Method and apparatus for processing tooth images

Abstract

The present disclosure provides a method and apparatus for processing dental images, which relates to the field of artificial intelligence technologies such as computer vision, deep learning, and model training. The specific implementation manners include: obtaining an image containing teeth; inputting the image into a tooth position determination network, and through the tooth position determination network, determining the probabilities of the position category distances of the teeth in the image reaching each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different. The present disclosure determines the position category of the teeth by determining the results obtained by the teeth in the image reaching the position category distances, so that a neural network that has learned the tooth position relationship can perform tooth position detection, thereby improving the accuracy of tooth position detection by utilizing the position relationship between teeth of different position categories.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, specifically to artificial intelligence technologies including computer vision, deep learning, and model training, and particularly to a method and apparatus for processing tooth images. Background Art

[0002] With the development of image processing technologies, more and more technical problems are solved through image processing. Image processing technologies generally input an image into a model for detection or recognition and obtain a detection result or recognition result output from the model.

[0003] The oral cavity includes various types of teeth, such as incisors, wisdom teeth, etc. During oral diagnosis, detecting the tooth position is of great significance for detecting and analyzing oral diseases. Usually, detecting the tooth position requires medical staff to judge and mark the tooth positions one by one based on panoramic films, which is rather cumbersome and time-consuming. In related technologies, the tooth position can be detected through a model. Summary of the Invention

[0004] A method, apparatus, electronic device, and storage medium for processing tooth images are provided.

[0005] According to a first aspect, a method for processing a tooth image is provided, including: obtaining an image including teeth; inputting the image into a tooth position determination network, and through the tooth position determination network, determining the probability that the position category distance of the teeth in the image reaches the position category distances of each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different.

[0006] According to a second aspect, a device for processing a tooth image is provided, including: an obtaining unit configured to obtain an image including teeth; a prediction unit configured to input the image into a tooth position determination network, and through the tooth position determination network, determine the probability that the position category distance of the teeth in the image reaches the position category distances of each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different.

[0007] According to a third aspect, an electronic device is provided, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method of any one of the embodiments of the method for processing a tooth image.

[0008] According to a fourth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, and the computer instructions are used to cause a computer to execute the method of any one of the embodiments of the method for processing a tooth image.

[0009] According to a fifth aspect, a computer program product is provided, including a computer program which, when executed by a processor, implements the method according to any one of the embodiments of the method for processing dental images.

[0010] According to the solution of the present disclosure, by determining the result obtained by the teeth in the image reaching the position category distance, the position category of the teeth can be determined, so that a neural network that has learned the tooth position relationship can perform tooth position detection, and thus the position relationship between teeth of different position categories can be utilized to improve the accuracy of tooth position detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects, and advantages of the present disclosure will become more apparent by reading the detailed description of the non-limiting embodiments with reference to the following drawings:

[0012] Figure 1 is an exemplary system architecture diagram to which some embodiments of the present disclosure can be applied;

[0013] Figure 2 is a flowchart of an embodiment of the method for processing dental images according to the present disclosure;

[0014] Figure 3 is a schematic diagram of an application scenario of the method for processing dental images according to the present disclosure;

[0015] Figure 4a is a flowchart of another embodiment of the method for processing dental images according to the present disclosure;

[0016] Figure 4b is a flowchart of another embodiment of the method for processing dental images according to the present disclosure;

[0017] Figure 5 is a schematic structural diagram of an embodiment of the apparatus for processing dental images according to the present disclosure;

[0018] Figure 6 is a block diagram of an electronic device for implementing the method for processing dental images of the embodiments of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to assist in understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0020] In the technical solution of the present disclosure, the acquisition, storage, and application of the user's personal information comply with the provisions of relevant laws and regulations, adopt necessary confidentiality measures, and do not violate public order and good customs.

[0021] It should be noted that, without conflict, the embodiments in the present disclosure and the features in the embodiments may be combined with each other. The present disclosure will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0022] Figure 1 An exemplary system architecture 100 is shown, which can apply the embodiments of the method for processing dental images or the apparatus for processing dental images of the present disclosure.

