Model training method and apparatus, and image recognition method and apparatus
By constructing an initial recognition model based on multiple feature extraction networks and integrated learning classifiers, the problem of poor overfitting and generalization ability in image feature category recognition of children's lung disease is solved, and the recognition accuracy and robustness are improved, and the recognition accuracy and robustness are achieved. Efficient recognition of childhood lung disease images.
Patent Information
- Application Number
- PCT/CN2024/088527
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-10
- Filing Date
- 2024-04-18
- Publication Date
- 2024-11-14
Smart Images

Figure CN2024088527_14112024_PF_FP_ABST
Abstract
Description
Model training method, image recognition method and device
[0001] This patent application claims priority to the Chinese patent application filed on May 10, 2023, with application number 202310524432.1 and invention name “Model training method, image recognition method and device”, the full text of which is incorporated by reference into this application. Technical Field
[0002] The present disclosure relates to the field of artificial intelligence technology, specifically to the field of medical and health image processing technology, and more particularly to a model training method, image recognition method and device. Background Art
[0003] Currently, there are two main approaches to automatic classification of pediatric lung disease image features based on deep learning: 1) Designing a corresponding deep neural network and directly training the network using a small-scale pediatric lung disease dataset to determine the classification of lung disease image features; 2) Using transfer learning, a classical neural network is pre-trained on the public dataset ImageNet and then trained on chest X-rays to achieve final prediction. However, the neural network in approach 1) is prone to overfitting, poor generalization, and low precision. In approach 2), the images in ImageNet are all real-world images, which can lead to low accuracy of the trained neural network in identifying lung disease image feature categories.
[0004] Summary of the Invention
[0005] The embodiments of the present disclosure provide a model training method, apparatus, device, and storage medium.
[0006] An embodiment of the present disclosure provides a model training method, which includes: obtaining a training sample set, constructing an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier, and training the initial recognition model using the training sample set to obtain a recognition model.
[0007] Another embodiment of the present disclosure provides an image recognition method, which includes: obtaining a first type of image including body parts; inputting the first type of image into a recognition model to obtain a disease image feature category corresponding to the body part, wherein the recognition model is a recognition model obtained by the method described in any implementation method of the first aspect above.
[0008] Another embodiment of the present disclosure provides a model training device, which includes: an acquisition module configured to acquire a training sample set, a construction module configured to construct an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier; and a training module configured to train the initial recognition model using the training sample set to obtain a recognition model.
[0009] Another embodiment of the present disclosure provides an image recognition device, which includes: an acquisition module, configured to obtain a first type of image including body parts; a recognition module, configured to input the first type of image into a recognition model to obtain a disease image feature category corresponding to the body part, wherein the recognition model is a recognition model obtained by the method described in any implementation method of the first aspect above.
[0010] An embodiment of the present disclosure provides an electronic device, which includes one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by the one or more processors, the one or more processors implement the model training method or image recognition method described in any of the above implementation methods.
[0011] An embodiment of the present disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the model training method or image recognition method described in any of the above implementations.
[0012] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] FIG1 is a diagram of an exemplary system architecture in which the present disclosure may be applied;
[0014] FIG2 a is a flow chart of an embodiment of a model training method according to the present disclosure;
[0015] FIG2 b is a schematic diagram of an embodiment of a model training method according to the present disclosure;
[0016] FIG3 is a schematic diagram of an application scenario of the model training method according to the present disclosure;
[0017] FIG4 is a flow chart of another embodiment of an image recognition method according to the present disclosure;
[0018] FIG5 is a schematic diagram of an embodiment of a model training apparatus according to the present disclosure;
[0019] FIG6 is a schematic diagram of an image recognition device according to an embodiment of the present disclosure;
[0020] FIG7 is a schematic diagram of the structure of a computer system suitable for implementing a server according to an embodiment of the present disclosure. DETAILED DESCRIPTION
[0021] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may 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.
