Laser energy control system for lithotripsy

Abstract

The present application relates to a kind of laser energy control systems for lithotripsy, it includes: image acquisition module, the image data of several patients is acquired in real time, and image data is preprocessed;Segmentation and identification module, based on deep learning, the image data after preprocessing is segmented and identified, and the stone image with stone position and shape characteristics is obtained;Feature extraction module, the stone position feature, shape and size feature in stone image are extracted;Intelligent analysis module is used to receive stone position feature, shape and size feature, with the horizontal center line in image as reference, determine whether stone position is located in the region to be processed;Control module, to determine the size of stone with stone shape and size feature, the output laser energy value is determined by combining stone size and stone position.The present application can accurately identify the position and size of stone, and then determine the corresponding laser energy, improve the accuracy of stone processing, avoid causing secondary injury.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical image processing, and particularly relates to a laser energy control system for crushing calculi. BACKGROUND

[0002] A calculus refers to a stone in the body, which is produced when biological material is solidified by organic or inorganic matter, and can block the bile duct or ureter, has a high incidence and recurrence rate, and has a great impact on the health of patients. Therefore, when a calculus is found to be large enough to reach a dangerous size, the calculus needs to be removed. Traditional calculus diagnosis is mainly performed by X-ray or ultrasound, but due to the limitations of the position, size and other factors of the calculus, the accuracy is limited.

[0003] With the rapid development of computer technology and artificial intelligence technology, how to accurately identify the position and size of the calculus through image recognition and big data processing and other technologies to accurately process the calculus has become a technical problem to be solved at present. SUMMARY

[0004] In view of the above problems, the purpose of the present application is to provide a laser energy control system for crushing calculi, which can accurately identify the position and size of the calculus, and then determine the corresponding laser energy, improve the accuracy of calculus processing, and avoid causing secondary damage.

[0005] To achieve the above purpose, the present application adopts the following technical scheme: a laser energy control system for crushing calculi, comprising: an image acquisition module, which acquires a plurality of image data of a patient in real time, and pre-processes the image data; a segmentation and recognition module, which segments and recognizes the pre-processed image data based on deep learning to obtain a calculus image with a calculus position and shape feature; a feature extraction module, which extracts a calculus position feature, shape and size feature in the calculus image; an intelligent analysis module, which is used for receiving the calculus position feature, shape and size feature, and determining whether the calculus position is located in a to-be-processed region based on a horizontal center line in the image; and a control module, which determines a calculus size based on the calculus shape and size feature, and determines an output laser energy value by combining the calculus size with the calculus position.

[0006] Further, in the segmentation and recognition module, the deep learning adopts a convolutional neural network model; and the convolutional neural network model is trained and optimized.

[0007] Further, the convolutional neural network model comprises a convolutional layer, a pooling layer and a fully connected layer, the model is trained by back propagation, and the weight and bias values are optimized in the training process to minimize the loss function.

[0008] Further, the training and optimization of the convolutional neural network model further comprises a data set division module and an enhancement processing module; the data set division module divides the preprocessed image data into a training set and a test set, and the image data in the training set is used to train the model; the enhancement processing module performs enhancement processing on the image data in the training set.

[0009] Further, the training and optimization of the convolutional neural network model further comprises a regularization module and a network pruning module, which improves the generalization ability of the model through the regularization module and removes unnecessary connections and nodes in the model through the network pruning module.

[0010] Further, the training of the convolutional neural network model adopts distributed training, and the training task is distributed to multiple computing nodes for parallel computing.

[0011] Further, the intelligent analysis module comprises a data receiving module and a data analysis module.

[0012] The data receiving module is configured to receive the stone location features, shape and size features transmitted by the feature extraction module.

[0013] The data analysis module is provided with a stone setting area, and the path of the laser is set as the horizontal center line in the image, and the path of the laser is located at the center position of the stone setting area; according to the stone location features, it is determined whether the stone is located in the stone setting area, and at the same time, it is determined whether the center line of the stone shape overlaps with the horizontal center line in the image, according to the stone location and the overlapping condition of the center line and the horizontal center line, the accurate stone location is obtained.

[0014] Further, if the stone is located in the stone setting area and the center line of the stone shape overlaps with the horizontal center line in the image, the first stone location information is obtained.

[0015] If the stone is located in the stone setting area and the center line of the stone shape deviates from the horizontal center line in the image, the second stone location information is obtained.

[0016] If the stone is located outside the stone setting area, the third stone location information is obtained.

[0017] Further, the control module adjusts the size of the laser energy according to the received first stone location information, second stone location information and third stone location information.

[0018] If the first stone location information is received, the control module adjusts the laser energy of the laser device to the normal use range.

[0019] If the second stone location information is received, the control module reduces the laser energy of the laser device by 1 to 2 levels.

[0020] If the second position information of the calculus is received, the control module adjusts the laser energy of the laser device to zero.