[0023] As Figure 1 shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0024] Users can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the terminal devices 101, 102, 103, such as video applications, live broadcast applications, instant messaging tools, email clients, social platform software, etc.

[0025] The terminal devices 101, 102, 103 here may be hardware or software. When the terminal devices 101, 102, 103 are hardware, they may be various electronic devices with a display screen, including but not limited to smart phones, tablet computers, e-book readers, laptop portable computers, and desktop computers, etc. When the terminal devices 101, 102, 103 are software, they may be installed in the above-listed electronic devices. It may be implemented as multiple software or software modules (such as multiple software or software modules for providing distributed services), or it may be implemented as a single software or software module. No specific limitation is made here.

[0026] The server 105 may be a server that provides various services, such as a background server that provides support for the terminal devices 101, 102, 103. The background server may analyze and process data such as dental images received, and feedback the processing results (such as the probability of the position category distance of the teeth in the image reaching the position category distance of each position category) to the terminal devices.

[0027] It should be noted that the method for processing dental images provided by the embodiments of the present disclosure can be executed by the server 105 or the terminal devices 101, 102, and 103. Correspondingly, the device for processing dental images can be disposed in the server 105 or the terminal devices 101, 102, and 103.

[0028] It should be understood that Figure 1 the numbers of terminal devices, networks, and servers in

[0029] Continuing to refer to Figure 2 , a flowchart 200 of an embodiment of the method for processing dental images according to the present disclosure is shown. The method for processing dental images includes the following steps:

[0030] Step 201, obtain an image including teeth.

[0031] In this embodiment, the execution entity (such as Figure 1 the server or terminal device shown) on which the method for processing dental images runs can obtain an image presenting teeth from this device or other electronic devices. In practice, the image here can be a pre-processed image.

[0032] Step 202, input the image into the tooth position determination network, and through the tooth position determination network, determine the probabilities of the position category distances of the teeth in the image reaching the position category distances of each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different.

[0033] In this embodiment, the above execution entity can input the above image into the tooth position determination network, and through this network, determine the probabilities of the position category distances of the teeth in the image reaching the position category distances of each position category. Specifically, for each position category among multiple position categories, the probability of reaching the position category distance of that position category can be determined.

[0034] The distance between the position category and the incisor position category is the position category distance of that position category. The position categories can include incisors, canines (i.e., premolars), molars, and wisdom teeth. The position category distance can be a preset distance or an actual distance. For example, in the case of a preset distance, the position category distance of incisors can be 0, and the position category distance of canines can be 1.

[0035] The number of determined probabilities can be at least two. The at least two probabilities correspond one by one to respective position categories, and these probabilities can respectively represent the probabilities that the position category of the teeth in the image reaches the position category distance of each position category in the respective position categories. In some cases, the determined probability can also be only one, representing the probability that the position category distance of the teeth in the image reaches the position category distances of all position categories in the respective position categories.

[0036] In this embodiment, the position category of the teeth in the image, that is, the tooth position, can be indicated by probabilities. Specifically, the above-mentioned execution subject can determine by whether the position category distance of the teeth in the image reaches each position category distance. Once the position category distance of a certain position category is reached, it indicates that the teeth in the image are of this position category or a position category after this position category. Here, "after" means farther from the front teeth.

[0037] The tooth position determination network is a deep neural network. For example, the tooth position determination network can be a specified residual network, such as ResNet (e.g., ResNet34), and can also be a densely connected convolutional neural network DenseNet. In addition, it can also be an EfficientNet.

[0038] The method provided by the above embodiment of the present disclosure can determine the position category of the teeth by the result obtained by determining that the teeth in the image reach the position category distance, and can enable the neural network that has learned the tooth position relationship to perform tooth position detection, so as to improve the accuracy of tooth position detection by using the position relationship between teeth of different position categories.

[0039] This application also provides a method for processing a tooth image, including: acquiring an image containing teeth; inputting the image into a tooth position determination network, and determining, through the tooth position determination network, whether the position category distance of the teeth in the image reaches the position category distances of each position category to obtain probabilities, where the position category distance refers to the distance between the position category and the front tooth position category, and the position category distances of different position categories are different; and determining an indication value of the position category of the teeth in the image based on the probabilities.