[0022] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0023] The present disclosure obtains a training sample set, constructs an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier, and trains the initial recognition model using the training sample set to obtain a trained recognition model, that is, based on multiple classic convolutional neural networks, a disease image feature category recognition network with strong generalization ability and good robustness is trained using a second category of image data set, and based on this recognition network, a feature extraction network is determined to perform feature transfer learning for the first category of images, so that the pathological knowledge learned in the disease image feature category recognition network can be generalized to the disease image feature category recognition of the first category of images, thereby achieving improved recognition accuracy for the disease image feature category of the first category of images without the need for a large amount of first category image data labeled with disease image feature categories.
[0024] FIG1 illustrates an exemplary system architecture 100 to which an embodiment of the model training method of the present disclosure may be applied.
[0025] As shown in Figure 1, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 is a medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.
[0026] The terminal devices 101 , 102 , and 103 interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications, such as image recognition applications and communication applications, may be installed on the terminal devices 101 , 102 , and 103 .
[0027] The terminal devices 101, 102, and 103 can be hardware or software. When the terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with display screens, including but not limited to mobile phones and laptops. When the terminal devices 101, 102, and 103 are software, they can be installed in the electronic devices listed above. They can be implemented as multiple software or software modules (for example, to provide services for model training), or as a single software or software module. No specific limitation is made here.
[0028] The server 105 may be a server that provides various services, for example, obtaining a training sample set, building an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier, and training the initial recognition model using the training sample set to obtain a recognition model.
[0029] It should be noted that the server 105 can be hardware or software. When the server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, for providing services for training models), or it can be implemented as a single software or software module. No specific limitations are made here.
[0030] It should be noted that the model training method provided in the embodiments of the present disclosure can be executed by the server 105, or by the terminal devices 101, 102, 103, or by the server 105 and the terminal devices 101, 102, 103 in cooperation with each other. Accordingly, the various parts (such as various units, sub-units, modules, and sub-modules) included in the model training apparatus can all be set in the server 105, or all be set in the terminal devices 101, 102, 103, or can be set in the server 105 and the terminal devices 101, 102, 103 respectively.
[0031] It should be understood that the number of terminal devices, networks, and servers in Figure 1 is merely illustrative and any number of terminal devices, networks, and servers may be provided as required.
[0032] FIG2a shows a flow chart 200 of an embodiment of a model training method that can be applied to the present disclosure. In this embodiment, the model training method includes the following steps:
[0033] Step 201: Obtain a training sample set.
[0034] In this embodiment, the execution entity (such as the server 105 or the terminal devices 101 , 102 , 103 as shown in FIG1 ) can obtain the training sample set from a local or remote server that stores the training sample set.
[0035] The training sample set may be a publicly available training sample set of medical images, wherein the training samples are first-category images labeled with disease image feature categories. The medical images may include one or more images acquired by various medical image acquisition devices (e.g., X-ray imaging devices, ultrasound imaging devices, and CT imaging devices), but are not limited thereto.
[0036] Here, the first type of image may include body parts, which may be lungs, heart, brain, etc., and this disclosure does not limit this.
[0037] Specifically, the training samples are images of children's lungs labeled with lung disease image feature categories, and the disease image feature categories may include category A (no disease manifestation), category B, and category C, etc.
[0038] In some optional manners, obtaining the training sample set includes: obtaining an initial training sample set; performing a preprocessing operation on the initial training sample set, and determining the initial training sample set after the preprocessing operation as the training sample set.
[0039] In this implementation, the execution entity may first obtain an initial training sample set, and divide the initial training sample set into a validation set, a test set, and a training set, that is, perform preprocessing operations according to different sets.
[0040] Furthermore, a preprocessing operation is performed on each set of the initial training sample set, and the initial training sample set after the preprocessing operation is determined as the training sample set.
[0041] Preprocessing operations here can include data augmentation and data normalization. Data augmentation mainly uses common data augmentation methods such as horizontal flipping, rotation, cropping, and adding noise to expand the training samples in the initial training sample set to increase the diversity of training data and prevent model overfitting.
