CN111080584B - Quality control method for medical image, computer device and readable storage medium - Google Patents
Quality control method for medical image, computer device and readable storage medium Download PDFInfo
- Publication number
- CN111080584B CN111080584B CN201911221084.0A CN201911221084A CN111080584B CN 111080584 B CN111080584 B CN 111080584B CN 201911221084 A CN201911221084 A CN 201911221084A CN 111080584 B CN111080584 B CN 111080584B
- Authority
- CN
- China
- Prior art keywords
- image
- tomographic
- classification
- result
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003908 quality control method Methods 0.000 title claims abstract description 44
- 238000001514 detection method Methods 0.000 claims abstract description 259
- 238000000034 method Methods 0.000 claims abstract description 51
- 238000010801 machine learning Methods 0.000 claims abstract description 15
- 238000013145 classification model Methods 0.000 claims description 67
- 238000012549 training Methods 0.000 claims description 37
- 238000004590 computer program Methods 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 20
- 238000005070 sampling Methods 0.000 claims description 19
- 238000002372 labelling Methods 0.000 claims description 16
- 238000002591 computed tomography Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 238000003384 imaging method Methods 0.000 description 5
- 210000004204 blood vessel Anatomy 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 210000004072 lung Anatomy 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 210000001165 lymph node Anatomy 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000013421 nuclear magnetic resonance imaging Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 208000031481 Pathologic Constriction Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 210000000621 bronchi Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000747 cardiac effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 230000010247 heart contraction Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 201000005202 lung cancer Diseases 0.000 description 1
- 208000020816 lung neoplasm Diseases 0.000 description 1
- 238000002595 magnetic resonance imaging Methods 0.000 description 1
- 210000001370 mediastinum Anatomy 0.000 description 1
- 230000001394 metastastic effect Effects 0.000 description 1
- 206010061289 metastatic neoplasm Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 208000018280 neoplasm of mediastinum Diseases 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000002600 positron emission tomography Methods 0.000 description 1
- 238000002601 radiography Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000036262 stenosis Effects 0.000 description 1
- 208000037804 stenosis Diseases 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 201000008827 tuberculosis Diseases 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The application relates to a quality control method of medical images, computer equipment and a readable storage medium. The method comprises the following steps: acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images; inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image; and displaying the detection result on a display interface of the medical image. The method is displayed to the corresponding technicians in time after the detection results are obtained, so that the technicians can learn the problems existing in shooting and correct the problems in time, and the quality of the obtained medical images can be greatly improved. Furthermore, the process of obtaining the medical image detection result is executed by the image detection model, manual participation is not needed, and the accuracy and the efficiency of quality control are greatly improved.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to a quality control method for medical images, a computer device, and a readable storage medium.
Background
With the continuous development of medical technology, the continuous perfection of social medical system and the increasing of the living standard of people, the electronic computer tomography (Computed Tomography, CT) image diagnosis technology has become an important device for diagnosing the illness state of human body, and the electronic computer tomography (Computed Tomography, CT) image diagnosis technology uses precisely collimated X-ray beams, gamma rays, ultrasonic waves and the like to scan one by one section around a certain part of the human body together with a detector with extremely high sensitivity, and has the characteristics of quick scanning time, clear image and the like. However, the quality of the CT image obtained by shooting is also irregular due to the different medical level at home, and the image with poor quality directly affects the diagnosis of the subsequent doctor and the treatment of the patient.
Conventional CT image quality control is usually performed by performing spot check on a CT image by a manual method, for example, whether the spot check includes an artifact or foreign matter, or whether the shooting range is complete, and then transmitting the spot check result to a corresponding technician for improvement.
However, the traditional CT image quality control method has strong subjective awareness, so that the accuracy of the obtained quality control result is low, and the efficiency is also low.
Disclosure of Invention
Based on this, it is necessary to provide a quality control method, a computer device and a readable storage medium for medical images, aiming at the problem of low accuracy and efficiency of quality control results in the conventional technology.
In a first aspect, an embodiment of the present application provides a quality control method for a medical image, including:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
and displaying the detection result on a display interface of the medical image.
In one embodiment, the method further comprises:
inputting a plurality of tomographic images into an image classification model to obtain a classification result of each tomographic image; the image classification model is a two-dimensional network model, and the classification result represents whether the acquisition range of the medical image is complete or not;
and displaying the classification result on a display interface of the medical image.
In one embodiment, inputting a plurality of tomographic images into an image detection model includes:
acquiring the number of each tomographic image;
dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval;
each set of tomographic images is input to an image detection model.
In one embodiment, the image detection model is a three-dimensional network model, and the detection result includes an artifact detection result and/or a foreign object detection result; obtaining a detection result of each tomographic image according to the continuity characteristic information between each tomographic image, including:
obtaining the labeling result of each tomographic image by utilizing the continuity characteristic information among the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result;
comparing the artifact classification result with a first preset threshold value to determine an artifact detection result;
and comparing the foreign matter classification result with a second preset threshold value to determine a foreign matter detection result.
In one embodiment, the classification result comprises an acquisition range classification result; inputting a plurality of tomographic images into an image classification model to obtain a classification result of each tomographic image, comprising:
Performing interval sampling on a plurality of tomographic images, and inputting each sampled tomographic image into an image classification model to obtain a classification value of each sampled tomographic image;
acquiring adjacent tomographic images of the sampled tomographic images with the classification values larger than a third preset threshold, and inputting the adjacent tomographic images into an image classification model to obtain classification values of the adjacent tomographic images;
and determining the acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold value.
In one embodiment, the method further comprises:
inputting the first training sample image into an initial image detection model to obtain an initial artifact detection result and/or an initial foreign matter detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions;
inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain the image classification model meeting the preset conditions.
