Method, device and medium for obtaining correction text based on oral images

Abstract

The present application relates to the field of oral image processing technology, and in particular to a method, device, and medium for obtaining correction text based on oral images. The method comprises: obtaining an initial feature vector of a target oral cavity based on an initial set of target oral cavity images; matching the initial feature vector of the target oral cavity in an oral feature vector sample library to obtain a matched first oral cavity feature vector and a second oral cavity feature vector; obtaining a target change in the oral cavity feature vector corresponding to different correction times of the target oral cavity based on the matched first oral cavity feature vector and the second oral cavity feature vector; and obtaining correction text based on the correction time corresponding to the target oral cavity update image set, the target oral cavity update feature vector, and the target change in the oral cavity feature vector corresponding to different correction times of the target oral cavity. The present invention can improve the efficiency and convenience of oral condition assessment for subjects undergoing oral correction.

Description

Technical Field

[0001] The present invention relates to the technical field of oral image processing, and in particular to a method, device and medium for obtaining correction text based on oral images. Background Art

[0002] The process of oral correction is a dynamic process, and the oral correction status needs to be continuously tracked during the process. In the existing technology, the subject of oral correction usually goes to a fixed place to seek professionals to evaluate their oral correction status. Each evaluation takes the subject of oral correction a long time; and the time for the subject of oral correction to conduct oral status evaluation is usually pre-set by professionals. Usually, the interval between two adjacent times is long, which makes it difficult to achieve intensive monitoring of the correction status of the subject of oral correction, and is not conducive to timely detection of abnormal situations in the oral correction process. How to improve the efficiency and convenience of oral status evaluation of the subject of oral correction is an urgent problem to be solved. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, device and medium for obtaining correction text based on oral images, so as to improve the efficiency and convenience of oral condition assessment of subjects undergoing oral correction.

[0004] According to a first aspect of the present invention, a method for obtaining corrected text based on an oral image is provided, the method comprising the following steps:

[0005] The target oral cavity initial feature vector is obtained according to the target oral cavity initial picture set; the target oral cavity initial picture set includes several target oral cavity initial pictures taken from different angles perpendicular to the tooth surface, and different target oral cavity initial pictures in the target oral cavity initial picture set include teeth at different positions in the target oral cavity.

[0006] The initial feature vector of the target oral cavity is matched in the oral feature vector sample library to obtain a matched first oral feature vector and a second oral feature vector; the oral feature vector sample library includes several oral feature vector sets corresponding to sample oral cavity, and the oral feature vector set corresponding to any sample oral cavity includes the oral feature vectors corresponding to the sample oral cavity at different correction times; the first oral feature vector is the oral feature vector corresponding to the sample oral cavity that has not undergone correction treatment, and the second oral feature vector is the oral feature vector corresponding to the sample oral cavity that has undergone correction treatment.

[0007] According to the matched first oral feature vector and the second oral feature vector, the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained; the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained based on the change amount of the oral feature vector corresponding to the first oral feature vector at different correction times and the change amount of the oral feature vector corresponding to the second oral feature vector at different correction times.

[0008] The correction text is obtained according to the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity; the correction text includes the judgment result of whether the target oral cavity is corrected abnormally.

[0009] Furthermore, obtaining the correction text according to the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector, and the target change amount of the oral cavity feature vector corresponding to different correction time periods of the target oral cavity includes:

[0010] According to the correction time corresponding to the target oral updated picture set, the target change of the oral feature vector corresponding to the specified correction time is selected from the target oral feature vector target changes corresponding to different correction times; the specified correction time is the difference between the upload time of the target oral updated picture set and the upload time of the target oral initial picture set.

[0011] The target oral cavity updated feature vector is compared with the target oral cavity initial feature vector to obtain the oral cavity feature vector change corresponding to the target oral cavity updated feature vector.

[0012] It is determined whether the correction of the target oral cavity is abnormal based on the change amount of the oral cavity feature vector corresponding to the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to the specified correction time.

[0013] Furthermore, judging whether the correction of the target oral cavity is abnormal according to the oral cavity feature vector change amount corresponding to the target oral cavity update feature vector and the oral cavity feature vector target change amount corresponding to the specified correction time includes:

[0014] If the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the oral feature vector target change amount corresponding to the specified correction time is less than or equal to the preset difference threshold, it is determined that the correction of the target oral cavity is normal and the correction text is output.

[0015] If the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the target oral feature vector change amount corresponding to the specified correction time is greater than a preset difference threshold, then determine whether the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the oral feature vector update change amount corresponding to the specified correction time is less than or equal to the preset difference threshold. If it is less than or equal to, then determine that the correction of the target oral cavity is normal and output the correction text; otherwise, determine that the correction of the target oral cavity is abnormal and output the correction text; the oral feature vector update change amount corresponding to the specified correction time is the change amount in the first change amount and the second change amount that has a smaller difference with the oral feature vector change amount corresponding to the target oral updated feature vector, the first change amount is the oral feature vector change amount corresponding to the first oral feature vector in the specified correction time, and the second change amount is the oral feature vector change amount corresponding to the second oral feature vector in the specified correction time.