[0021] Further, 0.1J is taken as a level.

[0022] The present application has the following advantages due to the above technical solutions:

[0023] 1. The present application can automatically complete the identification of calculus image data without manual intervention, reducing the workload of manual operation.

[0024] 2. The present application uses deep learning to efficiently process data, which can quickly obtain accurate results, improve the accuracy and precision of lithotripsy, and effectively avoid secondary damage.

[0025] 3. The present application can be widely applied to the treatment of various calculi and has high practical value. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a structure diagram of the laser energy control system for crushing calculus in the embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme of the embodiments of the present application will be described clearly and completely below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0028] It should be noted that the terms used here are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and furthermore, it should be understood that when the terms "comprise" and / or "include" are used in the specification, there is a feature, step, operation, device, component and / or combination thereof.

[0029] In order to solve the problem of limited accuracy of stone crushing in the prior art due to the limitations of the location, size and other factors of the stone, the present application provides a laser energy control system for crushing stones, which comprises: an image acquisition module, which acquires a plurality of image data of a patient in real time and pre-processes the image data; a segmentation and recognition module, which segments and recognizes the pre-processed image data based on deep learning to obtain a stone image with stone location and shape characteristics; a feature extraction module, which extracts stone location features, shape and size features in the stone image; an intelligent analysis module, which receives the stone location features, shape and size features, and determines whether the stone location is in a to-be-processed area based on a horizontal center line in the image; and a control module, which determines the stone size based on the stone shape and size features, combines the stone size with the stone location, and determines an output laser energy value. The present application can accurately identify the location and size of the stone, and then determine the corresponding laser energy, thereby improving the accuracy of stone processing and avoiding secondary damage.

[0030] In one embodiment of the present application, a laser energy control system for crushing stones is provided. In this embodiment, as shown in Figure 1 , the system comprises:

[0031] An image acquisition module, which acquires a plurality of image data of a patient in real time and pre-processes the image data;

[0032] A segmentation and recognition module, which segments and recognizes the pre-processed image data based on deep learning to obtain a stone image with stone location and shape characteristics;

[0033] A feature extraction module, which extracts stone location features, shape and size features in the stone image;

[0034] An intelligent analysis module, which receives the stone location features, shape and size features, and determines whether the stone location is in a to-be-processed area based on a horizontal center line in the image;

[0035] A control module, which determines the stone size based on the stone shape and size features, combines the stone size with the stone location, and determines an output laser energy value, so as to effectively crush the stone and avoid secondary damage to the human body.

[0036] In a feasible implementation, in the image acquisition module, the pre-processing includes gray scale processing and filtering processing of the image. The image data of the patient can be acquired by devices such as CT scanning, MRI and / or camera in an endoscope, which are not limited herein.

[0037] In a feasible implementation, in the segmentation and recognition module, the deep learning adopts a convolutional neural network (CNN) model. In this embodiment, the model parameters can be adjusted to improve the accuracy and stability of the model by training and optimizing the convolutional neural network model.

[0038] In this embodiment, the convolutional neural network model includes convolutional layers, pooling layers, and fully connected layers, etc. The model is trained by back propagation, and the weights and bias values are gradually optimized during the training process to minimize the loss function.

[0039] The data set division module divides the preprocessed image data into a training set and a test set, and the image data in the training set is used to train the model. The enhancement processing module performs enhancement processing on the image data in the training set to expand the training set and improve the robustness and generalization ability of the model. The enhancement processing includes random cropping, random translation, random rotation, random color disturbance, etc.

[0040] Optionally, in the training of the convolutional neural network model, a regularization module and a network pruning module are further included. The regularization module improves the generalization ability of the model to avoid overfitting of the model. The network pruning module removes unnecessary connections and nodes in the model to reduce the amount of calculation and storage space, and improve the efficiency and speed of the model.

[0041] In this embodiment, the regularization includes L1 regularization, L2 regularization, and dropout, etc., which are not limited herein.

[0042] Optionally, the training of the convolutional neural network model can also use distributed training to distribute the training tasks to multiple computing nodes for parallel computing to speed up the training and reduce the training time.

[0043] In a feasible implementation, the intelligent analysis module includes a data receiving module and a data analysis module. Specifically, the data receiving module receives the stone location features, shape and size features transmitted by the feature extraction module.

[0044] The data analysis module is provided with a stone setting area, and the path of the laser is set as the horizontal center line in the image. The path of the laser is located at the center position of the stone setting area. According to the stone location features, it is determined whether the stone is located in the stone setting area, and whether the center line of the stone shape overlaps with the horizontal center line in the image. According to the stone location and the overlap of the center line and the horizontal center line, the accurate stone location is obtained.

[0045] If the stone is located in the stone setting area and the center line of the stone shape overlaps with the horizontal center line in the image, the first stone location information is obtained.