[0042] Data standardization is mainly to reduce the impact of noise and grayscale changes in training samples on feature extraction and learning.
[0043] For example, the grayscale values of each channel of the data-enhanced training sample can be normalized to [-1, 1]. For each training sample I after data enhancement, its normalized expression can be expressed by the following formula: I = (I-127.5) / 127.5
[0044] This implementation method obtains an initial training sample set; performs a preprocessing operation on the initial training sample set, determines the initial training sample set after the preprocessing operation as the training sample set, and then uses the training sample set to train the initial recognition model to obtain a recognition model, which helps to further improve the recognition accuracy of the trained recognition model.
[0045] Step 202: construct an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier.
[0046] In this embodiment, the execution entity may construct an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier.
[0047] Among them, the preset ensemble learning classifier can be a multi-classifier system in existing technology or future development technology, such as a Stacking ensemble classifier, a Bagging ensemble classifier, etc., which is not limited in this disclosure.
[0048] Among them, the Stacking ensemble classifier is a hierarchical model integration framework, which specifically includes two layers. The first layer consists of multiple base classifiers, whose input is the original training set. The second layer consists of meta-classifiers. The output of the first-layer base classifier is added as a feature to the training set to retrain the meta-classifier of the second layer to obtain the trained Stacking ensemble classifier.
[0049] Here, the number of at least two different feature extraction networks is the same as the number of base classifiers in the preset ensemble learning classifier, one feature extraction network corresponds to one base classifier, and the output of the feature extraction network is the input of the corresponding base classifier.
[0050] Among them, the feature extraction network is determined based on the recognition network, and the recognition network can be trained based on the second type of images labeled with disease image feature categories in a public dataset. The first type of images and the second type of images include the same body parts.
[0051] Here, the body part may be the lungs, heart, brain, etc., and the present disclosure is not limited thereto.
[0052] It should be pointed out that the first type of images and the second type of images are different images. The second type of images are easier to obtain than the first type of images, and the amount of data of the second type of images is usually much larger than that of the first type of images.
[0053] For example, the age of the subject whose body part is included in the second category of images is different from the age of the subject whose body part is included in the first category of images. The first category of images is an X-ray image of a child's lungs, and the second category of images is an X-ray image of an adult's lungs.
[0054] Here, the recognition network may be an image recognition network in existing technology or future development technology, such as VGG16, resent50, inception-v3, etc.
[0055] It should be noted that different feature extraction networks correspond to different recognition networks. The feature extraction network can be a recognition network, that is, the feature extraction network is equivalent to the recognition network, or it can be a part of the recognition network used to extract features.
[0056] Specifically, the execution entity builds an initial recognition model based on two different feature extraction networks, such as a first feature extraction network and a second feature extraction network, and a Stacking ensemble classifier. The first feature extraction network is based on the first recognition network, such as VGG16 (VGG16 includes 13 convolutional layers and 3 fully connected layers), and is determined, for example, by using the convolutional layers and the first two fully connected layers in VGG16 as the first feature extraction network, that is, the features output by the second fully connected layer as the extracted features. The second feature extraction network is based on the second recognition network, such as Resnet50, and is determined, for example, by using Resnet50 as the entirety as the second feature extraction network.
[0057] In some optional embodiments, the preset ensemble learning classifier is a Stacking ensemble classifier.
[0058] In this implementation, the Stacking ensemble classifier is a hierarchical model integration framework, which specifically includes two layers. The first layer consists of multiple base classifiers, and the second layer consists of meta-classifiers. It can effectively combat overfitting, does not require too much parameter adjustment and feature selection, and has a good learning effect.
[0059] In some optional embodiments, constructing an initial recognition model based on at least two different feature extraction networks and preset ensemble learning classifiers includes: constructing an initial recognition model based on at least three different feature extraction networks and preset ensemble learning classifiers.