[0016] Furthermore, obtaining target changes of the oral cavity feature vector corresponding to different correction times of the target oral cavity according to the matched first oral cavity feature vector and the second oral cavity feature vector includes:

[0017] Acquire oral cavity feature vectors of the sample oral cavity corresponding to the first oral cavity feature vector, whose correction time is longer than the correction time corresponding to the first oral cavity feature vector, to form a first oral cavity feature vector subset.

[0018] For any oral feature vector in the first oral feature vector subset, the correction time difference between the oral feature vector and the first oral feature vector is determined as the correction time interval corresponding to the oral feature vector, and the oral feature vector change corresponding to the oral feature vector is obtained based on the oral feature vector and the first oral feature vector.

[0019] Acquire oral cavity feature vectors of the sample oral cavity corresponding to the second oral cavity feature vector, whose correction time is longer than the correction time corresponding to the second oral cavity feature vector, to form a second oral cavity feature vector subset.

[0020] For any oral feature vector in the second oral feature vector subset, the correction time difference between the oral feature vector and the second oral feature vector is determined as the correction time interval corresponding to the oral feature vector, and the oral feature vector change corresponding to the oral feature vector is obtained based on the oral feature vector and the second oral feature vector.

[0021] According to the oral feature vector change amount corresponding to the oral feature vector in the first oral feature vector subset corresponding to the same correction time interval and the oral feature vector change amount corresponding to the oral feature vector in the second oral feature vector subset corresponding to the same correction time interval, the target change amount of the oral feature vector corresponding to the same correction time interval is obtained.

[0022] Each identical correction time interval is determined as a correction time, and the target change amount of the oral cavity feature vector corresponding to each identical correction time interval is determined as the target change amount of the oral cavity feature vector corresponding to the corresponding correction time.

[0023] Furthermore, the target oral cavity initial feature vector is matched in the oral cavity feature vector sample library to obtain the matched first oral cavity feature vector and second oral cavity feature vector, including:

[0024] The oral feature vectors in the oral feature vector sample library are clustered to obtain several oral feature vector clusters.

[0025] Obtain the similarity between the target oral cavity initial feature vector and the center of each oral cavity feature vector cluster.

[0026] The cluster corresponding to the maximum similarity is divided into a first subcluster and a second subcluster; the oral feature vector in the first subcluster is the oral feature vector corresponding to the sample oral cavity that has not been corrected, and the oral feature vector in the second subcluster is the oral feature vector corresponding to the sample oral cavity that has been corrected.

[0027] The oral cavity feature vector in the first subcluster that is most similar to the target oral cavity initial feature vector is determined as the first oral cavity feature vector, and the oral cavity feature vector in the second subcluster that is most similar to the target oral cavity initial feature vector is determined as the second oral cavity feature vector.

[0028] Furthermore, the oral cavity feature vector includes at least one of the following feature vectors: the width of the gap between adjacent teeth, the degree of tooth arrangement, and the inclination angle of teeth.

[0029] Furthermore, the difference between the oral feature vector change amount corresponding to the target oral update feature vector and the target oral feature vector change amount corresponding to the specified correction time is the absolute value of the difference between the oral feature vector change amount corresponding to the target oral update feature vector and the target oral feature vector change amount corresponding to the specified correction time.

[0030] Furthermore, the target change of the oral cavity feature vector corresponding to any correction time of the target oral cavity is obtained by weighted summing the change of the oral cavity feature vector corresponding to the correction time of the first oral cavity feature vector and the change of the oral cavity feature vector corresponding to the correction time of the second oral cavity feature vector.

[0031] According to a second aspect of the present invention, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for obtaining corrected text based on oral images when executing the computer program.

[0032] According to a third aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for obtaining corrected text based on an oral image is implemented.

[0033] Compared with the prior art, the present invention has at least the following beneficial effects:

[0034] The present invention obtains the initial feature vector of the target oral cavity based on oral cavity pictures of the target oral cavity at different angles, and matches the initial feature vector of the target oral cavity with the oral cavity feature vectors in the oral cavity feature vector sample library to obtain a first oral cavity feature vector and a second oral cavity feature vector that match the initial feature vector of the target oral cavity. Based on the oral cavity feature vector set of the sample oral cavity corresponding to the first oral cavity feature vector and the second oral cavity feature vector, the present invention obtains the target change amount of the sample oral cavity feature vector corresponding to different correction time lengths of the target oral cavity; thus, after subsequently obtaining the difference between the target oral cavity updated feature vector and the target oral cavity initial feature vector, the relationship between the difference and the corresponding target change amount can be used to evaluate whether the target oral cavity is abnormally corrected. Compared with the existing technology, the present invention only requires the subject of oral correction to take oral pictures and upload them to obtain the evaluation results, which is more efficient and the process is more convenient. It can quickly monitor the oral correction status and is conducive to timely detection of abnormalities in the oral correction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0036] Figure 1 A flowchart of a method for obtaining corrected text based on an oral image provided in Example 1 of the present invention;