In one embodiment, the method further comprises:
storing coordinates of the artifact detection frame and/or the foreign object detection frame and numbers of tomographic images including the artifact and/or the foreign object;
and storing the serial numbers of the tomographic images with incomplete acquisition ranges.
In one embodiment, the method further comprises:
counting the numbers of the stored tomographic images according to the counting requirement to obtain a quality counting result of the medical image; the statistical requirements include at least one of artifact statistical requirements, foreign object statistical requirements, and acquisition range insufficiency statistical requirements.
In a second aspect, an embodiment of the present application provides a quality control apparatus for medical images, including:
the acquisition module is used for acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
the processing module is used for inputting a plurality of tomographic images into the image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
And the display module is used for displaying the detection result on a display interface of the medical image.
In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
and displaying the detection result on a display interface of the medical image.
In a fourth aspect, embodiments of the present application provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
And displaying the detection result on a display interface of the medical image.
The quality control method, the quality control device, the computer equipment and the readable storage medium for the medical image can acquire the medical image to be detected, wherein the medical image comprises a plurality of tomographic images; inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image; and displaying the detection result on a display interface of the medical image. In the method, the image detection model detects the medical image according to the continuity characteristic information among each tomographic image, so that the accuracy of a detection result can be improved; and the detection result is displayed to the corresponding technician in time after being obtained, so that the technician can learn the problems existing in shooting and correct the problems in time, and the quality of the obtained medical image can be greatly improved. Furthermore, the process of obtaining the medical image detection result by the method is executed by the image detection model, manual participation is not needed, and the accuracy and the efficiency of quality control are greatly improved.
Drawings
Fig. 1 is a flow chart of a quality control method for medical images according to an embodiment;
FIG. 1a is a schematic illustration of labeling of detection results provided in one embodiment;
fig. 2 is a flowchart of a quality control method for medical images according to another embodiment;
fig. 3 is a flowchart of a quality control method for medical images according to another embodiment;
fig. 3a is a flowchart illustrating a quality control method of a medical image according to another embodiment;
fig. 4 is a flowchart of a quality control method for medical images according to another embodiment;
fig. 5 is a schematic structural diagram of a quality control device for medical images according to an embodiment;
fig. 6 is a schematic diagram of an internal structure of a computer device according to an embodiment.
Detailed Description
The quality control method of the medical image provided by the embodiment of the application can be suitable for the quality control process of the medical image, and the medical image can be direct digital flat-panel X-ray imaging (Digital Radiography, DR), electronic computer tomography imaging (Computed Tomography, CT), nuclear magnetic resonance imaging (Nuclear Magnetic Resonance Imaging, MRI), positron emission tomography imaging (Positron Emission Computed Tomography, PET) and the like. For example, CT is usually used for contrast enhanced scanning to determine whether there is a tumor or lymph node enlargement of the mediastinum and the hilum, whether there is a stenosis or obstruction of the bronchi, which greatly helps to detect primary and metastatic mediastinum tumors, tuberculosis of lymph nodes, central lung cancer, and the like, and the interstitial and parenchymal lesions in the lung can be well displayed. However, because the current medical level is still good, the CT photographing technology of the lower level hospital is not well trained and standardized, and the photographed image may not be able to generate a diagnosis conclusion for the lower level hospital due to poor quality, and is unfavorable for the review of the expert of the upper level hospital. Therefore, quality control of the photographed medical image is required to improve the quality of the medical image.
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application are further described in detail by the following embodiments with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
It should be noted that, the execution body of the method embodiment described below may be a quality control device of a medical image, and the device may be implemented as part or all of a computer device by software, hardware, or a combination of software and hardware. The following method embodiments are described taking an execution subject as an example of a computer device, where the computer device may be a terminal, a server, a separate computing device, or integrated on a medical imaging device, and this embodiment is not limited to this.
Fig. 1 is a flow chart of a quality control method for medical images according to an embodiment. The embodiment relates to a specific process that computer equipment detects medical images to be detected, obtains detection results and displays the detection results. As shown in fig. 1, the method includes:
s101, acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images.
Specifically, the medical image to be detected acquired by the computer device may be a medical image that has been shot by the current technician, that is, when the technician shoots a film, the computer device automatically acquires the medical image that has been shot at the current time and performs a subsequent detection process.
The medical image comprises a plurality of tomographic images, taking a CT image as an example, a plurality of tomographic images can be obtained by one CT shooting, each tomographic image is a two-dimensional image, continuity structure information is arranged between adjacent tomographic images, and the whole CT image formed by the continuity structure information is a three-dimensional image.
S102, inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image.
Specifically, the computer device may input a plurality of tomographic images into an image detection model, where the image detection model is a machine learning model, and optionally, the image detection model may be a three-dimensional network model, and then the image detection model processes the entire three-dimensional image, and may be a neural network model, such as a VNet-FPN, faster-RCNN, or RetinaNet model, or may be another machine learning model. The image detection model can comprise a series of convolution layers, pooling layers and full-connection layers, and feature extraction is carried out on each tomographic image in the whole three-dimensional image through the convolution layers and the pooling layers, so that feature information of each tomographic image can be obtained, and continuity of feature information between adjacent tomographic images is obtained; and then the full-connection layer can obtain the detection result of each tomographic image according to the continuity characteristic information between each tomographic image. Alternatively, the image detection model may be a plurality of two-dimensional network models, and each tomographic image is processed respectively to obtain a detection result of each tomographic image.