[0040] In this embodiment, the above-mentioned execution subject can determine the indication value of the position category of the teeth in the image based on the probabilities in various ways. For example, the above-mentioned execution subject can obtain the correspondence between the probabilities and the indication values of the position categories, and obtain the indication value of the position category corresponding to the probability according to this correspondence.

[0041] The indication value is a value that indicates the position category. For example, the indication value of the position category of the front teeth can be 0 or a, and correspondingly, the indication value of the position category of the canine teeth can be 1 or b.

[0042] The process of determining the indication value can be carried out within the tooth position determination network. For example, the indication value can be the output of the tooth position determination network. Alternatively, the indication value can also be determined using the probabilities output by the tooth position determination network.

[0043] This embodiment can further process the result of tooth position classification, making the tooth position classification result more readable to people through the indication value.

[0044] In some alternative implementation manners of this embodiment, the indication value includes at least two indication values, and the at least two indication values are respectively used to indicate whether the position category distances of the teeth in the image reach each of the at least two position category distances, where the at least two position category distances include the position category distances of position categories other than the incisor position category.

[0045] In these implementation manners, the indication value can include at least two. For example, the respective indication values can be arranged according to the respective position categories, for example, arranged from the indication value of the position category of the canine tooth to the indication value of the position category of the wisdom tooth. For example, in the order of the indication values from left to right, the indication value represents the position category distances from near to far from the incisor. Each indication value corresponds to one of the at least two position category distances. Each indication value is at one value position. There may be no indication value directly indicating the incisor position category among the at least two indication values.

[0046] For example, the indication value can be [0, 0, 0], and this indication value indicates the incisor position category. The three 0s therein indicate that the position category distances of the teeth in the image do not reach the position category distance of the canine tooth, do not reach the position category distance of the molar, and do not reach the position category distance of the wisdom tooth. Again, for example, the indication value can be [1, 1, 0], and this indication value indicates the molar position category. The first 1 counted from the left indicates that the position category distance of the teeth in the image reaches the position category distance of the canine tooth, and the second 1 indicates that the position category distance of the teeth in the image reaches the position category distance of the molar. The 0 therein indicates that the position category distance of the teeth in the image does not reach the position category distance of the wisdom tooth.

[0047] The indication value in these implementation manners does not adopt the one-hot encoding method, but provides a new representation method of the position category through the indication value. Whether the position category distance is reached is used to represent the position category of the teeth. And by omitting the incisor position category, the character length of the indication value is reduced, and the indication value is streamlined.

[0048] Continue to refer to Figure 3 , Figure 3 is a schematic diagram of an application scenario of the method for processing a tooth image according to this embodiment. In Figure 3In the application scenario, the execution entity 301 obtains an image 302 including teeth. The execution entity 301 inputs the image 302 into the tooth position determination network 303, and through the tooth position determination network 303, determines the probability 304 that the position category distance of the teeth in the image reaches the position category distances of each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different.

[0049] In some alternative implementation manners of any embodiment of the present disclosure, based on the probability, determining an indication value of the position category of the teeth in the image includes: comparing the magnitude relationship between each probability corresponding to at least two position category distances and a probability threshold; for each probability among the probabilities, generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold, where, for the indication value corresponding to the largest position category distance among the indication values indicating reaching the probability threshold, the position category indicated by the position category distance corresponding to the indication value is the position category of the teeth in the image.

[0050] In these alternative implementation manners, the probability can be a probability. Each position category distance in at least two position categories corresponds to one probability among the probabilities. The above-mentioned execution entity can compare each probability with the probability threshold. Each position category distance among the position category distances has a probability corresponding to the position category distance. The probability corresponding to the position category distance refers to the probability of determining whether the position category distance of the teeth in the image reaches the position category distance. The greater the probability, that is, the closer to 1, the greater the possibility that the position category distance of the teeth in the image reaches the position category distances of each position category.

