local feature curve and each local feature curve in the template character in the preset template character is calculated according to different priorities, to determine the template character corresponding to the Dongba structure type single element character.

[0051] Referring to Figure 1 The embodiments of the present application disclose a Dongba structure type single element character recognition method, which comprises the following steps:

[0052] In step S11, a Dongba structure type single element character to be recognized is determined. The Dongba structure type single element character to be recognized contains several types of character structures.

[0053] It can be understood that most of the Dongba structural element characters are composed of line skeletons, which contain different local lines, that is, different text structures. It should be pointed out that in the present application, before processing the Dongba structural element characters to be identified, it can also include: pre-processing the Dongba character image containing the Dongba structural element characters to obtain the processed character image of the Dongba structural element characters; further, the pre-processing of the Dongba character image containing the Dongba structural element characters to obtain the processed character image of the Dongba structural element characters can include: adjusting the resolution of the Dongba character image based on the preset image resolution to obtain the processed character image of the Dongba structural element characters. Specifically, in the pre-processing stage, the interference caused by the redundancy of the text strokes can be reduced by reducing the text resolution, thereby highlighting the structural characteristics of the text. For example, Figure 2 As shown, the connected domain priority marking algorithm is used to extract feature curves for character images with resolutions of 100px, 75px, 50px and 25px respectively. It is easy to see that when the resolution of the text is too high, the redundancy of the text strokes is relatively high; and when the resolution of the text is too low, the accuracy of the extracted feature curve will also be affected. In a specific embodiment, after many attempts, the resolution of the Dongba structural single element character can be set to 50px. At the same time, considering the different lengths and widths of structural single element characters of different forms, for a Dongba structural single element character, if the length of the character is len pixels and the height is high pixels, the reset rule of the character size is:

[0054]

[0055] Step S12: extracting the initial feature curve of the Dongba structural monomorphic character to be identified based on the connected component priority marking algorithm.

[0056] It should be noted that in specific embodiments, after obtaining the processed character picture through step S11, the initial feature curve of the Dongba structural single-element character extracted based on the connected domain priority marking algorithm can include: extracting the character feature curve in the processed character picture based on the connected domain marking algorithm to obtain the initial feature curve. Specifically, after pre-processing the character picture of the Dongba structural single-element character, the initial feature curve of the processed character picture can be extracted based on the connected domain marking algorithm; accordingly, in the embodiments of the present application, the initial feature curve of the Dongba structural single-element character to be recognized extracted based on the connected domain priority marking algorithm can include: determining the starting point of the character corresponding to the Dongba structural single-element character according to the connected domain priority marking algorithm, and taking the starting point as the current known point; calculating the priority weight values of a plurality of vertices adjacent to the current known point based on the connected domain priority marking algorithm; determining the next known point according to the priority weight values, and determining the next known point as a new current known point, and then jumping to the step of calculating the priority weight values of a plurality of vertices adjacent to the current known point based on the connected domain priority marking algorithm to obtain the initial feature curve of the Dongba structural single-element character. Specifically, the connected domain priority marking algorithm can determine the starting point of the character and the initial direction vector, and then determine the next vertex by calculating the priority weight values of the adjacent vertices, and so on, so that the starting point can return to the starting point through all the lines of the character, so that the initial feature curve corresponding to the Dongba structural single-element character can be obtained. As shown in FIG. 8, it is a schematic diagram of a Dongba structural single-element character and a corresponding initial feature curve. Figure 3

[0057] Step S13, simplifying the initial feature curve based on a discrete curve evolution algorithm to obtain a simplified feature curve of the Dongba structural single-element character to be recognized.