The detection result represents the quality result of the medical image, such as whether the medical image has artifact or foreign matter, whether the exposure is good, whether the spine is inclined, and the like. The detection result may include a position, a type, and the like, where the image quality is not acceptable. The continuity characteristic information characterizes the continuity of the structural information between the adjacent tomograms, the pixel attribute of the same tissue or structure in the previous tomogram and the next tomogram should have certain regularity, and if the regularity is not satisfied, the pixel attribute may be artifact or foreign matter; if the shot cardiac coronary CT image generally contains blood vessels, and the distribution of the blood vessels has position regularity, if the positions of corresponding pixel points of the same blood vessel in two adjacent tomographic images are different, whether the blood vessel is an artifact caused by heart beating can be judged.
Optionally, before inputting the tomographic images into the image detection model for processing, image preprocessing may be performed on each tomographic image: cutting the size of the tomographic image into the size which can be processed by the image detection model, adjusting the window width and the window level of the tomographic image to reach a proper image angle, and then carrying out normalization and standardization operation on the adjusted tomographic image so that each tomographic image is converted into a standard image.
S103, displaying the detection result on a display interface of the medical image.
Specifically, after the above detection result is obtained by the computer device, the detection result may be displayed on a display interface when a technician takes a film, where the technician may display the captured medical image in a 2D or 3D view manner when taking the film, where the 2D view manner is to tile and display each tomographic image, and the 3D view manner is to display the whole CT image in a three-dimensional image. Then, the computer device may display the detection result of each tomographic image on the 2D view interface or may display the detection result on the 3D view interface, so as to remind the technician of the current shooting problem and correct the shooting operation. Optionally, the computer device may also store the obtained detection result in a database for later invoking in quality statistics.
Optionally, the computer device may label the detection result at a corresponding position of the medical image for display (see fig. 1a, where the rectangular frame is a foreign object labeling frame), or may display the detection result in a list form or a text form, which is not limited in this embodiment.
Alternatively, the medical images may be medical images already captured and stored in the image archiving and communication system (Picture Archiving and Communication Systems, PACS), and then the detection result of each medical image is obtained through the image detection model and then stored in a database corresponding to the medical images.
According to the quality control method for the medical image, the computer equipment firstly obtains the medical image comprising a plurality of tomographic images, then inputs the tomographic images into the image detection model, extracts the characteristics of each tomographic image through the image detection model, obtains the detection result of each tomographic image according to the continuity characteristic information among each tomographic image, and displays the detection result on the display interface of the medical image. In the method, the image detection model detects the medical image according to the continuity characteristic information among each tomographic image, so that the accuracy of a detection result can be improved; and the detection result is displayed to the corresponding technician in time after being obtained, so that the technician can learn the problems existing in shooting and correct the problems in time, and the quality of the obtained medical image can be greatly improved. Furthermore, the process of obtaining the medical image detection result by the method is executed by the image detection model, manual participation is not needed, and the accuracy and the efficiency of quality control are greatly improved.
Optionally, in some of these embodiments, the computer device may classify the medical image in addition to detecting the captured medical image. As shown in fig. 2, the method further includes:
S201, inputting a plurality of tomographic images into an image classification model to obtain a classification result of each tomographic image; the image classification model is a two-dimensional network model, and the classification result represents whether the acquisition range of the medical image is complete or not.
S202, displaying the classification result on a display interface of the medical image.
Specifically, the image classification model is a two-dimensional network model, and then the image classification model is used for processing each tomographic image respectively to obtain a classification result of each tomographic image, and at this time, the continuity information between the adjacent layer images can be not considered. The classification result can represent whether the acquisition range of the shot medical image is complete, for example, if a lung image is to be shot, the medical image needs to contain the image of the whole lung, and if the acquisition is not complete, the quality of the medical image is considered to be unqualified.
Alternatively, the image classification model may be a neural network model, such as ResNet, denseNet or SENet model, or may be another machine learning model. The image classification model can also comprise a series of convolution layers, pooling layers and full-connection layers, the feature extraction is carried out on each tomographic image through the convolution layers and the pooling layers, the feature information of each tomographic image can be obtained, and then the full-connection layers can obtain the classification result of each tomographic image according to the feature information of each tomographic image. The method and the manner for displaying the classification result can be referred to as a method and a manner for displaying the detection result, and are not described herein.
Optionally, for the classification result of the obtained medical image, the computer device may also store the classification result, for example, in a database, for later recall in the quality statistics analysis.
According to the quality control method for the medical image, the computer equipment inputs a plurality of tomographic images into the image classification model, and classification results of each tomographic image are obtained and displayed. In the method, the image classification model outputs the detection result and then displays the detection result to the corresponding technician in time, so that the technician can learn the problems existing in shooting and correct the problems in time, and the quality of the obtained medical image can be greatly improved. Furthermore, the process of obtaining the medical image classification result by the method is executed by the image classification model, manual participation is not needed, and the accuracy and the efficiency of quality control are greatly improved.
However, in practical applications, tens or hundreds of tomographic images are obtained for each medical image photographing, and if all tomographic images are used as a set of input image detection models, the model processing efficiency is reduced, and then the computer device may group all tomographic images first and then process them in groups. Fig. 3 is a flowchart of a quality control method for medical images according to still another embodiment. The present embodiment relates to a specific procedure in which a computer device groups a plurality of tomographic images and inputs the images into an image detection model. Based on the above embodiment, optionally, as shown in fig. 3, S102 may include:
S301, the number of each tomographic image is acquired.
Specifically, each tomographic image has a unique number, and typically, the medical image is stored in a digital imaging and communication (Digital Imaging and Communications in Medicine, DICOM) file, including an inspection identifier of the medical image, a sequence identifier of the medical image, a series identifier of the medical image, and a file identifier of the medical image.
S302, dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the n-1 group and the number of the first tomographic image in the n-1 group have a preset interval.