[0051] In practice, the probability can include the probabilities corresponding to position categories other than the incisor position category. For example, the probabilities corresponding to the canine position category, the molar position category, and the wisdom tooth position category are 0.7, 0.2, and 0.3 respectively, and the probability threshold is 0.5. The probability 0.7 is greater than the probability threshold 0.5, that is, it reaches the probability threshold, the probability 0.2 is less than the probability threshold, that is, it does not reach the probability threshold, and the probability 0.3 is less than the probability threshold, that is, it does not reach the probability threshold.

[0052] The above-mentioned execution entity can adopt various methods to generate an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold. For example, the above-mentioned execution entity can input the result of whether the probability reaches the probability threshold into a preset formula or model, and use the output of the preset formula or model as the indication value.

[0053] Among the indication values that reach the probability threshold, the indication value with the largest corresponding position category distance, and the position category indicated by this indication value is the position category of the teeth in the image.

[0054] These implementation methods can generate accurate indication values by comparing the probability that the position category distance of the teeth in the image reaches the position category distance of each position category with the probability threshold.

[0055] In some optional application scenarios of these implementation methods, the tooth position determination network is a conditional probability network, and the position category distance is equal to the value of the position category to which it belongs. The closer it is to the incisor position category, the smaller the value of the position category; the conditional probability of the conditional probability network refers to using whether each probability corresponding to each position category distance reaches the probability threshold as a judgment condition to judge the probability that the position category distance of the teeth in the image reaches the value of each position category.

[0056] In these application scenarios, both the position category distance and the position category can be represented by numerical values. Each position category distance belongs to a position category, that is, it corresponds to this position category. The position category distance is equal to the value of the position category to which the position category distance belongs. For example, the position category distances of incisors, canines, molars, and wisdom teeth are equal to 0, 1, 2, and 3 respectively. Correspondingly, the values of the position categories of incisors, canines, molars, and wisdom teeth can also be equal to 0, 1, 2, and 3 respectively.

[0057] The conditional probability in the conditional probability network can refer to using whether the probability can reach the probability threshold as a judgment condition to judge the probability P that the position category distance y of the teeth in the image reaches the value k of the position category.

[0058] The probability P of the k-th element in the probability P k can be expressed by the following formula:

[0059] P k = Pr(y≥k|x), k = 1, 2…, C - 1

[0060] where y≥k|x means that given x, the position category distance y of the teeth in the image reaches the value k of the position category, and the probability of reaching here is P k . Pr indicates P k is a function representing the relationship between P, y, and k. There are C categories of the position categories of the teeth.

[0061] In some application scenarios, there can be the following corresponding relationship, representing the corresponding relationship between the following two: the function value under each position category and the probability that the position category distance of the position category of the teeth in the image reaches the value of this position category:

[0062] Function value Pr(y≥1|x) Pr(y≥2|x) Pr(y≥3|x) Probability 0.7 0.2 0.3

[0063] These application scenarios can enhance the understanding of the meaning during network learning and improve the interpretability of the network through a conditional probability network.

[0064] In some alternative application scenarios of these implementation manners, comparing the magnitudes of the respective probabilities with the probability threshold may include: comparing the magnitudes of the respective probabilities with the probability threshold in ascending order of the position category distances corresponding to the probabilities; and for each of the respective probabilities, generating an indication value indicating whether the position category distance of the tooth in the image reaches the position category distance corresponding to the probability based on whether the probability reaches the probability threshold, including: for the currently compared probability, in response to the comparison result that the probability reaches the probability threshold, determining a first indication value indicating that the position category distance of the tooth in the image reaches the position category distance corresponding to the probability.

[0065] In these application scenarios, the above-mentioned execution entity may compare each of the respective probabilities with the probability threshold in ascending order of the position category distances corresponding to the probabilities. Specifically, if the probability reaches the probability threshold, a first indication value indicating that the position category distance of the tooth in the image reaches the position category distance corresponding to the probability, such as 1, is determined.

[0066] These application scenarios can determine the first indication value when the comparison result is that the probability corresponding to the position category distance reaches the threshold, so as to accurately determine that the position category distance of the tooth in the image is at least not less than the position category distance.