[0037] Figure 2 A flowchart of obtaining a first oral cavity feature vector and a second oral cavity feature vector provided in the first embodiment of the present invention;

[0038] Figure 3 This is a flowchart of obtaining target changes in the oral cavity feature vector corresponding to different correction times provided by the first embodiment of the present invention;

[0039] Figure 4 This is a flowchart of obtaining corrected text provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 any creative efforts shall fall within the scope of protection of the present invention. Example

[0041] According to this embodiment, Figure 1 As shown, a method for obtaining correction text based on oral images is provided, and the method includes the following steps:

[0042] S100, obtaining an initial feature vector of the target oral cavity according to an initial image set of the target oral cavity; the initial image set of the target oral cavity includes a plurality of initial images of the target oral cavity taken perpendicular to the tooth surface from different angles, and different initial images of the target oral cavity in the initial image set of the target oral cavity include teeth at different positions in the target oral cavity.

[0043] In this embodiment, the target oral cavity initial image set refers to a set of images taken from different angles perpendicular to the tooth surface for obtaining the target oral cavity initial feature vector, where different images cover teeth at different locations in the target oral cavity. As a specific embodiment, if no orthodontic devices are present on the teeth of the target oral cavity in the target oral cavity initial image set, and information on whether the target oral cavity had undergone oral corrections before uploading the target oral cavity initial image set node is unavailable, then the target oral cavity corresponding to the subject undergoing correction is either a subject who has not undergone oral corrections or a subject who has previously undergone oral corrections but whose orthodontic devices have been removed.

[0044] Those skilled in the art know that any image shooting method in the prior art falls within the scope of protection of the present invention. For example, a dedicated camera is used to shoot from different angles (such as the front, the side) perpendicular to (or approximately perpendicular to) the tooth surface to ensure that each picture clearly covers teeth in different positions (such as the front teeth, the back teeth, the left teeth, and the right teeth) to form an initial set of images of the target oral cavity. Vertical shooting can avoid feature measurement errors caused by perspective deformation, and more accurate oral features can be obtained without constructing a three-dimensional oral model. Optionally, for any picture, its corresponding oral feature value is obtained, and the oral feature values ​​are spliced ​​according to the order of the tooth positions in the oral cavity to obtain the oral feature vector of the target oral cavity. For example, the oral feature vector of the target oral cavity is F=(F1, F2,…, F i ,…,F n ), F i is the subvector corresponding to the i-th oral feature, the value range of i is 1 to n, and n is the number of preset oral features; F i =(F i,1 ,F i,2,…,F i,j ,…,F i,m ), F i,j is the jth value corresponding to the i-th oral feature, where j ranges from 1 to m, and m is the number of values ​​corresponding to the i-th oral feature. For example, if the i-th oral feature is the width of the gap between adjacent teeth, then the number of values ​​corresponding to the i-th oral feature is the number of adjacent tooth pairs. For another example, if the i-th oral feature is the tooth inclination angle, then the number of values ​​corresponding to the i-th oral feature is the number of teeth.

[0045] In this embodiment, the initial features of the target oral cavity refer to characteristic parameters extracted from the initial set of target oral cavity images and capable of reflecting the dental status of the target oral cavity. It should be understood that the initial feature vector of the target oral cavity, i.e., the target oral cavity feature vector obtained based on the initial set of target oral cavity images, is initially used only to distinguish the updated feature vector of the target oral cavity. As a specific embodiment, the oral cavity features include at least one of the following: interdental gap width, tooth alignment, and tooth inclination angle. Interdental gap width refers to the gap distance between two adjacent teeth, typically measured in millimeters, and can be obtained by measuring the shortest straight-line distance between the tooth edge contours. Tooth alignment refers to the regularity of the teeth's arrangement on the dental arch, and can be quantified by calculating the standard deviation of the angle between the tooth's long axis and the dental arch reference line, or the average distance that the tooth position deviates from the dental arch midline. Tooth inclination angle refers to the angle between the tooth's long axis and a reference line perpendicular to the alveolar bone plane, and can be used to assess the degree of tooth inclination. Those skilled in the art will appreciate that the extraction of the above features can be achieved using image processing algorithms, such as edge detection algorithms, curve fitting methods, and Hough transforms, which will not be described in detail here.

[0046] This embodiment acquires comprehensive oral image information through multi-angle shooting, and extracts quantitative oral feature vectors in combination with image processing technology, providing an objective data basis for subsequent matching and analysis.