Specifically, the computer device may divide the preset number of tomograms with consecutive numbers into the same group, for example, if 9 tomograms are provided, the numbers are 001, 002, 003..009, and the preset number of tomograms in each divided group is 3, 4 groups of tomograms are obtained: (001, 002, 003), (003, 004, 005), (005, 006, 007), (007, 008, 009). The number of the first tomogram in the n-1 group and the number of the first tomogram in the n-1 group have a preset interval, and as in the above example, the interval between the number of the first tomogram in the n-1 group and the number of the first tomogram in the n-1 group is 2.
Wherein the above-mentioned preset number and preset interval may be set according to an actual image type, which may be set by a technician according to experience and continuity information of each tomographic image.
S303, each group of tomographic images is input to the image detection model.
Specifically, after the computer device obtains a plurality of groups of tomographic images, each group of tomographic images may be input into the image detection model, so as to obtain a detection result of each tomographic image.
Optionally, the image detection model is a three-dimensional network model, and the detection result may include an artifact detection result and/or a foreign object detection result, as shown in fig. 3a, the computer device inputs each group of tomographic images into the image detection model, and the image detection model obtains the detection result of each tomographic image according to the continuity feature information between each tomographic image, where the step includes:
s303a, obtaining a labeling result of each tomographic image by utilizing the continuous characteristic information among the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result.
Specifically, because whether the detection is an artifact or a foreign object can be determined by continuously analyzing a plurality of front and back tomographic images, the computer equipment can analyze the continuous characteristic information among each tomographic image in each group of tomographic images to obtain the labeling result of each tomographic image. The labeling result comprises a detection frame of the artifact and/or the foreign object, and an artifact classification result and a foreign object classification result in the detection frame. Alternatively, the artifact classification result and the foreign object classification result may be represented by probabilities that each pixel point in the detection frame belongs to an artifact and a foreign object.
And S303b, comparing the artifact classification result with a first preset threshold value to determine an artifact detection result.
S303c, comparing the foreign matter classification result with a second preset threshold value to determine a foreign matter detection result.
Specifically, a probability threshold may be set for whether a pixel point in a medical image belongs to an artifact or a foreign object, where the probability threshold of the artifact is referred to as a first preset threshold, and the probability threshold of the foreign object is referred to as a second preset threshold. Comparing the artifact classification result with a first preset threshold value, if the artifact classification result is greater than or equal to the first preset threshold value, determining that the artifact detection result in the detection frame is artifact, otherwise, determining that the artifact detection result is not artifact; and comparing the foreign matter classification result with a second preset threshold value, if the foreign matter classification result is greater than or equal to the second preset threshold value, determining that the foreign matter detection result in the detection frame is foreign matter, otherwise, determining that the foreign matter detection result is not foreign matter.
According to the quality control method for the medical images, the computer equipment divides the serial number of the preset number of the tomographic images into the same group, so that a plurality of groups of tomographic images are obtained, an image detection model is input, and a detection result of each tomographic image can be obtained. By grouping the tomographic images, the image detection model can process one group of tomographic images at a time, so that the processing efficiency can be further improved; and the continuity characteristic information among a group of internal tomographic images is considered, so that the obtained detection result is more accurate than the continuity characteristic information among all tomographic images, and the efficiency and the accuracy of the quality control process are improved.
Optionally, the classification result includes a collection range classification result, that is, whether the collection range is complete; then, as shown in fig. 4, the above-mentioned inputting a plurality of tomographic images into the image classification model, obtaining a classification result of each tomographic image includes:
s401, sampling a plurality of tomographic images at intervals, and inputting each sampled tomographic image into an image classification model to obtain a classification value of each sampled tomographic image.
Specifically, the computer device may sample the plurality of tomographic images at intervals to obtain a plurality of sampled tomographic images, for example, 9 tomographic images in the above example, may obtain (001, 003, 005, 007, 009) 5 tomographic images at intervals by sampling one time, may obtain (001, 004, 007) 3 tomographic images at intervals by sampling 2 times, and the specific number of intervals may be set according to the layer thickness of the tomographic images. And then inputting each sampling tomographic image into the image classification model to obtain a classification value of each sampling tomographic image, wherein the classification value can be expressed by the probability that the image belongs to the complete acquisition range.
S402, acquiring adjacent tomographic images of the sampled tomographic images with the classification values larger than a third preset threshold, and inputting the adjacent tomographic images into an image classification model to obtain classification values of the adjacent tomographic images.
Specifically, a probability threshold can be set for whether the pixel points in the medical image belong to the complete acquisition range, the probability threshold with the complete acquisition range is called a third preset threshold, then the classification value is compared with the third preset threshold, if the classification value is larger than the third preset threshold, the current fault image acquisition range is determined to be incomplete, otherwise, the current fault image acquisition range is determined to be complete. Because the adjacent tomographic images have the continuous characteristic information, if the current tomographic image is not complete in the acquisition range, the adjacent tomographic image can be considered to be also not complete in the acquisition range, and if the current tomographic image is complete in the acquisition range, the adjacent tomographic image can be considered to be also complete in the acquisition range.
Therefore, next, the adjacent tomograms of the sampled tomograms (i.e., tomograms with incomplete acquisition ranges) with classification values greater than the third preset threshold are input into the image classification model again to obtain classification values of the adjacent tomograms. And the adjacent tomograms of the sampling tomograms (namely, the tomograms with complete acquisition ranges) with the classification values smaller than or equal to the third preset threshold value are not needed to be judged, so that the calculation amount of an image classification model can be reduced, and the image classification processing efficiency is improved.
S403, determining the acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold.