[0058] In the embodiments of the present application, after obtaining the initial feature curve corresponding to the Dongba structural single-element character, the initial feature curve can be simplified based on the discrete curve evolution algorithm; it can be understood that in order to reduce the complexity of the similarity comparison, the initial feature curve of the character can also be simplified before comparison to remove redundant and abnormal feature points. In a specific embodiment, in each stage of evolution, a new line segment is used to replace a pair of adjacent original line segments s1 and s2, and the new line segment is obtained by connecting the two end points of s1∪s2. In the evolution process, the merging order of the line segments is determined by the metric K, that is:

[0059]

[0060] ​Among them, β(s1, s2) is the vertex turning angle of segments s1 and s2, and l represents the normalized segment length. By continuously merging redundant segments in the local feature curve, the number of curve vertices is reduced and the curve features are preserved, while the complexity of the curve can be greatly reduced. The simplified results of the local feature curve of the structural single element character are as follows: Figure 4 As shown in the figure, the dot representation is the simplified result of the initial characteristic curve, and the dotted-line representation is for ease of understanding. Through simplification, 70% of the redundant points in the characteristic curve are removed. It can be seen that although a large number of original characteristic points have been removed, the curve shape has not changed significantly, which shows that the characteristic curve simplification process before similarity comparison is necessary.

[0061] Step S14: extracting local feature curves corresponding to the character structures from the simplified feature curve, and calculating the curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure.

[0062] It is understandable that, since the characteristic curve of Dongba structural single element character is composed of the local characteristic curves representing the structure of the character, the more complex the structure of the character, the more local characteristic curves that make up the character. As the number of local curves increases, some "universal" curves that do not have unique characteristics may be produced, such as straight line segments or oblique line segments. Since the vast majority of local curves may be characteristic curves composed of universal elements such as straight lines, curves and oblique lines, these local curves with "universality" may produce larger similarity values ​​when compared with any characteristic curve, and the similarity comparison algorithm of the curves will not be able to give play to its original advantages. In the present application, the proportion of each local characteristic curve in the entire Dongba structural single element character, i.e., the curve weight, can be calculated based on the number of characteristic points in the local characteristic curves.

[0063] Step S15: Based on the matching relationship between the preset weights of the template characters and the curve weights of the local feature curves, a plurality of groups of template characters corresponding to the local feature curves are selected from the preset template character library, and the similarity between each local feature curve and each template character in the corresponding group of template characters is calculated, so as to determine the target template character corresponding to the Dongba structural single element character to be identified from the preset template character library based on the similarity.

[0064] In this embodiment, after obtaining the curve weights corresponding to each local feature curve in the simplified feature curve, a plurality of groups of template characters corresponding to the local feature curves of the simplified feature curve can be selected from a preset template character library based on the matching relationship between the preset weights of the local feature curves of the template characters and the curve weights of the local feature curves of the simplified feature curve. Furthermore, the similarity between each local feature curve in the simplified feature curve and the corresponding local feature curves in the plurality of groups of template characters can be calculated, so that the target template character corresponding to the simplified feature curve of the Dongba structural single element character to be recognized, i.e., the recognition result of the Dongba structural single element character, can be determined from the preset template character library based on the similarity between the local feature curves. In a specific embodiment, the curve weights of each local feature curve in the template character can be compared one by one with the local feature curve of the Dongba structural single element character to be recognized, i.e., the test curve, to calculate the similarity between each pair of local feature curves. It should be pointed out that in order to ensure that there are only 0 or 1 corresponding local feature curves between the local feature curves of the template character or the local feature curves of the character to be tested, when a set of corresponding curves is found, they will be uniformly removed from the local feature curve lists of the template character and the character to be tested.

[0065] As can be seen from this, in this application, the interference caused by redundant strokes can be reduced by adjusting the character image resolution of Dongba structural monomorphic characters to highlight the structural characteristics of the characters. Moreover, the initial feature curve can be simplified using a discrete curve evolution algorithm, which can remove most of the redundant points in the feature curve while maintaining the shape of the feature curve, reducing the complexity of subsequent similarity comparisons. Furthermore, this application can use the number of feature points to calculate the curve weight of the local feature curve, and based on the curve weight, perform similarity calculations between subsequent characters, which can improve the recognition accuracy of Dongba structural monomorphic characters.