[0067] Optionally, for each of the respective probabilities, generating an indication value indicating whether the position category distance of the tooth in the image reaches the position category distance corresponding to the probability based on whether the probability reaches the probability threshold may further include: for the currently compared probability, in response to the comparison result that the probability does not reach the probability threshold, stopping the comparison and determining a second indication value for the probability whose corresponding position category distance is greater than the probability, where the second indication value indicates that the position category distance of the tooth in the image does not reach the position category distance corresponding to the probability.

[0068] Specifically, during the comparison, it is not necessary to traverse and compare all the probabilities with the probability threshold. Once it is compared that the probability does not reach the probability threshold, the above-mentioned execution entity can stop the comparison. The above-mentioned execution entity can determine the second indication value for the probability whose comparison result is not reached and the probabilities that have not been compared with the probability threshold. Finally, the indication value may include the first indication value and / or the second indication value. The position category distance corresponding to the probability that has not been compared with the probability threshold is greater than the position category distance of the above-mentioned probability.

[0069] These implementation methods can directly determine the comparison result after finding the probability that does not reach the probability threshold, shortening the comparison process and time consumption.

[0070] Further reference is made to Figure 4a , which shows the process 400 of an embodiment of the training method of the tooth position determination network. The tooth position determination network that has completed training is any one of the above networks. The process 400 includes the following steps:

[0071] Step 401, obtain training samples, where the training samples include images containing teeth, and the probabilities of the position category distances of the teeth reaching each position category.

[0072] In this embodiment, the execution entity (such as Figure 1 the server or terminal device shown) on which the processing method of the tooth image runs can obtain training samples. The training samples include images that contain teeth. In addition, the training samples also include the above probabilities. In some cases, the probabilities in the samples can include, for example, 1 or 0.

[0073] Step 402, use the image as the input and the probability as the target output to train the tooth position determination network to be trained, and obtain the tooth position determination network that has completed training.

[0074] In this embodiment, the above execution entity can use the image as the input and the probability as the target output to train the tooth position determination network to be trained. Specifically, the above execution entity can first input the image into the tooth position determination network to be trained for forward propagation to obtain the probability output by the tooth position determination network to be trained. Then, the above execution entity can use this probability, the probability in the training samples, and a preset loss function to determine the loss value, and use the loss value to train the tooth position determination network to be trained.

[0075] This embodiment can train a model that outputs a prediction result indicating the position category distance of the tooth. Moreover, in this embodiment, by determining the result obtained when the tooth in the image reaches the position category distance, the position category of the tooth can be determined, enabling the neural network that has learned the tooth position relationship to perform tooth position detection, thereby improving the accuracy of tooth position detection by the trained network by utilizing the position relationship between teeth of different position categories.

[0076] In some optional implementation methods of this embodiment, training the tooth position determination network to be trained to obtain the tooth position determination network that has completed training may include: training the tooth position determination network to be trained to obtain the trained tooth position determination network; using the balanced mean square error loss function to update the trained tooth position determination network to obtain the tooth position determination network that has completed training.

[0077] Among these alternative implementations, the loss function can be the Balance Mean Squared Error (Balance MSE) loss function. The results of a full oral scan generally include 1 incisor image, 4 canine images, and 4 molar images. Since the incidence rate of wisdom teeth is 5%, generally, there are 5 wisdom tooth images in every 100 tooth images. Therefore, from the perspective of the tooth image distribution, it can be seen that the data volume ratio of wisdom teeth is very small, and thus the various types of samples are very unbalanced.

[0078] To solve the long-tail problem of wisdom tooth classification, in this embodiment, the Balance Mean Squared Error loss function is used to determine the loss value. This method can balance the long-tail problem, thereby solving the problem of unbalanced various types of samples.

[0079] Optionally, the generation step of at least one training sample includes: for each original sample among at least one original sample containing teeth, perform an elimination process on at least one tooth in the original sample to obtain a toothless image of the original sample; use the toothless images of at least one original sample containing teeth as at least one training sample.

[0080] Specifically, the above-mentioned execution entity can perform instance annotation to implement the elimination process on at least one tooth in the sample image to obtain a toothless sample image.