Specifically, according to the classification value of the adjacent tomograms and the third preset threshold value, whether the acquisition range of the adjacent tomograms is complete or not can be determined, if the acquisition range is not complete and the adjacent tomograms are still not judged, the adjacent tomograms are continuously acquired and are input into the image classification model for processing until all the tomograms with the incomplete acquisition range are obtained. And those tomographic images in the image classification model are not input, but are images with complete acquisition ranges, so that the acquisition range classification result of each tomographic image can be obtained.
According to the quality control method for the medical image, the computer equipment performs interval sampling on a plurality of tomographic images, inputs each sampled tomographic image into the image classification model to obtain the classification value of each sampled tomographic image, inputs the adjacent tomographic image of the sampled tomographic image with the classification value larger than the third preset threshold into the image classification model to obtain the classification value of the adjacent tomographic image, and therefore the acquisition range classification result of each tomographic image can be determined. In the method, each tomographic image is not required to be input into the image classification model for classification, so that the classification efficiency of the image classification model can be improved, and the quality control efficiency can be further improved.
In addition, the image detection model and the image classification model need to be model trained before being used, and then the method further comprises the following steps: inputting the first training sample image into an initial image detection model to obtain an initial artifact detection result and/or an initial foreign matter detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions; inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain an image classification model meeting preset conditions.
Specifically, before the training sample image is input into the initial image detection model or the initial image classification model, the image preprocessing mode in the above embodiment may be used to preprocess the training sample image, and after preprocessing is completed, a doctor with rich experience may label the artifact and/or foreign object detection frame and classification in the first training sample image, and label whether the second training sample image is a gold standard in training. The training sample image can also comprise normal medical images so as to ensure the diversity of training data. Optionally, the computer device may further add random disturbance to the training sample image, and select, within a certain range, a voxel value on a voxel cube adjacent to the current voxel point as the current value, so as to improve robustness of the model.
And then, the computer equipment inputs the first training sample image into an initial image detection model, compares the output detection frame and detection classification with the marked artifact and/or foreign object detection frame and classification to calculate loss, and updates the initial image detection model by utilizing the loss back gradient propagation, so as to obtain a training convergence image detection model through a continuous iteration process. And inputting the second training sample image into the initial image classification model, comparing the output acquisition range classification result with the marked acquisition range to calculate loss, updating the initial image classification model by utilizing the loss back gradient propagation, and obtaining the training convergence image classification model through continuous iterative process.
Optionally, in some embodiments, after the computer device obtains the artifact detection frame and/or the foreign object detection frame, the computer device may obtain corresponding coordinates according to the position of the artifact detection frame and/or the foreign object detection frame in the tomographic image, and then store the coordinates of the artifact detection frame and/or the foreign object detection frame, the serial numbers of the tomographic images including the artifact and/or the foreign object, and store the serial numbers of the tomographic images with incomplete acquisition range.
Optionally, the computer device may perform statistics on the numbers of the stored tomographic images according to the statistics requirement, so as to obtain statistics results of the medical images, such as statistics of medical images including artifacts or medical images including foreign matters or medical images with incomplete collection range, and display the statistics results to a technician, so as to know shooting problems of different technicians, so as to correct shooting operations of the technicians, and improve quality of the obtained medical images.
It should be understood that, although the steps in the flowcharts of fig. 1-3, 3a and 4 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps of fig. 1-3, 3a, and 4 may include multiple sub-steps or phases that are not necessarily performed at the same time, but may be performed at different times, nor does the order in which the sub-steps or phases are performed necessarily occur in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or phases of other steps.
Fig. 5 is a schematic structural diagram of a quality control device for medical images according to an embodiment. As shown in fig. 5, the apparatus includes: an acquisition module 11, a processing module 12 and a display module 13.
Specifically, the acquiring module 11 is configured to acquire a medical image to be detected, where the medical image includes a plurality of tomographic images.
The processing module 12 is configured to input a plurality of tomographic images into the image detection model, perform feature extraction on each tomographic image through the image detection model, and obtain a detection result of each tomographic image according to the continuity feature information between each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image.
And the display module 13 is used for displaying the detection result on a display interface of the medical image.
The quality control device for medical images provided in this embodiment may execute the above method embodiments, and its implementation principle and technical effects are similar, and will not be described herein.
In one embodiment, the processing module 12 is further configured to input a plurality of tomographic images into the image classification model to obtain a classification result of each tomographic image; the image classification model is a two-dimensional network model, and the classification result represents whether the acquisition range of the medical image is complete or not; the display module 13 is further configured to display the classification result on a display interface of the medical image.
In one embodiment, the processing module 12 is specifically configured to acquire a number of each tomographic image; dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval; each set of tomographic images is input to an image detection model.
In one embodiment, the image detection model is a three-dimensional network model, and the detection result includes an artifact detection result and/or a foreign object detection result; the processing module 12 is specifically configured to obtain a labeling result of each tomographic image by using the continuous feature information between the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result; comparing the artifact classification result with a first preset threshold value to determine an artifact detection result; and comparing the foreign matter classification result with a second preset threshold value to determine a foreign matter detection result.
In one embodiment, the classification result comprises an acquisition range classification result; the processing module 12 is specifically configured to sample a plurality of tomograms at intervals, and input each sampled tomogram into the image classification model to obtain a classification value of each sampled tomogram; acquiring adjacent tomographic images of the sampled tomographic images with the classification values larger than a third preset threshold, and inputting the adjacent tomographic images into an image classification model to obtain classification values of the adjacent tomographic images; and determining the acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold value.
In one embodiment, the apparatus further includes a training module, configured to input a first training sample image into an initial image detection model, to obtain an initial artifact detection result and/or an initial foreign object detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions; inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain the image classification model meeting the preset conditions.
In one embodiment, the apparatus further includes a storage module configured to store coordinates of the artifact detection frame and/or the foreign object detection frame and a number of a tomographic image including the artifact and/or the foreign object; and storing the serial numbers of the tomographic images with incomplete acquisition ranges.