[0066] See also Figure 5 As shown, the embodiment of the present invention discloses a Dongba structured single-element character recognition method, comprising:

[0067] Step S21 , determining Dongba structural element characters to be identified; the Dongba structural element characters to be identified include several types of character structures.

[0068] Step S22: extracting the initial characteristic curve of the Dongba structural single-element character to be identified based on the connected component priority marking algorithm.

[0069] Step S23: simplifying the initial characteristic curve based on the discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single-element character to be recognized.

[0070] Step S24: extracting local feature curves corresponding to the text structures from the simplified feature curves.

[0071] Step S25: Count the number of all feature points in the simplified feature curve to obtain the total number of feature points.

[0072] In an embodiment of the present application, in order to facilitate the similarity calculation of local feature curves, the curve weight of the local feature curve can be calculated, that is, the contribution rate of a certain local feature curve to the entire character feature; in a specific embodiment, it is first necessary to count the number of all feature points in the simplified feature curve, that is, to obtain the total number of feature points of the entire character, so as to calculate the curve weight of a single local feature curve.

[0073] Step S26 : Calculate the curve weight corresponding to each local feature curve according to the ratio between the number of feature points in each local feature curve in the simplified feature curve and the total number of feature points.

[0074] Furthermore, the curve weight corresponding to each local feature curve can be calculated based on the ratio between the number of feature points in each local feature curve and the total number of feature points in the simplified feature curve. In a specific embodiment, for a feature curve Curve representing a Dongba structural monogram, if it includes n local feature curves and m feature points, then for a feature curve including m i The local feature curve partCurve of feature points i , the following relationship exists:

[0075]

[0076] Since the character feature curve obtained at this time is already a simplified feature curve, most of the feature points in the curve contain the corresponding features of the character. Therefore, this embodiment can use the ratio of the feature points in the local feature curve to the total feature points in the simplified feature curve as the weight of the local feature curve, so each local feature curve partCurve i The curve weight weight i The calculation formula is:

[0077]

[0078] Step S27: calculating the similarity between the local feature curve with the largest curve weight in the Dongba structural single element character and several local feature curves with the largest curve weight in the template character library, so as to screen out several groups of template characters whose similarity is greater than a first preset similarity threshold.

[0079] Furthermore, after obtaining the curve weights of each local feature curve in the simplified feature curve, the similarity between the local feature curve with the largest curve weight in the corresponding Dongba structural single element character and several local feature curves with the largest curve weights in several characters in the template character library can be calculated. In this way, several groups of template characters can be screened out based on whether the calculated similarity is greater than the first preset similarity threshold. It can be understood that at this time, a group of template characters can include the various local feature curves of a single template character.

[0080] Step S28: calculating similarities between all local feature curves in the Dongba structural single-element character and all local feature curves in the plurality of template character groups based on the curve weights, and calculating an average of the similarities between the local feature curves to obtain an average similarity.

[0081] In an embodiment of the present application, after screening out several groups of template characters based on the similarities between the local feature curves under the calculated curve weights, the similarities between all the local feature curves in the Dongba structural single element characters and all the local feature curves in the several groups of template characters can be calculated based on the curve weights corresponding to the local feature curves. In a specific embodiment, the similarities between the local feature curves under the curve weights can be calculated in descending order; then the average value of the similarities between the local feature curves is calculated to obtain the average similarity. In a specific embodiment, as Figure 6 The figure shows a schematic diagram of similarity results, including the similarity between the calculated local feature curve in the template character and the local feature curve of the character to be tested (i.e., the Dongba structural single element character to be recognized) and the average similarity between each local feature curve.

[0082] Step S29: If the average similarity is greater than a second preset similarity threshold, it is determined that the corresponding template character corresponds to the Dongba structural single-element character.

[0083] Furthermore, after obtaining the average similarity between the local feature curve of the Dongba structural single element character and the local feature curves in the corresponding groups of template characters through step S28, it is possible to determine which template character corresponds to the current Dongba structural single element character by judging whether the average similarity is greater than a second preset similarity threshold, that is, the recognition result of the Dongba structural single element character to be recognized.