[0081] As Figure 4b shown, the figure shows the original sample (on the left) and the toothless image (on the right).

[0082] These implementations can obtain toothless images and perform sample enhancement through the toothless images, so as to ensure that the model still has a high prediction accuracy when there are missing teeth in the image.

[0083] Further referring to Figure 5 , as an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a device for processing tooth images. This device embodiment corresponds to the method embodiment shown in Figure 2 . Except for the features described below, this device embodiment may further include the same or corresponding features or effects as the method embodiment shown in Figure 2 . This device can be specifically applied to various electronic devices.

[0084] As Figure 5As shown, the tooth image processing device 500 of this embodiment includes: an acquisition unit 501 and a prediction unit 502. The acquisition unit 501 is configured to acquire an image containing teeth; the prediction unit 502 is configured to input the image into a tooth position determination network, and through the tooth position determination network, determine the probability that the position category distance of the teeth in the image reaches the position category distance of each position category, wherein the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different.

[0085] In this embodiment, the specific processing of the acquisition unit 501 and the prediction unit 502 of the tooth image processing device 500 and the technical effects thereof can be referred to in Figure 2 The relevant descriptions of step 201 and step 202 in the corresponding embodiment are not repeated here.

[0086] In some optional implementations of this embodiment, the device further includes: a category determination unit configured to determine an indication value of a position category of teeth in the image based on probability.

[0087] In some optional implementations of this embodiment, the indication value includes at least two indication values, and the at least two indication values ​​are respectively used to indicate whether the position category distance of the teeth in the image reaches each position category distance of at least two position category distances, wherein the at least two position category distances include position category distances of position categories other than the incisor position category.

[0088] In some optional implementations of the present embodiment, in some optional implementations of the present embodiment, the category determination unit is further configured to perform probability-based determination of an indication value of the position category of teeth in the image in the following manner: for each probability corresponding to at least two position category distances, compare the size relationship between each probability and a probability threshold; for each of the probabilities, according to whether the probability reaches the probability threshold, generate an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability, wherein, for the indication value having the largest position category distance among the indication values ​​indicating reaching the probability threshold, the position category indicated by the position category distance corresponding to the indication value is the position category of the teeth in the image.

[0089] In some optional implementations of the present embodiment, the tooth position determination network is a conditional probability network, the position category distance is equal to the numerical value of the position category to which it belongs, and the closer to the incisor position category, the smaller the numerical value of the position category; the conditional probability network indicates whether each probability corresponding to each position category distance reaches a probability threshold as a judgment condition, and judges the probability that the position category distance of the teeth in the image reaches the numerical value of each position category.

[0090] In some alternative implementation manners of this embodiment, the category determination unit is further configured to compare the magnitudes of each probability and the probability threshold in the following manner: compare the magnitudes of each probability and the probability threshold in ascending order of the position category distance corresponding to the probability; and the category determination unit is further configured to generate an indication value indicating whether the position category distance of the tooth in the image reaches the position category distance corresponding to the probability for each probability among the probabilities in the following manner: for the currently compared probability, if the comparison result is that the probability reaches the probability threshold, then determine a first indication value indicating that the position category distance of the tooth in the image reaches the position category distance corresponding to the probability.

[0091] In some alternative implementation manners of this embodiment, the category determination unit is further configured to generate an indication value indicating whether the position category distance of the tooth in the image reaches the position category distance corresponding to the probability for each probability among the probabilities in the following manner: for the currently compared probability, if the comparison result is that the probability does not reach the probability threshold, then stop the comparison, and determine a second indication value for the probability and the probabilities whose corresponding position category distances are greater than the probability, where the second indication value indicates that the position category distance of the tooth in the image does not reach the position category distance corresponding to the probability.

[0092] The present disclosure further provides a training device for a tooth position determination network. The trained tooth position determination network is any of the above networks. The device may include: a sample acquisition unit configured to acquire training samples, where the training samples include images containing teeth and the probabilities that the position category distances of the teeth reach the position category distances of each position category; a training unit configured to train the to-be-trained tooth position determination network with the image as the input and the probability as the target output to obtain the trained tooth position determination network.