[0047] S200, matching the initial feature vector of the target oral cavity in the oral cavity feature vector sample library to obtain a matched first oral cavity feature vector and a second oral cavity feature vector; the oral cavity feature vector sample library includes several oral cavity feature vector sets corresponding to sample oral cavity, and the oral cavity feature vector set corresponding to any sample oral cavity includes the oral cavity feature vectors corresponding to the sample oral cavity at different correction times; the first oral cavity feature vector is the oral cavity feature vector corresponding to the sample oral cavity that has not undergone correction processing, and the second oral cavity feature vector is the oral cavity feature vector corresponding to the sample oral cavity that has undergone correction processing.

[0048] In this embodiment, the oral feature vector sample library is pre-constructed, which stores oral feature vector sets corresponding to several sample oral cavity. The feature vector set of each sample oral cavity contains the oral feature vectors corresponding to the sample oral cavity at different correction durations. It should be understood that the oral feature vector corresponding to any sample oral cavity when the correction duration is 0 is the oral feature vector of the sample oral cavity before correction. Any oral feature vector of any sample oral cavity has the same dimension as the initial feature vector of the target oral cavity, and the elements in the same position represent the same physical meaning, for example, they are all the inclination angles of the teeth at the same position.

[0049] As a specific embodiment, Figure 2 As shown, S200 includes:

[0050] S210 , clustering the oral feature vectors in the oral feature vector sample library to obtain a number of oral feature vector clusters.

[0051] As a specific embodiment, each oral feature vector in the oral feature vector sample library is in the form of a vector. Those skilled in the art know that any method for comparing similarities between vectors in the prior art falls within the protection scope of the present invention.

[0052] Those skilled in the art will appreciate that any clustering method in the prior art falls within the scope of protection of the present invention. It should be understood that oral feature vectors within the same cluster have a high similarity, while oral feature vectors between different clusters have a large difference.

[0053] S220 , obtaining the similarity between the target oral cavity initial feature vector and the center of each oral cavity feature vector cluster.

[0054] It should be understood that the center of any oral cavity feature vector cluster is obtained by averaging the oral cavity feature vectors in the oral cavity feature vector cluster.

[0055] In this embodiment, the target oral cavity initial feature vector and the center of any oral cavity feature vector cluster are both in vector form. The similarity between the target oral cavity initial feature vector and the center of each oral cavity feature vector cluster can be obtained using conventional methods for obtaining similarity between vectors. For example, a cosine similarity algorithm can be used. It should be understood that a greater similarity indicates more similar oral cavity features.

[0056] S230, dividing the cluster corresponding to the maximum similarity into a first sub-cluster and a second sub-cluster; the oral feature vectors in the first sub-cluster are the oral feature vectors corresponding to the sample oral cavity that has not been corrected, and the oral feature vectors in the second sub-cluster are the oral feature vectors corresponding to the sample oral cavity that has been corrected.

[0057] In this embodiment, the cluster corresponding to the maximum similarity refers to the cluster whose center has the greatest similarity with the initial feature vector of the target oral cavity, and the oral cavity features corresponding to the feature vectors in this cluster have a greater similarity with the initial feature of the target oral cavity.

[0058] In this embodiment, the oral feature vector corresponding to the sample oral cavity that has not undergone correction processing refers to the oral feature vector whose corresponding correction time is 0, and the oral feature vector corresponding to the sample oral cavity that has undergone correction processing refers to the oral feature vector whose corresponding correction time is greater than 0.

[0059] S240 , determining the oral cavity feature vector in the first subcluster that is most similar to the target oral cavity initial feature vector as the first oral cavity feature vector, and determining the oral cavity feature vector in the second subcluster that is most similar to the target oral cavity initial feature vector as the second oral cavity feature vector.

[0060] Based on S210-S240, this embodiment reduces the amount of calculation during matching through the clustering step and improves the matching efficiency; by first finding the most similar cluster and then dividing the cluster into a first sub-cluster and a second sub-cluster, the oral feature vector corresponding to the uncorrected sample mouth and the oral feature vector corresponding to the corrected sample mouth that are most similar to the target oral initial feature vector in the most similar cluster are obtained, providing a reliable reference for the subsequent formulation of correction targets.

[0061] S300, obtaining the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times based on the matched first oral feature vector and the second oral feature vector; the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained based on the change amount of the oral feature vector corresponding to the first oral feature vector at different correction times and the change amount of the oral feature vector corresponding to the second oral feature vector at different correction times.

[0062] In this embodiment, the target change in the oral cavity feature vector corresponding to any correction duration of the target oral cavity is obtained by weighted summing the change in the oral cavity feature vector corresponding to the first oral cavity feature vector and the change in the oral cavity feature vector corresponding to the second oral cavity feature vector. Those skilled in the art will appreciate that this weighted summation process is conventional and will not be further described here. Optionally, the weights used in the weighted summation process may be empirical values.