In one embodiment, the device further comprises a statistics module, which is used for counting the numbers of the stored tomographic images according to the statistics requirement to obtain the quality statistics result of the medical images; the statistical requirements include at least one of artifact statistical requirements, foreign object statistical requirements, and acquisition range insufficiency statistical requirements.
For specific limitations of the quality control apparatus for medical images, reference may be made to the above limitations of the quality control method for medical images, and no further description is given here. The above-mentioned respective modules in the quality control apparatus for medical images may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a terminal, and the internal structure of which may be as shown in fig. 6. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of quality control of medical images. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in FIG. 6 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
and displaying the detection result on a display interface of the medical image.
The computer device provided in this embodiment has similar implementation principles and technical effects to those of the above method embodiment, and will not be described herein.
In one embodiment, the processor when executing the computer program further performs the steps of:
inputting a plurality of tomographic images into an image classification model to obtain a classification result of each tomographic image; the image classification model is a two-dimensional network model, and the classification result represents whether the acquisition range of the medical image is complete or not;
and displaying the classification result on a display interface of the medical image.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring the number of each tomographic image;
dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval;
each set of tomographic images is input to an image detection model.
In one embodiment, the image detection model is a three-dimensional network model, and the detection results include artifact detection results and/or foreign object detection results; the processor when executing the computer program also implements the steps of:
obtaining the labeling result of each tomographic image by utilizing the continuity characteristic information among the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result;
Comparing the artifact classification result with a first preset threshold value to determine an artifact detection result;
and comparing the foreign matter classification result with a second preset threshold value to determine a foreign matter detection result.
In one embodiment, the classification result comprises an acquisition range classification result; the processor when executing the computer program also implements the steps of:
performing interval sampling on a plurality of tomographic images, and inputting each sampled tomographic image into an image classification model to obtain a classification value of each sampled tomographic image;
acquiring adjacent tomographic images of the sampled tomographic images with the classification values larger than a third preset threshold, and inputting the adjacent tomographic images into an image classification model to obtain classification values of the adjacent tomographic images;
and determining the acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold value.
In one embodiment, the processor when executing the computer program further performs the steps of:
inputting the first training sample image into an initial image detection model to obtain an initial artifact detection result and/or an initial foreign matter detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions;
Inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain the image classification model meeting the preset conditions.
In one embodiment, the processor when executing the computer program further performs the steps of:
storing coordinates of the artifact detection frame and/or the foreign object detection frame and numbers of tomographic images including the artifact and/or the foreign object;
and storing the serial numbers of the tomographic images with incomplete acquisition ranges.
In one embodiment, the processor when executing the computer program further performs the steps of:
counting the numbers of the stored tomographic images according to the counting requirement to obtain a quality counting result of the medical image; the statistical requirements include at least one of artifact statistical requirements, foreign object statistical requirements, and acquisition range insufficiency statistical requirements.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
Inputting a plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
and displaying the detection result on a display interface of the medical image.
The computer readable storage medium provided in this embodiment has similar principles and technical effects to those of the above method embodiment, and will not be described herein.
In one embodiment, the computer program when executed by the processor further performs the steps of:
inputting a plurality of tomographic images into an image classification model to obtain a classification result of each tomographic image; the image classification model is a two-dimensional network model, and the classification result represents whether the acquisition range of the medical image is complete or not;
and displaying the classification result on a display interface of the medical image.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring the number of each tomographic image;
dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval;
Each set of tomographic images is input to an image detection model.
In one embodiment, the image detection model is a three-dimensional network model, and the detection results include artifact detection results and/or foreign object detection results; the computer program when executed by the processor also performs the steps of:
obtaining the labeling result of each tomographic image by utilizing the continuity characteristic information among the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result;
comparing the artifact classification result with a first preset threshold value to determine an artifact detection result;
and comparing the foreign matter classification result with a second preset threshold value to determine a foreign matter detection result.
In one embodiment, the classification result comprises an acquisition range classification result; the computer program when executed by the processor also performs the steps of:
performing interval sampling on a plurality of tomographic images, and inputting each sampled tomographic image into an image classification model to obtain a classification value of each sampled tomographic image;
acquiring adjacent tomographic images of the sampled tomographic images with the classification values larger than a third preset threshold, and inputting the adjacent tomographic images into an image classification model to obtain classification values of the adjacent tomographic images;
And determining the acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold value.
In one embodiment, the computer program when executed by the processor further performs the steps of:
inputting the first training sample image into an initial image detection model to obtain an initial artifact detection result and/or an initial foreign matter detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions;
inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain the image classification model meeting the preset conditions.
In one embodiment, the computer program when executed by the processor further performs the steps of:
storing coordinates of the artifact detection frame and/or the foreign object detection frame and numbers of tomographic images including the artifact and/or the foreign object;
And storing the serial numbers of the tomographic images with incomplete acquisition ranges.
In one embodiment, the computer program when executed by the processor further performs the steps of:
counting the numbers of the stored tomographic images according to the counting requirement to obtain a quality counting result of the medical image; the statistical requirements include at least one of artifact statistical requirements, foreign object statistical requirements, and acquisition range insufficiency statistical requirements.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.