[0084] For more specific processing procedures of the above steps S21, S22, S23 and S24, reference may be made to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.

[0085] It can be seen that in the embodiment of the present application, the ratio between the number of feature points in each local feature curve and the total number of feature points in the simplified feature curve of the entire Dongba structural single element character can be regarded as the curve weight of the local feature curve, and then the correspondence between the template character and the local feature curves with different curve weights in the Dongba structural single element character to be identified is calculated to calculate the similarity between each local feature curve, and the similarity between each local feature curve can be calculated in order from large to small according to the curve weight. In this way, the similarity is calculated in combination with the curve weight of the local feature curve, and the similarity between each local feature curve in the Dongba structural single element character to be identified and the corresponding local feature curve in the template character can be accurately calculated, which can avoid the false high similarity when comparing the straight line segments or oblique line segments in the characters, and thus improve the reliability of Dongba structural single element character recognition.

[0086] like Figure 7 As shown, the present application discloses a Dongba structured single-element character recognition device, comprising:

[0087] The character determination module 11 is used to determine a Dongba structural element character to be identified; the Dongba structural element character to be identified includes several character structures;

[0088] A characteristic curve extraction module 12 is used to extract the initial characteristic curve of the Dongba structural type single element character to be identified based on a connected domain priority marking algorithm;

[0089] a characteristic curve simplification module 13, configured to simplify the initial characteristic curve based on a discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single-element character to be identified;

[0090] a weight calculation module 14, configured to extract local feature curves corresponding to the character structures from the simplified feature curve, and calculate a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure;

[0091] The template character determination module 15 is used to screen out a plurality of groups of template characters corresponding to each of the local feature curves from a preset template character library based on a matching relationship between the preset weights of the template characters and the curve weights of the local feature curves, and calculate the similarity between each of the local feature curves and each template character in the corresponding group of template characters, so as to determine the target template character corresponding to the Dongba structural single element character to be identified from the preset template character library based on the similarity.

[0092] Thus, the present application can calculate the curve weight of the local characteristic curve in the corresponding simplified characteristic curve based on the number of characteristic points in the Dongba structural single element character, and then calculate the similarity of the local characteristic curve between the Dongba structural single element character and the preset template character based on the curve weight, and determine the template character corresponding to the Dongba structural single element character from the preset template character library based on the similarity. In this way, by calculating the similarity between the local characteristic curves of the Dongba structural single element character and the preset template character based on the different curve weights, the matching degree between the various local characteristic curves when calculating the similarity can be improved, thereby improving the recognition efficiency of the Dongba structural single element character.

[0093] In a specific embodiment, the device may further include:

[0094] a character image preprocessing submodule, configured to preprocess the Dongba character image containing the Dongba structural element character to obtain a processed character image of the Dongba structural element character;

[0095] Accordingly, the characteristic curve extraction module 12 may include:

[0096] The first feature curve extraction unit is configured to extract the character feature curve in the processed character image based on the connected component labeling algorithm to obtain the initial feature curve.

[0097] In another specific embodiment, the character image preprocessing submodule may include:

[0098] The resolution adjustment unit is used to adjust the resolution of the Dongba character image based on a preset image resolution to obtain the processed character image of the Dongba structural element character.

[0099] In a specific embodiment, the characteristic curve extraction module 12 may include:

[0100] a starting point determining unit, configured to determine the starting point of the character corresponding to the Dongba structural element character according to the connected domain priority marking algorithm, and use the starting point as the current known point;

[0101] A vertex priority calculation unit, configured to calculate priority weights of a plurality of vertices adjacent to the current known point based on the connected domain priority marking algorithm;

[0102] The second characteristic curve extraction unit is used to determine the next known point according to the priority weight, and determine the next known point as the new current known point, and then jump to the step of calculating the priority weights of several vertices adjacent to the current known point based on the connected domain priority marking algorithm to obtain the initial characteristic curve of the Dongba structural single element character.