[0060] In this implementation, the execution entity may construct an initial recognition model based on at least three different feature extraction networks and a preset ensemble learning classifier.
[0061] Specifically, the execution entity constructs an initial recognition model based on three different feature extraction networks, such as a first feature extraction network, a second feature extraction network, and a third feature extraction network, and a preset ensemble learning classifier. The first feature extraction network is based on a first recognition network, such as VGG16. The second feature extraction network is based on a second recognition network, such as Resnet50. The third feature extraction network is based on a third recognition network, such as Inception-v3.
[0062] This implementation method constructs an initial recognition model based on at least three different feature extraction networks and a preset ensemble learning classifier, and then trains the initial recognition model to obtain a trained recognition model, which helps to further improve the recognition accuracy and robustness of the trained recognition model.
[0063] Step 203: Use the training sample set to train the initial recognition model to obtain a recognition model.
[0064] In this embodiment, after determining the initial recognition model, the execution subject may use the training sample set to train the initial recognition model, and terminate the training when the training termination condition is met to obtain the recognition model.
[0065] Among them, the conditions for the end of training may include but are not limited to at least one of the following: the training time exceeds the preset duration; the number of training times exceeds the preset number of times; the loss value calculated using the target loss function is less than the preset loss value threshold.
[0066] Specifically, as shown in Figure 2b, the initial recognition model includes two feature extraction networks, such as the first feature extraction network and the second feature extraction network, and a Stacking ensemble classifier. Two different feature extraction networks are used to extract features from training samples, such as children's lung X-ray images marked with lung disease image feature categories, to construct a transfer feature training set T = {T1, T2}, where T k ={(X k,1 ,y k,1 ), (X k,2 ,y k,2 )…(X k,N ,y k,N )},(X k,i ,y k,i ) represents the features extracted by the feature extraction network k for training sample i and the corresponding disease image feature category information, k∈{1,2}, y k,i ∈{1,2,…,M}, M is the number of disease categories, and N is the total number of training samples.
[0067] The stacking ensemble classifier consists of two base classifiers H1 and H2 and a meta-classifier C. The base classifier is composed of a Logistic regression model. Two features in the migration feature training set are used to train a Logistic regression model respectively, that is, T1 trains H1 and T2 trains H2.
[0068] For any data X k,i ∈T k , using the base classifier H k Make predictions and get probability distribution in, Indicates that training sample i passes through the base classifier H k Predict the probability that it belongs to category j, and then train a meta-classifier C consisting of an SVM based on the dataset constructed based on the probability distribution.
[0069] In addition, the training sample set can be divided into a validation set, a test set and a training set, and the validation set, the test set and the training set are used to train the initial recognition model.
[0070] Specifically, the training sample set includes a total of 5,856 training samples. The training samples are X-ray images of children's lungs labeled with lung disease image feature categories. There are three categories of lung disease image feature categories, namely Class A, Class B and Class C. For the training samples corresponding to each category, 10% are randomly selected as the validation set, 10% as the test set, and the remaining 80% as the training set.
[0071] Here, the validation set is used to evaluate the convergence of the model and whether the model is overfitting during the model training process. That is, after each training of the model on the training set for a certain number of times, the model will be predicted and evaluated on the validation set. If the effect of the training set continues to increase, but the effect of the validation set decreases instead of increasing, it indicates overfitting. When the effects of the model on the training set and the validation set are consistent and tend to be stable, it indicates that the model training has converged.
[0072] In some optional embodiments, the initialization parameters in the training process of the recognition network are determined by adopting model parameters trained with a preset image data set as the initialization parameters of the recognition network.
[0073] In this implementation, the execution entity may use model parameters trained using a preset image dataset as initialization parameters of the recognition network.
[0074] Among them, the preset image dataset can be any existing image dataset, for example, a medical library image dataset, an ImageNet dataset, etc., and this disclosure does not limit this.
[0075] Among them, the ImageNet dataset is a public dataset, and the images in the ImageNet dataset are all natural images in the real world.