Claims (10)
1. A method for quality control of medical images, comprising:
acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
inputting the plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
Displaying the detection result on a display interface of the medical image;
performing interval sampling on the plurality of tomographic images, and inputting each sampled tomographic image into an image classification model to obtain a classification value of each sampled tomographic image;
acquiring adjacent tomograms of the sampled tomograms with classification values larger than a third preset threshold, and inputting the adjacent tomograms into an image classification model to obtain classification values of the adjacent tomograms;
determining an acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold; the collection range classification result comprises whether the collection range is complete or not;
displaying the classification result on a display interface of the medical image;
wherein the determining the collection range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold value includes:
determining whether the acquisition range of the adjacent tomograms is complete according to the classification value of the adjacent tomograms and the third preset threshold value; if the acquisition range of the adjacent tomograms is incomplete and the adjacent tomograms are not judged, continuously acquiring the adjacent tomograms and inputting the adjacent tomograms into the image classification model for processing.
2. The method of claim 1, wherein the inputting the plurality of tomographic images into an image detection model comprises:
acquiring the number of each tomographic image;
dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval;
each set of tomographic images is input to the image detection model.
3. The method according to claim 2, wherein the image detection model is a three-dimensional network model, and the detection results include artifact detection results and/or foreign object detection results; and obtaining a detection result of each tomographic image according to the continuous characteristic information between each tomographic image, wherein the detection result comprises the following steps:
obtaining the labeling result of each tomographic image by utilizing the continuous characteristic information among the tomographic images in each group of tomographic images; the labeling result comprises a detection frame, an artifact classification result and a foreign object classification result;
comparing the artifact classification result with a first preset threshold value to determine the artifact detection result;
And comparing the foreign matter classification result with a second preset threshold value to determine the foreign matter detection result.
4. The method according to claim 1, wherein the method further comprises:
inputting the first training sample image into an initial image detection model to obtain an initial artifact detection result and/or an initial foreign matter detection result; training the initial image detection model according to the loss between the initial artifact detection result and the artifact label and the loss between the initial foreign object detection result and the foreign object label to obtain an image detection model meeting preset conditions;
inputting the second training sample image into an initial image classification model to obtain an initial acquisition range classification result; and training the initial image classification model according to the initial acquisition range classification result and the loss between the acquisition range labels to obtain an image classification model meeting preset conditions.
5. The method according to claim 2, wherein the method further comprises:
storing coordinates of the artifact detection frame and/or the foreign object detection frame and numbers of tomographic images including the artifact and/or the foreign object;
and storing the serial numbers of the tomographic images with incomplete acquisition ranges.
6. The method of claim 5, wherein the method further comprises:
counting the numbers of the stored tomographic images according to the counting requirement to obtain a quality counting result of the medical image; the statistical requirements include at least one of artifact statistical requirements, foreign object statistical requirements, and acquisition range non-complete statistical requirements.
7. A quality control apparatus for medical images, comprising:
the acquisition module is used for acquiring a medical image to be detected, wherein the medical image comprises a plurality of tomographic images;
the processing module is used for inputting the plurality of tomographic images into an image detection model, extracting the characteristics of each tomographic image through the image detection model, and obtaining the detection result of each tomographic image according to the continuity characteristic information among each tomographic image; the image detection model is a machine learning model, and the detection result represents the quality result of the medical image;
the display module is used for displaying the detection result on a display interface of the medical image;
the processing module is specifically configured to sample the plurality of tomograms at intervals, and input each sampled tomogram into an image classification model to obtain a classification value of each sampled tomogram; acquiring adjacent tomograms of the sampled tomograms with classification values larger than a third preset threshold, and inputting the adjacent tomograms into an image classification model to obtain classification values of the adjacent tomograms; determining an acquisition range classification result of each tomographic image according to the classification value of each sampling tomographic image, the classification value of the adjacent tomographic image and the third preset threshold; the collection range classification result comprises whether the collection range is complete or not;
The display module is further used for displaying the classification result on a display interface of the medical image;
the processing module is specifically configured to determine whether the acquisition range of the adjacent tomograms is complete according to the classification value of the adjacent tomograms and the third preset threshold; if the acquisition range of the adjacent tomograms is incomplete and the adjacent tomograms are not judged, continuously acquiring the adjacent tomograms and inputting the adjacent tomograms into the image classification model for processing.
8. The quality control device of claim 7, wherein,