[0063] As a specific embodiment, Figure 3 As shown, S300 includes:

[0064] S310 , obtaining oral cavity feature vectors of sample oral cavity corresponding to the first oral cavity feature vector, whose correction time is longer than the correction time corresponding to the first oral cavity feature vector, to form a first oral cavity feature vector subset.

[0065] S320, for any oral feature vector in the first oral feature vector subset, determine the correction time difference between the oral feature vector and the first oral feature vector as the correction time interval corresponding to the oral feature vector, and obtain the oral feature vector change corresponding to the oral feature vector based on the oral feature vector and the first oral feature vector.

[0066] In this embodiment, the oral feature vector change corresponding to any oral feature vector in the first oral feature vector subset is the difference between the oral feature vector and the first oral feature vector. q p=1 (w p ×f p ), f p is the absolute value of the difference between the pth element in the oral feature vector and the pth element in the first oral feature vector, w p is the preset weight of the pth element, 0 <w p <1.

[0067] S330 , obtaining oral cavity feature vectors of the sample oral cavity corresponding to the second oral cavity feature vector, whose correction time is longer than the correction time corresponding to the second oral cavity feature vector, to form a second oral cavity feature vector subset.

[0068] S340, for any oral feature vector in the second oral feature vector subset, determine the correction time difference between the oral feature vector and the second oral feature vector as the correction time interval corresponding to the oral feature vector, and obtain the oral feature vector change corresponding to the oral feature vector based on the oral feature vector and the second oral feature vector.

[0069] In this embodiment, the oral feature vector change corresponding to any oral feature vector in the second oral feature vector subset is the difference between the oral feature vector and the second oral feature vector. The method for obtaining the difference between the oral feature vector and the second oral feature vector is similar to the method for obtaining the difference between the oral feature vector and the first oral feature vector described above, and will not be repeated here.

[0070] S350, obtaining the target change amount of the oral feature vector corresponding to the same correction time interval according to the change amount of the oral feature vector corresponding to the oral feature vector in the first oral feature vector subset and the change amount of the oral feature vector corresponding to the oral feature vector in the second oral feature vector subset corresponding to the same correction time interval.

[0071] As a specific embodiment, a weighted summation is performed on the oral feature vector changes corresponding to the oral feature vectors in the first oral feature vector subset and the oral feature vector changes corresponding to the oral feature vectors in the second oral feature vector subset corresponding to the same correction time interval to obtain the target oral feature vector change corresponding to the same correction time interval. Those skilled in the art will appreciate that the weighted summation process is conventional and will not be further described here; optionally, the weights in the weighted summation are empirical values.

[0072] S360: Determine each identical correction time interval as a correction time interval, and determine the target change amount of the oral cavity feature vector corresponding to each identical correction time interval as the target change amount of the oral cavity feature vector corresponding to the corresponding correction time interval.

[0073] Based on S310-S360, the target changes corresponding to each correction duration are integrated to form the target changes in the oral feature vectors of the target oral cavity at different correction durations. Therefore, this embodiment takes into account two situations: the target oral cavity has not undergone oral correction or has undergone oral correction when uploading the target oral initial image set node. Furthermore, since the oral feature changes corresponding to the same correction duration vary in different situations, this embodiment fuses the correction data of the sample oral cavity corresponding to the first oral feature vector and the second oral feature vector, making the target changes applicable to both situations, providing a more accurate progress reference for the correction process.

[0074] As an optional embodiment, S300 further includes outputting correction parameters of the sample oral cavity corresponding to the first oral cavity feature vector and correction parameters of the sample oral cavity corresponding to the second oral cavity feature vector, to provide professionals with a reference related to the correction parameters. The correction parameters refer to the magnitude and direction of the force applied during the oral correction process.

[0075] S400, obtaining a correction text based on the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector, and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity; the correction text includes a judgment result on whether the target oral cavity is abnormally corrected.

[0076] As a specific embodiment, Figure 4 As shown, S400 includes:

[0077] S410, based on the correction time corresponding to the target oral cavity updated picture set, the target change amount of the oral cavity feature vector corresponding to the specified correction time is selected from the target oral cavity oral cavity target change amounts corresponding to different correction time periods; the specified correction time period is the difference between the upload time of the target oral cavity updated picture set and the upload time of the target oral cavity initial picture set.

[0078] In this embodiment, the updated target oral cavity feature vector has the same dimension as the initial target oral cavity feature vector, and the elements in the same position represent the same physical meaning. It should be understood that the updated target oral cavity feature vector is obtained from the updated target oral cavity image set. If there are orthodontic devices on the teeth of the target oral cavity, the orthodontic devices are first segmented using a segmentation method based on color features, morphological features, or deep learning, and then filled with pixels near the teeth; finally, the oral cavity features are extracted based on the image without the orthodontic devices.

[0079] As a specific implementation method, the correction time corresponding to the target oral update picture set is compared with the correction time corresponding to the target change of each oral feature vector, and the target change of the oral feature vector corresponding to the minimum correction time difference is determined as the target change of the oral feature vector corresponding to the specified correction time.