[0076] This implementation method uses model parameters trained using a preset image dataset as initialization parameters of the recognition network, which helps the recognition network converge to the global optimum more quickly.
[0077] Continuing to refer to FIG3 , FIG3 is a schematic diagram of an application scenario of the model training method according to this embodiment.
[0078] In the application scenario of Figure 3, the execution entity 301 can obtain a training sample set 302 from a local or remote server that stores the training sample set, where the training samples are first-category images labeled with disease image feature categories, such as X-ray images of children's lungs labeled with lung disease image feature categories; based on at least two different feature extraction networks 303 and Stacking ensemble classifiers 304, an initial recognition model 305 is constructed, wherein the number of at least two different feature extraction networks is the same as the number of base classifiers in the Stacking ensemble classifier, one feature extraction network corresponds to one base classifier, the output of the feature extraction network is the input of the corresponding base classifier, the feature extraction network is based on the recognition network, such as VGG16 and resnet50, and the recognition network is trained based on the second-category images labeled with disease image feature categories, such as X-ray images of adult lungs labeled with lung disease image feature categories; the initial recognition model 305 is trained using the training sample set 302 to obtain a recognition model 306.
[0079] The model training method disclosed in the present invention obtains a training sample set, constructs an initial recognition model based on at least two different feature extraction networks and a Stacking ensemble classifier, and uses the training sample set to train the initial recognition model to obtain a recognition model, thereby effectively improving the recognition accuracy of the trained recognition model for the types of diseases in children's body parts.
[0080] Continuing to refer to FIG4 , a process 400 of an embodiment of an image recognition method according to the present disclosure is shown. The image recognition method includes the following steps:
[0081] Step 401: Acquire a first type of image including body parts.
[0082] In this embodiment, the execution entity may obtain the first type of image containing body parts through a wired or wireless connection.
[0083] Here, the body part may be the lungs, heart, brain, etc.
[0084] Among them, the wireless connection method may include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or developed in the future. For example, an X-ray imaging device collects images of body parts and sends them to a server via a wired or wireless connection method. The server or a computer connected to the server processes the images of the body parts to form a first type of image. The first type of image can be stored in a server, a computer connected to the server, etc., for the above-mentioned execution entity (such as a preset computer, server, computer cluster or server cluster) to obtain through a wired or wireless connection method.
[0085] Step 402: Input the first type of image into the recognition model to obtain the disease image feature category corresponding to the body part.
[0086] In this embodiment, after acquiring a first category of images including body parts, the executing entity may input the first category of images into a recognition model to obtain the disease image feature categories corresponding to the body parts. The recognition model is obtained using the method described in the embodiment corresponding to FIG. 2 and is not further described here.
[0087] Specifically, the execution entity may obtain an image to be identified containing a child's lungs in a wired or wireless manner, input the image to be identified into a recognition model, and obtain a disease image feature category corresponding to the child's lungs.
[0088] The image recognition method provided by the embodiment of the present disclosure obtains an image to be recognized including a child's body parts; inputs the image to be recognized into a recognition model to obtain disease image feature categories corresponding to the child's body parts, thereby effectively improving the recognition accuracy of the disease image feature categories of the child's body part images.
[0089] Further referring to FIG5 , as an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a model training device. The device embodiment corresponds to the method embodiment shown in FIG2 , and the device can be specifically applied to various electronic devices.
[0090] As shown in FIG5 , the model training device 500 of this embodiment includes: an acquisition module 501 , a construction module 502 and a training module 503 .
[0091] The acquisition module 501 may be configured to acquire a training sample set.
[0092] The construction module 502 may be configured to construct an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier.
[0093] The training module 503 may be configured to train the initial recognition model using the training sample set to obtain a recognition model.
[0094] In some optional aspects of this embodiment, the construction module is further configured to: construct an initial recognition model based on at least three different feature extraction networks and a preset ensemble learning classifier.