[0093] In some alternative implementation manners of this embodiment, the training unit is further configured to train the to-be-trained tooth position determination network in the following manner to obtain the trained tooth position determination network: train the to-be-trained tooth position determination network to obtain a post-training tooth position determination network; and update the post-training tooth position determination network using the balanced mean square error loss function to obtain the trained tooth position determination network.

[0094] In some alternative implementation manners of this embodiment, the generation step of at least one training sample includes: for each original sample among at least one original sample containing teeth, perform an elimination process on at least one tooth in the original sample to obtain a toothless image of the original sample; and use the toothless images of at least one original sample containing teeth as at least one training sample.

[0095] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0096] Figure 6 A schematic block diagram of an exemplary electronic device 600 that can be used to implement the embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as, for example, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, for example, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely exemplary and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0097] As Figure 6 shown, the device 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0098] Multiple components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a magnetic disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0099] The computing unit 601 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 executes the various methods and processes described above, such as the method for processing dental images. For example, in some embodiments, the method for processing dental images can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the method for processing dental images described above can be executed. Alternatively, in other embodiments, the computing unit 601 can be configured to execute the method for processing dental images in any other suitable manner (e.g., by means of firmware).

[0100] Various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-a-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0101] The program code for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program codes can be executed entirely on the machine, partially on the machine, as an independent software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0102] In the context of this disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0103] For purposes of providing an interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic, speech, or tactile input).

[0104] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of a communication network include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0105] A computer system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

[0106] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and no limitations are imposed herein.

[0107] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for processing a tooth image, the method comprising: Obtaining an image containing teeth; Inputting the image into a tooth position determination network, and through the tooth position determination network, determining the probabilities of the position category distances of the teeth in the image reaching the position category distances of each position category, wherein the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different; Wherein, it further includes: For each probability corresponding to at least two position category distances respectively, comparing the magnitude relationship between each probability and a probability threshold; For each probability among the various probabilities, generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold, wherein, for the indication value corresponding to the largest position category distance among the indication values indicating reaching the probability threshold, the position category indicated by the position category distance corresponding to the indication value is the position category of the teeth in the image.

2. The method according to claim 1, wherein, The indication value includes at least two indication values, and the at least two indication values are respectively used to indicate whether the position category distance of the teeth in the image reaches each position category distance among at least two position category distances, wherein the at least two position category distances include the position category distances of position categories other than the incisor position category.

3. The method according to any one of claims 1-2, wherein The tooth position determination network is a conditional probability network, the position category distance is equal to the value of the position category to which it belongs, and the closer the position category is to the incisor position category, the smaller the value of the position category; The conditional probability network indicates using whether each probability corresponding to each position category distance reaches the probability threshold as a judgment condition to judge the probability that the position category distance of the teeth in the image reaches the value of each position category.

4. The method according to claim 1, wherein, The comparing the magnitude relationship between each probability and the probability threshold includes: Comparing the magnitude relationship between each probability and the probability threshold in ascending order of the position category distance corresponding to the probability; and The generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability for each probability among the various probabilities includes: For the currently compared probability, in response to the comparison result being that the probability reaches the probability threshold, determining a first indication value indicating that the position category distance of the teeth in the image reaches the position category distance corresponding to the probability.

5. The method according to claim 4, wherein The generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability for each probability among the various probabilities further includes: For the currently compared probability, in response to the comparison result being that the probability does not reach the probability threshold, stopping the comparison, and determining a second indication value for the probability and the probabilities corresponding to position category distances greater than the probability, wherein the second indication value indicates that the position category distance of the teeth in the image does not reach the position category distance corresponding to the probability.

6. A training method for a tooth position determination network, where the trained tooth position determination network is the network described in any one of claims 1-5. The method includes: Obtaining training samples, where the training samples include images containing teeth, and the probabilities of the position category distances of the teeth reaching each position category; Using the images as inputs and the probabilities as target outputs to train the tooth position determination network to be trained, obtaining a trained tooth position determination network.