[0103] In a specific embodiment, the weight calculation module 14 may include:

[0104] A total feature point number counting unit, configured to count the number of all feature points in the simplified feature curve to obtain a total feature point number;

[0105] The curve weight calculation unit is configured to calculate the curve weight corresponding to each local feature curve according to the ratio between the number of feature points in each local feature curve in the simplified feature curve and the total number of feature points.

[0106] In a specific embodiment, the template character determination module 15 may include:

[0107] The first similarity calculation unit is used to calculate the similarity between the local feature curve with the largest curve weight in the Dongba structural single element character and several local feature curves with the largest curve weight in the template character library, so as to screen out several groups of template characters whose similarity is greater than a first preset similarity threshold.

[0108] In another specific embodiment, the template character determination module 15 may include:

[0109] a second similarity calculation unit, configured to calculate similarities between all local feature curves in the Dongba structural single element character and all local feature curves in the plurality of groups of template characters based on curve weights, and calculate an average of the similarities between the local feature curves to obtain an average similarity;

[0110] The character judgment unit is used to judge that the corresponding template character corresponds to the Dongba structural type single element character when the average similarity is greater than a second preset similarity threshold.

[0111] Furthermore, the embodiment of the present application also discloses an electronic device, Figure 8 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content in the diagram should not be considered as any limitation to the scope of application of the present application.

[0112] Figure 8 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps of the Dongba structured monomorphic character recognition method disclosed in any of the aforementioned embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0113] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and the external device. The communication protocol it follows is any communication protocol that can be applied to the technical solution of this application and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world. Its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0114] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or CD, etc. The resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0115] The operating system 221 is used to manage and control the hardware devices and computer program 222 on the electronic device 20. The operating system 221 can be Windows Server, NetWare, Unix, Linux, etc. In addition to including a computer program capable of performing the Dongba structured monomorphic character recognition method performed by the electronic device 20 as disclosed in any of the aforementioned embodiments, the computer program 222 can further include computer programs capable of performing other specific tasks.

[0116] Furthermore, this application discloses a computer-readable storage medium for storing a computer program; wherein, when executed by a processor, the computer program implements the aforementioned method for recognizing Dongba structural monomorphic characters. The specific steps of this method can be found in the corresponding contents disclosed in the aforementioned embodiments and will not be further described here.

[0117] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from the other embodiments. Reference can be made to the descriptions of the identical or similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and the relevant parts can be referred to the descriptions of the methods.

[0118] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0119] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0120] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0121] The above is a detailed introduction to the technical solution provided by the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for those skilled in the art, according to the ideas of the present application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A Dongba structural single-element character recognition method, characterized in that: include: Determining a Dongba structural single element character to be identified; wherein the Dongba structural single element character to be identified includes several character structures; Extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm; Simplifying the initial characteristic curve based on a discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single-element character to be identified; Extracting local feature curves corresponding to the character structures from the simplified feature curve, and calculating a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure; Based on a matching relationship between preset weights of the template characters and the curve weights of the local feature curves, a plurality of groups of template characters corresponding to the local feature curves are screened from a preset template character library, and similarities are calculated between each local feature curve and the feature curves of each template character in a corresponding group of template characters, so as to determine a target template character corresponding to the Dongba structural type single element character to be recognized from the preset template character library based on the similarities; The method of selecting a plurality of groups of template characters corresponding to the local feature curves from a preset template character library based on a matching relationship between the preset weights of the template characters and the curve weights of the local feature curves comprises: Calculating the similarity between the local characteristic curve with the largest curve weight in the Dongba structural single element character and several local characteristic curves with the largest curve weight in the template character library, so as to screen out several groups of template characters whose similarity is greater than a first preset similarity threshold; The method of respectively calculating the similarity between each of the local feature curves and the feature curves of each template character in a corresponding set of template characters, and determining a target template character corresponding to the Dongba structural element character to be identified from the preset template character library based on the similarity, comprises: Calculating similarities between all local feature curves in the Dongba structural single element character and all local feature curves in the plurality of template character groups based on the curve weights, and calculating an average of the similarities between the local feature curves to obtain an average similarity; If the average similarity is greater than a second preset similarity threshold, it is determined that the corresponding template character corresponds to the Dongba structural type single element character.