[0095] In some optional aspects of this embodiment, the acquisition module is further configured to acquire an initial training sample set; perform a preprocessing operation on the initial training sample set, and determine the initial training sample set after the preprocessing operation as the training sample set.
[0096] In some optional aspects of this embodiment, the initialization parameters in the training process of the recognition network are determined in the following manner: model parameters trained using a preset image data set are used as the initialization parameters of the recognition network.
[0097] In some optional aspects of this embodiment, the preset ensemble learning classifier is a Stacking ensemble classifier.
[0098] With further reference to FIG6 , as an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an image recognition device. This device embodiment corresponds to the method embodiment shown in FIG4 , and the device can be specifically applied to various electronic devices.
[0099] As shown in FIG6 , the image recognition device 600 of this embodiment includes an acquisition module 601 and a recognition module 602 .
[0100] The acquisition module 601 may be configured to acquire a first type of image including body parts.
[0101] The recognition module 602 may be configured to input the first type of image into a recognition model to obtain a disease image feature category corresponding to a body part.
[0102] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device and a readable storage medium.
[0103] As shown in FIG7 , it is a block diagram of an electronic device according to the model training method of an embodiment of the present disclosure.
[0104] 700 is a block diagram of an electronic device according to a method for model training according to an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.
[0105] As shown in Figure 7, the electronic device includes: one or more processors 701, a memory 702, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). In Figure 7, a processor 701 is taken as an example.
[0106] Memory 702 is a non-transitory computer-readable storage medium provided by the present disclosure. The memory stores instructions executable by at least one processor to cause the at least one processor to perform the model training method provided by the present disclosure. The non-transitory computer-readable storage medium of the present disclosure stores computer instructions for causing a computer to perform the model training method provided by the present disclosure.
[0107] The memory 702, as a non-transient computer-readable storage medium, can be used to store non-transient software programs, non-transient computer executable programs, and modules, such as the program instructions / modules corresponding to the model training method in the embodiments of the present disclosure (for example, the acquisition module 501, the construction module 502, and the training module 503 shown in FIG5). The processor 701 executes the various functional applications and data processing of the server by running the non-transient software programs, instructions, and modules stored in the memory 702, that is, implementing the model training method in the above method embodiment.
[0108] The memory 702 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created by the use of the electronic device for model training, etc. In addition, the memory 702 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory 702 may optionally include a memory remotely located relative to the processor 701, and these remote memories may be connected to the electronic device for model training via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.
[0109] The electronic device of the model training method may further include: an input device 703 and an output device 704. The processor 701, the memory 702, the input device 703 and the output device 704 may be connected via a bus or other means, and FIG7 takes the bus connection as an example.
[0110] The input device 703 can receive input digital or character information, such as an input device such as a touch screen, a keypad, a mouse, a trackpad, a touch pad, a pointer, one or more mouse buttons, a trackball, a joystick, etc. The output device 704 can include a display device, an auxiliary lighting device (e.g., an LED), and a tactile feedback device (e.g., a vibration motor). The display device can include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display, and a plasma display. In some embodiments, the display device can be a touch screen.
[0111] Various implementations of the systems and techniques described herein can be realized in digital electronic circuit systems, integrated circuit systems, dedicated ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0112] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and / or device (e.g., a magnetic disk, an optical disk, a memory, a programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.
[0113] To provide 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 pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the 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 input, voice input, or tactile input).
[0114] 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 implementations 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 communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.
[0115] Computer systems may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The client and server relationship arises through computer programs running on the respective computers and having a client-server relationship to each other.
[0116] According to the technical solution of the embodiment of the present disclosure, the recognition accuracy of the trained recognition model for the feature categories of body part diseases in the first category of images is effectively improved.
[0117] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions of this disclosure can be achieved, and this document is not limited here.
[0118] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on 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 application.