[0080] S420 , comparing the target oral cavity updated feature vector with the target oral cavity initial feature vector to obtain an oral cavity feature vector change amount corresponding to the target oral cavity updated feature vector.

[0081] As a specific embodiment, the difference between the target oral cavity updated feature vector and the target oral cavity initial feature vector is determined as the change in the oral cavity feature vector corresponding to the target oral cavity updated feature vector. The process of determining the difference between the target oral cavity updated feature vector and the target oral cavity initial feature vector is similar to the method for determining the difference between the oral cavity feature vector and the first oral cavity feature vector described above and is not further described here.

[0082] S430 , judging whether the correction of the target oral cavity is abnormal based on the change amount of the oral cavity feature vector corresponding to the updated feature vector of the target oral cavity and the target change amount of the oral cavity feature vector corresponding to the specified correction time.

[0083] As a specific implementation, S430 includes:

[0084] S431, if the difference between the oral cavity feature vector change amount corresponding to the target oral cavity update feature vector and the oral cavity feature vector target change amount corresponding to the specified correction time is less than or equal to a preset difference threshold, it is determined that the correction of the target oral cavity is normal and the correction text is output.

[0085] In this embodiment, the difference between the change in the oral feature vector corresponding to the target oral update feature vector and the target change in the oral feature vector corresponding to the specified correction time is the absolute value of the difference between the change in the oral feature vector corresponding to the target oral update feature vector and the target change in the oral feature vector corresponding to the specified correction time.

[0086] Optionally, the preset difference threshold is an empirical value.

[0087] S432, if the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the target oral feature vector change amount corresponding to the specified correction time is greater than a preset difference threshold, then determine whether the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the oral feature vector update change amount corresponding to the specified correction time is less than or equal to the preset difference threshold, if it is less than or equal to, then determine that the correction of the target oral cavity is normal, and output the correction text; otherwise, determine that the correction of the target oral cavity is abnormal, and output the correction text; the oral feature vector update change amount corresponding to the specified correction time is the change amount in the first change amount and the second change amount that has a smaller difference with the oral feature vector change amount corresponding to the target oral updated feature vector, the first change amount is the oral feature vector change amount corresponding to the first oral feature vector in the specified correction time, and the second change amount is the oral feature vector change amount corresponding to the second oral feature vector in the specified correction time.

[0088] In this embodiment, taking into account the target change amount of the oral feature vector corresponding to the specified correction time, a comprehensive consideration is given to the two situations that the target oral cavity has not undergone oral correction or has undergone oral correction when uploading the target oral initial picture set node. When the difference between the oral feature vector change amount corresponding to the target oral update feature vector and the target change amount of the oral feature vector corresponding to the specified correction time is greater than a preset difference threshold, the difference between the oral feature vector change amount corresponding to the target oral update feature vector and the oral feature vector update change amount corresponding to the specified correction time is further compared. This can better fit the individual situation of whether the target oral cavity has undergone oral correction or has undergone oral correction when uploading the target oral initial picture set node, which is conducive to improving the accuracy of abnormality judgment.

[0089] This embodiment obtains the initial feature vector of the target oral cavity based on oral images of the target oral cavity from different angles, and matches the initial feature vector of the target oral cavity with the oral feature vectors in the oral feature vector sample library to obtain a first oral feature vector and a second oral feature vector that match the initial feature vector of the target oral cavity. Based on the oral feature vector set of the sample oral cavity corresponding to the first oral feature vector and the second oral feature vector, this embodiment obtains the target change amount of the sample oral feature vector corresponding to different correction time lengths of the target oral cavity; thus, after subsequently obtaining the difference between the updated feature vector of the target oral cavity and the initial feature vector of the target oral cavity, the relationship between the difference and the corresponding target change amount can be used to evaluate whether the correction of the target oral cavity is abnormal. Compared with the existing technology, this embodiment only requires the subject of oral correction to take oral pictures and upload them to obtain the evaluation results, which is more efficient and the process is more convenient. It can quickly monitor the oral correction status and is conducive to timely detection of abnormalities in the oral correction process. Example

[0090] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0091] The target oral cavity initial feature vector is obtained according to the target oral cavity initial picture set; the target oral cavity initial picture set includes several target oral cavity initial pictures taken from different angles perpendicular to the tooth surface, and different target oral cavity initial pictures in the target oral cavity initial picture set include teeth at different positions in the target oral cavity.

[0092] The initial feature vector of the target oral cavity is matched in the oral feature vector sample library to obtain a matched first oral feature vector and a second oral feature vector; the oral feature vector sample library includes several oral feature vector sets corresponding to sample oral cavity, and the oral feature vector set corresponding to any sample oral cavity includes the oral feature vectors corresponding to the sample oral cavity at different correction times; the first oral feature vector is the oral feature vector corresponding to the sample oral cavity that has not undergone correction treatment, and the second oral feature vector is the oral feature vector corresponding to the sample oral cavity that has undergone correction treatment.