7. The method according to claim 6, wherein, The training of the tooth position determination network to be trained to obtain a trained tooth position determination network includes: Training the tooth position determination network to be trained to obtain a trained tooth position determination network; Using a balanced mean squared error loss function to update the trained tooth position determination network to obtain a trained tooth position determination network.

8. The method according to claim 6 or 7, wherein The generation step of at least one training sample includes: For each of at least one original sample containing teeth, performing an elimination process on at least one tooth in the original sample to obtain an edentulous image of the original sample; Using the edentulous images of at least one original sample containing teeth as the at least one training sample.

9. A processing device for tooth images, the device includes: An acquisition unit configured to acquire images containing teeth; A prediction unit configured to input the images into a tooth position determination network, and through the tooth position determination network, determine the probabilities of the position category distances of the teeth in the images reaching each position category, where the position category distance refers to the distance between the position category and the incisor position category, and the position category distances of different position categories are different; Wherein, the device further includes a category determination unit, and the category determination unit is configured to: Compare the magnitudes of each of the probabilities corresponding to at least two position category distances with a probability threshold; For each of the probabilities, generate an indication value indicating whether the position category distance of the tooth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold. Among the indication values indicating reaching the probability threshold, for the indication value corresponding to the largest position category distance, the position category indicated by the position category distance corresponding to the indication value is the position category of the tooth in the image.

10. The apparatus according to claim 9, wherein, The indication value includes at least two indication values, and the at least two indication values are respectively used to indicate whether the position category distance of the tooth in the image reaches each of at least two position category distances, where the at least two position category distances include the position category distances of position categories other than the incisor position category.

11. The device according to one of claims 9 - 10, wherein, The tooth position determination network is a conditional probability network, and the position category distance is equal to the value of the position category to which it belongs. The closer the position category is to the incisor position category, the smaller the value of the position category; The conditional probability network indicates using whether each of the probabilities corresponding to the respective position category distances reaches the probability threshold as a judgment condition to judge the probabilities of the position category distances of the teeth in the image reaching the values of each position category.

12. The apparatus according to claim 9, wherein, The category determination unit is further configured to perform the comparison of the respective probabilities with the probability threshold in the following manner: The category determination unit is further configured to perform, for each of the respective probabilities, generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold in the following manner: For the currently compared probability, in response to the result of the comparison being that the probability reaches the probability threshold, determine a first indication value indicating that the position category distance of the teeth in the image reaches the position category distance corresponding to the probability.

13. The device according to claim 12, wherein, The category determination unit is further configured to perform, for each of the respective probabilities, generating an indication value indicating whether the position category distance of the teeth in the image reaches the position category distance corresponding to the probability according to whether the probability reaches the probability threshold in the following manner: For the currently compared probability, in response to the result of the comparison being that the probability does not reach the probability threshold, stop the comparison, and determine a second indication value for the probability and the probabilities corresponding to position category distances greater than the probability, where the second indication value indicates that the position category distance of the teeth in the image does not reach the position category distance corresponding to the probability.

14. A training device for a tooth position determination network, the trained tooth position determination network being the network according to any one of claims 9-13, the device comprising: A sample acquisition unit configured to acquire training samples, where the training samples include images containing teeth and the probabilities that the position category distances of the teeth reach the position category distances of respective position categories; A training unit configured to train the to-be-trained tooth position determination network with the image as the input and the probability as the target output to obtain a trained tooth position determination network.

15. The apparatus according to claim 14, wherein, The training unit is further configured to perform the training of the to-be-trained tooth position determination network to obtain a trained tooth position determination network in the following manner: Train the to-be-trained tooth position determination network to obtain a post-training tooth position determination network; Update the post-training tooth position determination network using a balanced mean square error loss function to obtain a trained tooth position determination network.

16. The device according to claim 14 or 15, wherein, The generation step of at least one training sample includes: For each of at least one original sample containing teeth, perform an elimination process on at least one tooth in the original sample to obtain a toothless image of the original sample; Use the toothless images of at least one original sample containing teeth as the at least one training sample.

17. An electronic device, comprising: At least one processor; And A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1-8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-8.

19. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 - 8.

Images