Claims
1. A model training method, the method comprising: Obtaining a training sample set, wherein the training samples are first-category images labeled with disease image feature categories; Based on at least two different feature extraction networks and a preset ensemble learning classifier, an initial recognition model is constructed, wherein the number of the at least two different feature extraction networks is the same as the number of base classifiers in the preset ensemble learning classifier, one feature extraction network corresponds to one base classifier, the output of the feature extraction network is the input of the corresponding base classifier, the feature extraction network is determined based on the recognition network, the recognition network is trained based on a second type of image labeled with a disease image feature category, and the first type of image and the second type of image include the same body parts; The initial recognition model is trained using the training sample set to obtain a recognition model.
2. The method according to claim 1, wherein: The initial recognition model is constructed based on at least two different feature extraction networks and a preset integrated learning classifier, including: An initial recognition model is constructed based on at least three different feature extraction networks and a preset ensemble learning classifier.
3. The method according to claim 1, wherein: The step of obtaining a training sample set includes: Obtain an initial training sample set; A preprocessing operation is performed on the initial training sample set, and the initial training sample set after the preprocessing operation is determined as the training sample set, wherein the preprocessing operation includes data enhancement and data standardization.
4. The method according to claim 1, wherein: The initialization parameters in the training process of the recognition network are determined in the following way: The model parameters trained by the preset image data set are used as the initialization parameters of the recognition network.
5. The method according to any one of claims 1 to 4, wherein: The preset ensemble learning classifier is a Stacking ensemble classifier.
6. The method according to any one of claims 1 to 5, wherein: The age of the object to which the body parts included in the second category of images belong is different from the age of the object to which the body parts included in the first category of images belong.
7. An image recognition method, the method comprising: Acquire a first type of image including a body part; The first type of image is input into a recognition model to obtain a disease image feature category corresponding to the body part, wherein the recognition model is a recognition model obtained by the method according to any one of claims 1-6.
8. A model training device, comprising: An acquisition module is configured to acquire a training sample set, wherein the training samples are first-category images labeled with disease image feature categories; A construction module is configured to construct an initial recognition model based on at least two different feature extraction networks and a preset ensemble learning classifier, wherein the number of the at least two different feature extraction networks is the same as the number of base classifiers in the preset ensemble learning classifier, one feature extraction network corresponds to one base classifier, the output of the feature extraction network is the input of the corresponding base classifier, the feature extraction network is determined based on the recognition network, the recognition network is trained based on a second type of image labeled with a disease image feature category, and the first type of image and the second type of image include the same body parts; The training module is configured to train the initial recognition model using the training sample set to obtain a recognition model.
9. The device according to claim 8, wherein: The building blocks are further configured to: An initial recognition model is constructed based on at least three different feature extraction networks and a preset ensemble learning classifier.
10. The device according to claim 8, wherein: The acquisition module is further configured to: Obtain an initial training sample set; A preprocessing operation is performed on the initial training sample set, and the initial training sample set after the preprocessing operation is determined as the training sample set, wherein the preprocessing operation includes data enhancement and data standardization.
11. The device according to claim 8, wherein: The initialization parameters in the training process of the recognition network are determined in the following way: The model parameters trained by the preset image data set are used as the initialization parameters of the recognition network.
12. The device according to any one of claims 8 to 11, wherein: The preset ensemble learning classifier is a Stacking ensemble classifier.
13. An image recognition device, comprising: An acquisition module configured to acquire a first type of image including a body part; The recognition module is configured to input the first type of image into a recognition model to obtain a disease image feature category corresponding to the body part, wherein the recognition model is a recognition model obtained by the method according to one of claims 1-6.
14. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor to enable the at least A processor is capable of executing the method according to any one of claims 1 to 7.
15. 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 to 7.
Citation Information
Patent Citations
Arrhythmia prediction method
CN111329469A
Detection device, training method, training device, equipment and medium
CN113409306A
Automatic chromosome segmentation and classification method based on deep learning
CN113658150A
Model training method, image recognition method and device
CN116645546A
System for interpreting medical images through machine learnings
KR1020180040287A