[0093] According to the matched first oral feature vector and the second oral feature vector, the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained; the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained based on the change amount of the oral feature vector corresponding to the first oral feature vector at different correction times and the change amount of the oral feature vector corresponding to the second oral feature vector at different correction times.

[0094] The correction text is obtained according to the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity; the correction text includes the judgment result of whether the target oral cavity is corrected abnormally. Example

[0095] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the following steps are implemented:

[0096] The target oral cavity initial feature vector is obtained according to the target oral cavity initial picture set; the target oral cavity initial picture set includes several target oral cavity initial pictures taken from different angles perpendicular to the tooth surface, and different target oral cavity initial pictures in the target oral cavity initial picture set include teeth at different positions in the target oral cavity.

[0097] The initial feature vector of the target oral cavity is matched in the oral feature vector sample library to obtain a matched first oral feature vector and a second oral feature vector; the oral feature vector sample library includes several oral feature vector sets corresponding to sample oral cavity, and the oral feature vector set corresponding to any sample oral cavity includes the oral feature vectors corresponding to the sample oral cavity at different correction times; the first oral feature vector is the oral feature vector corresponding to the sample oral cavity that has not undergone correction treatment, and the second oral feature vector is the oral feature vector corresponding to the sample oral cavity that has undergone correction treatment.

[0098] According to the matched first oral feature vector and the second oral feature vector, the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained; the target change amount of the oral feature vector corresponding to the target oral cavity at different correction times is obtained based on the change amount of the oral feature vector corresponding to the first oral feature vector at different correction times and the change amount of the oral feature vector corresponding to the second oral feature vector at different correction times.

[0099] The correction text is obtained according to the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity; the correction text includes the judgment result of whether the target oral cavity is corrected abnormally.

[0100] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0101] Although some specific embodiments of the present invention have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for obtaining correction text based on oral images, characterized in that: The method comprises the following steps: obtaining an initial feature vector of the target oral cavity according to an initial image set of the target oral cavity; the initial image set of the target oral cavity includes a plurality of initial images of the target oral cavity taken perpendicular to the tooth surface from different angles, and different initial images of the target oral cavity in the initial image set of the target oral cavity include teeth at different positions in the target oral cavity; Matching the initial feature vector of the target oral cavity in an oral cavity feature vector sample library to obtain a matched first oral cavity feature vector and a second oral cavity feature vector; the oral cavity feature vector sample library includes oral cavity feature vector sets corresponding to several sample oral cavity, and the oral cavity feature vector set corresponding to any sample oral cavity includes oral cavity feature vectors corresponding to the sample oral cavity at different correction times; the first oral cavity feature vector is the oral cavity feature vector corresponding to the sample oral cavity that has not undergone correction treatment, and the second oral cavity feature vector is the oral cavity feature vector corresponding to the sample oral cavity that has undergone correction treatment; Obtain target changes in the oral cavity feature vector corresponding to different correction times for the target oral cavity based on the matched first oral cavity feature vector and the second oral cavity feature vector; target changes in the oral cavity feature vector corresponding to different correction times for the target oral cavity are obtained based on changes in the oral cavity feature vector corresponding to different correction times for the first oral cavity feature vector and changes in the oral cavity feature vector corresponding to different correction times for the second oral cavity feature vector; The correction text is obtained according to the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity; the correction text includes the judgment result of whether the target oral cavity is corrected abnormally.

2. The method for obtaining correction text based on oral images according to claim 1, characterized in that: The correction text is obtained based on the correction time corresponding to the target oral cavity update picture set, the target oral cavity update feature vector, and the target change amount of the oral cavity feature vector corresponding to different correction times of the target oral cavity, including: According to the correction time corresponding to the target oral cavity updated picture set, the target oral cavity feature vector target change corresponding to the specified correction time is selected from the target oral cavity feature vector target changes corresponding to different correction time periods; the specified correction time period is the difference between the upload time of the target oral cavity updated picture set and the upload time of the target oral cavity initial picture set; Comparing the target oral cavity updated feature vector with the target oral cavity initial feature vector to obtain the oral cavity feature vector change corresponding to the target oral cavity updated feature vector; It is determined whether the correction of the target oral cavity is abnormal based on the change amount of the oral cavity feature vector corresponding to the target oral cavity update feature vector and the target change amount of the oral cavity feature vector corresponding to the specified correction time.