2. The Dongba structural single-element character recognition method according to claim 1, characterized in that: Before extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm, the method further includes: Preprocessing the Dongba character image containing the Dongba structural element character to obtain a processed character image of the Dongba structural element character; Accordingly, the method for extracting the initial characteristic curve of Dongba structural single-element characters based on the connected domain priority marking algorithm includes: The character feature curve in the processed character image is extracted based on the connected component priority marking algorithm to obtain the initial feature curve.

3. The Dongba structural single-element character recognition method according to claim 2, characterized in that: The pre-processing of the Dongba character image containing the Dongba structural element character to obtain a processed character image of the Dongba structural element character comprises: The resolution of the Dongba character image is adjusted based on a preset image resolution to obtain the processed character image of the Dongba structural element character.

4. The Dongba structural single-element character recognition method according to claim 1, characterized in that: The method of extracting the initial characteristic curve of the Dongba structural type single element character to be identified based on the connected domain priority marking algorithm includes: Determine the starting point of the character corresponding to the Dongba structural single element character according to the connected domain priority marking algorithm, and use the starting point as the current known point; Calculating priority weights of several vertices adjacent to the current known point based on the connected domain priority marking algorithm; The next known point is determined according to the priority weight, and the next known point is determined as the new current known point, and then the step of calculating the priority weights of several vertices adjacent to the current known point based on the connected domain priority marking algorithm is jumped to obtain the initial characteristic curve of the Dongba structural single element character.

5. The Dongba structural single-element character recognition method according to claim 1, characterized in that: The calculating the curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure includes: Counting the number of all feature points in the simplified feature curve to obtain the total number of feature points; The curve weight corresponding to each local feature curve is calculated according to the ratio between the number of feature points in each local feature curve in the simplified feature curve and the total number of feature points.

6. A Dongba structured single-element character recognition device, characterized in that: include: A character determination module is used to determine the Dongba structural single-element character to be identified; The Dongba structural single element characters to be identified include several character structures; A feature curve extraction module is used to extract the initial feature curve of the Dongba structural single element character to be identified based on a connected domain priority marking algorithm; a characteristic curve simplification module, configured to simplify the initial characteristic curve based on a discrete curve evolution algorithm to obtain a simplified characteristic curve of the Dongba structural single element character to be identified; a weight calculation module, configured to extract local feature curves corresponding to the character structures from the simplified feature curve, and calculate a curve weight of each local feature curve in the simplified feature curve based on the number of feature points in each local feature curve corresponding to each character structure; a template character determination module, configured to select, from a preset template character library, a plurality of groups of template characters corresponding to the respective local feature curves based on a matching relationship between preset weights of the template characters and curve weights of the local feature curves, and respectively calculate a similarity between each local feature curve and a feature curve of each template character in a corresponding group of template characters, so as to determine, from the preset template character library, a target template character corresponding to the to-be-recognized Dongba structural type single element character based on the similarity; a first similarity calculation unit, configured to calculate the similarity between the local characteristic curve with the largest curve weight in the Dongba structural single element character and a plurality of local characteristic curves with the largest curve weight in the template character library, so as to screen out a plurality of groups of template characters having similarities greater than a first preset similarity threshold; a second similarity calculation unit, configured to calculate similarities between all local feature curves in the Dongba structural single element character and all local feature curves in the plurality of groups of template characters based on curve weights, and calculate an average of the similarities between the local feature curves to obtain an average similarity; The character judgment unit is used to judge that the corresponding template character corresponds to the Dongba structural type single element character when the average similarity is greater than a second preset similarity threshold.

7. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the Dongba structural monomorphic character recognition method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that Used to store a computer program, which, when executed by a processor, implements the Dongba structural monomorphic character recognition method according to any one of claims 1 to 5.

Images