[0046] If the stone is located in the stone setting area and the center line of the stone shape deviates from the horizontal center line in the image, the second stone location information is obtained.

[0047] If the stone is located outside the stone setting area, the third stone position information is obtained.

[0048] In one possible implementation, the control module is connected with the laser device for crushing the stone, and is configured to control the size of the laser energy output by the laser device.

[0049] The control module adjusts the size of the laser energy according to the received first stone position information, second stone position information and third stone position information, so as to accurately crush the stone and improve the accuracy of stone crushing, and avoid secondary damage to the human body.

[0050] Specifically, if the first stone position information is received, the control module adjusts the laser energy of the laser device to a normal use range; the laser pulse energy in the normal use range is 0.2-6.0J.

[0051] If the second stone position information is received, the control module reduces the laser energy of the laser device by 1-2 levels, and in this embodiment, 0.1J is taken as one level.

[0052] If the second stone position information is received, the control module adjusts the laser energy of the laser device to zero.

[0053] In the above embodiments, the control module can be arranged in a computer and / or a cloud server, and can be adjusted in function according to actual needs.

[0054] In summary, the present application processes the stone image based on deep learning and intelligent analysis module, efficiently segments, identifies and extracts features from the input image data, can adjust the laser energy value output by the laser device in real time, effectively improves the accuracy of stone crushing, can provide more accurate diagnosis and treatment plan for clinicians, and further promotes the development and progress of medical technology.

[0055] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A laser energy control system for lithotripsy, characterized by, The application relates to a kidney stone treatment system. The system comprises: an image acquisition module, which acquires a plurality of image data of a patient in real time and pre-processes the image data; a segmentation and recognition module, which segments and recognizes the pre-processed image data based on deep learning to obtain a kidney stone image with a kidney stone position and shape feature; a feature extraction module, which extracts a kidney stone position feature, shape and size feature in the kidney stone image; an intelligent analysis module, which receives the kidney stone position feature, shape and size feature, and determines whether the kidney stone position is located in a to-be-processed region based on a horizontal center line in the image; a control module, which determines a kidney stone size based on the kidney stone shape and size feature, combines the kidney stone size with the kidney stone position, and determines an output laser energy value. The intelligent analysis module comprises a data receiving module and a data analysis module. The data receiving module receives the kidney stone position feature, shape and size feature transmitted by the feature extraction module. The data analysis module is provided with a kidney stone setting region, and a path of laser is set as the horizontal center line in the image, and the path of laser is located at a center position of the kidney stone setting region; whether the kidney stone is located in the kidney stone setting region is determined based on the kidney stone position feature, and whether a center line of the kidney stone shape overlaps the horizontal center line in the image is determined, and an accurate kidney stone position is obtained based on the kidney stone position and the overlap of the center line and the horizontal center line. If the kidney stone is located in the kidney stone setting region and the center line of the kidney stone shape overlaps the horizontal center line in the image, first position information of the kidney stone is obtained. If the kidney stone is located in the kidney stone setting region and the center line of the kidney stone shape deviates from the horizontal center line in the image, second position information of the kidney stone is obtained. If the kidney stone is located outside the kidney stone setting region, third position information of the kidney stone is obtained. The control module adjusts the size of laser energy based on the received first position information, second position information and third position information of the kidney stone. If the first position information of the kidney stone is received, the control module adjusts the laser energy of the laser device to a normal use range. If the second position information of the kidney stone is received, the control module reduces the laser energy of the laser device by 1 to 2 levels.

2. The laser energy control system for lithotripsy of claim 1, wherein, If the third position information of the kidney stone is received, the control module adjusts the laser energy of the laser device to zero.

3. The system for controlling laser energy for breaking up a calculus of claim 2, wherein In the segmentation and recognition module, a convolutional neural network model is used for deep learning, and the convolutional neural network model is trained and optimized.

4. The system for controlling laser energy for breaking up a calculus according to claim 2, wherein The convolutional neural network model comprises a convolutional layer, a pooling layer and a full connection layer, the model is trained through back propagation, and weight and bias values are optimized in the training process to minimize a loss function. The training and optimization of the convolutional neural network model further comprise a data set division module and an enhancement processing module.

5. The system for controlling laser energy for breaking up a calculus of claim 2, wherein, The data set division module divides the pre-processed image data into a training set and a test set, and the model is trained by image data in the training set; the enhancement processing module performs enhancement processing on the image data in the training set. The training and optimization of the convolutional neural network model further comprise a regularization module and a network pruning module, the model generalization ability is improved through the regularization module, and unnecessary connections and nodes in the model are removed through the network pruning module.

6. The system for controlling laser energy for breaking up a calculus of claim 2, wherein, The training of the convolutional neural network model adopts distributed training, and the training task is distributed to multiple computing nodes for parallel computation.

7. The system for controlling laser energy for breaking up a calculus of claim 1, wherein, 0.1 J is taken as one level.

Images