3. The method for obtaining correction text based on oral images according to claim 2, characterized in that: Judging whether the target oral cavity is corrected abnormally based on the oral cavity feature vector change corresponding to the target oral cavity update feature vector and the oral cavity feature vector target change corresponding to the specified correction time includes: If the difference between the oral cavity feature vector change amount corresponding to the target oral cavity update feature vector and the oral cavity feature vector target change amount corresponding to the specified correction time is less than or equal to a preset difference threshold, it is determined that the correction of the target oral cavity is normal and the correction text is output; If the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the target oral feature vector change amount corresponding to the specified correction time is greater than a preset difference threshold, then determine whether the difference between the oral feature vector change amount corresponding to the target oral updated feature vector and the oral feature vector update change amount corresponding to the specified correction time is less than or equal to the preset difference threshold. If it is less than or equal to, then determine that the correction of the target oral cavity is normal and output the correction text; otherwise, determine that the correction of the target oral cavity is abnormal and output the correction text; the oral feature vector update change amount corresponding to the specified correction time is the change amount in the first change amount and the second change amount that has a smaller difference with the oral feature vector change amount corresponding to the target oral updated feature vector, the first change amount is the oral feature vector change amount corresponding to the first oral feature vector in the specified correction time, and the second change amount is the oral feature vector change amount corresponding to the second oral feature vector in the specified correction time.

4. The method for obtaining correction text based on oral images according to claim 1, characterized in that: Obtaining target changes in the oral cavity feature vector corresponding to different correction times of the target oral cavity according to the matched first oral cavity feature vector and second oral cavity feature vector includes: Obtaining oral cavity feature vectors of sample oral cavity corresponding to the first oral cavity feature vector, whose correction time is longer than the correction time corresponding to the first oral cavity feature vector, to form a subset of the first oral cavity feature vector; For any oral feature vector in the first oral feature vector subset, determine the correction time difference between the oral feature vector and the first oral feature vector as the correction time interval corresponding to the oral feature vector, and obtain the oral feature vector change corresponding to the oral feature vector based on the oral feature vector and the first oral feature vector; Obtain oral cavity feature vectors of the sample oral cavity corresponding to the second oral cavity feature vector, the correction time of which is longer than the correction time corresponding to the second oral cavity feature vector, to form a subset of the second oral cavity feature vector; For any oral feature vector in the second oral feature vector subset, determine the correction time difference between the oral feature vector and the second oral feature vector as the correction time interval corresponding to the oral feature vector, and obtain the oral feature vector change corresponding to the oral feature vector based on the oral feature vector and the second oral feature vector; Obtaining a target change amount of the oral feature vector corresponding to the same correction time interval according to the change amount of the oral feature vector corresponding to the oral feature vector in the first oral feature vector subset and the change amount of the oral feature vector corresponding to the oral feature vector in the second oral feature vector subset corresponding to the same correction time interval; Each identical correction time interval is determined as a correction time, and the target change amount of the oral cavity feature vector corresponding to each identical correction time interval is determined as the target change amount of the oral cavity feature vector corresponding to the corresponding correction time.

5. The method for obtaining correction text based on oral images according to claim 1, characterized in that: Matching the target oral cavity initial feature vector in the oral cavity feature vector sample library to obtain the matched first oral cavity feature vector and second oral cavity feature vector includes: Clustering the oral feature vectors in the oral feature vector sample library to obtain several oral feature vector clusters; Obtaining the similarity between the target oral cavity initial feature vector and the center of each oral cavity feature vector cluster; The cluster corresponding to the maximum similarity is divided into a first sub-cluster and a second sub-cluster; the oral cavity feature vector in the first sub-cluster is the oral cavity feature vector corresponding to the sample oral cavity that has not been corrected, and the oral cavity feature vector in the second sub-cluster is the oral cavity feature vector corresponding to the sample oral cavity that has been corrected; The oral cavity feature vector in the first subcluster that is most similar to the target oral cavity initial feature vector is determined as the first oral cavity feature vector, and the oral cavity feature vector in the second subcluster that is most similar to the target oral cavity initial feature vector is determined as the second oral cavity feature vector.

6. The method for obtaining correction text based on oral images according to claim 1, characterized in that: Oral characteristics include the following: At least one of the following: the width of the space between adjacent teeth, the neatness of tooth arrangement and the angle of tooth inclination.

7. The method for obtaining correction text based on oral images according to claim 3, characterized in that: The difference between the oral feature vector change amount corresponding to the target oral update feature vector and the target oral feature vector change amount corresponding to the specified correction time is the absolute value of the difference between the oral feature vector change amount corresponding to the target oral update feature vector and the target oral feature vector change amount corresponding to the specified correction time.

8. The method for obtaining correction text based on oral images according to claim 1, characterized in that: The target change amount of the oral cavity feature vector corresponding to any correction time of the target oral cavity is obtained by weighted summing the change amount of the oral cavity feature vector corresponding to the correction time of the first oral cavity feature vector and the change amount of the oral cavity feature vector corresponding to the correction time of the second oral cavity feature vector.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for obtaining corrected text based on oral images as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for obtaining corrected text based on an oral image as described in any one of claims 1 to 8 is implemented.

Images