the processing module is specifically used for acquiring the serial number of each tomographic image; dividing the serial number of the preset number of the tomographic images into the same group to obtain a plurality of groups of tomographic images; the number of the first tomographic image in the nth group and the number of the first tomographic image in the n-1 th group have a preset interval; each set of tomographic images is input to the image detection model.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1-6 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911221084.0A CN111080584B (en) | 2019-12-03 | 2019-12-03 | Quality control method for medical image, computer device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911221084.0A CN111080584B (en) | 2019-12-03 | 2019-12-03 | Quality control method for medical image, computer device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111080584A CN111080584A (en) | 2020-04-28 |
| CN111080584B true CN111080584B (en) | 2023-10-31 |
Family
ID=70312588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911221084.0A Active CN111080584B (en) | 2019-12-03 | 2019-12-03 | Quality control method for medical image, computer device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111080584B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111640093A (en) * | 2020-05-20 | 2020-09-08 | 上海联影智能医疗科技有限公司 | Quality control method for medical image and computer-readable storage medium |
| CN111798439A (en) * | 2020-07-11 | 2020-10-20 | 大连东软教育科技集团有限公司 | Medical image quality interpretation method and system for online and offline fusion and storage medium |
| CN112767307A (en) * | 2020-12-28 | 2021-05-07 | 上海联影智能医疗科技有限公司 | Image processing method, image processing device, computer equipment and storage medium |
| CN112652031A (en) * | 2020-12-30 | 2021-04-13 | 上海联影智能医疗科技有限公司 | Artifact correction analysis method, electronic device and storage medium |
| CN112862778A (en) * | 2021-02-04 | 2021-05-28 | 上海联影智能医疗科技有限公司 | Scanning image detection method, electronic device and computer readable medium |
| CN113033650A (en) * | 2021-03-22 | 2021-06-25 | Oppo广东移动通信有限公司 | Image classification method, training method and device of classification model and storage medium |
| CN113299371B (en) * | 2021-07-05 | 2022-04-26 | 数坤(北京)网络科技股份有限公司 | Medical image display method and device, computer equipment and storage medium |
| CN114566263A (en) * | 2022-04-24 | 2022-05-31 | 杭州太美星程医药科技有限公司 | Marking method, device and system for medical examination image sequence and computer equipment |
| CN114742836B (en) * | 2022-06-13 | 2022-09-09 | 浙江太美医疗科技股份有限公司 | Medical image processing method and device and computer equipment |
| CN116777900A (en) * | 2023-07-31 | 2023-09-19 | 广州医科大学附属第一医院（广州呼吸中心） | Medical image quality scoring method, system and readable storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09299366A (en) * | 1996-05-10 | 1997-11-25 | Fujitsu Ltd | Region extract device |
| JPH11306264A (en) * | 1998-04-27 | 1999-11-05 | Toshiba Iyo System Engineering Kk | Computer aided diagnostic device |
| CN107301640A (en) * | 2017-06-19 | 2017-10-27 | 太原理工大学 | A kind of method that target detection based on convolutional neural networks realizes small pulmonary nodules detection |
| WO2018122632A1 (en) * | 2016-12-26 | 2018-07-05 | Abb Schweiz Ag | A method for detecting inrush and ct saturation and an inteligent electronic device therfor |
| CN109741316A (en) * | 2018-12-29 | 2019-05-10 | 成都金盘电子科大多媒体技术有限公司 | Medical image intelligently comments piece system |
| DE102017221297A1 (en) * | 2017-11-28 | 2019-05-29 | Siemens Healthcare Gmbh | Method and device for the automated evaluation of at least one image data set recorded with a medical image recording device, computer program and electronically readable data carrier |
| CN110232688A (en) * | 2019-07-12 | 2019-09-13 | 上海联影智能医疗科技有限公司 | Medical image system, method, computer equipment and readable storage medium storing program for executing |
| CN110264460A (en) * | 2019-06-24 | 2019-09-20 | 科大讯飞股份有限公司 | A kind of discrimination method of object detection results, device, equipment and storage medium |
| CN110458817A (en) * | 2019-08-05 | 2019-11-15 | 上海联影医疗科技有限公司 | Qualitative forecasting method, device, equipment and the storage medium of medical image |
| CN110473226A (en) * | 2019-07-18 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Training method, computer equipment and the readable storage medium storing program for executing of image processing network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10489907B2 (en) * | 2017-11-13 | 2019-11-26 | Siemens Healthcare Gmbh | Artifact identification and/or correction for medical imaging |
-
2019
- 2019-12-03 CN CN201911221084.0A patent/CN111080584B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09299366A (en) * | 1996-05-10 | 1997-11-25 | Fujitsu Ltd | Region extract device |
| JPH11306264A (en) * | 1998-04-27 | 1999-11-05 | Toshiba Iyo System Engineering Kk | Computer aided diagnostic device |
| WO2018122632A1 (en) * | 2016-12-26 | 2018-07-05 | Abb Schweiz Ag | A method for detecting inrush and ct saturation and an inteligent electronic device therfor |
| CN107301640A (en) * | 2017-06-19 | 2017-10-27 | 太原理工大学 | A kind of method that target detection based on convolutional neural networks realizes small pulmonary nodules detection |
| DE102017221297A1 (en) * | 2017-11-28 | 2019-05-29 | Siemens Healthcare Gmbh | Method and device for the automated evaluation of at least one image data set recorded with a medical image recording device, computer program and electronically readable data carrier |
| CN109741316A (en) * | 2018-12-29 | 2019-05-10 | 成都金盘电子科大多媒体技术有限公司 | Medical image intelligently comments piece system |
| CN110264460A (en) * | 2019-06-24 | 2019-09-20 | 科大讯飞股份有限公司 | A kind of discrimination method of object detection results, device, equipment and storage medium |
| CN110232688A (en) * | 2019-07-12 | 2019-09-13 | 上海联影智能医疗科技有限公司 | Medical image system, method, computer equipment and readable storage medium storing program for executing |
| CN110473226A (en) * | 2019-07-18 | 2019-11-19 | 上海联影智能医疗科技有限公司 | Training method, computer equipment and the readable storage medium storing program for executing of image processing network |
| CN110458817A (en) * | 2019-08-05 | 2019-11-15 | 上海联影医疗科技有限公司 | Qualitative forecasting method, device, equipment and the storage medium of medical image |
Non-Patent Citations (4)
| Title |
|---|
| Fast Convolutional Neural Network Training Using Selective Data Sampling: Application to Hemorrhage Detection in Color Fundus Images;Mark J. J. P. van Grinsven et al.;《 IEEE Transactions on Medical Imaging 》;第35卷(第5期);第1273-1283页 * |
| YW Wang ; et al..An Efficient Segmentation Framework for Continuous CT Images.《Journal of Image Processing and Communication》.2010,全文. * |
| 连续断层图像立体重采样评估;肖嵛等;《第三军医大学学报》;第29卷(第3期);第193-195页 * |
| 陈蕾蕾等.PET/CT显像常见伪影及质量控制.《中国医疗设备》.2019,第34卷(第02期),第17-24页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111080584A (en) | 2020-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111080584B (en) | Quality control method for medical image, computer device and readable storage medium | |
| CN109493328B (en) | Medical image display method, viewing device and computer device | |
| CN111160367B (en) | Image classification method, apparatus, computer device, and readable storage medium | |
| CN109308728B (en) | Positron emission computed tomography image processing method and device | |
| US10825149B2 (en) | Defective pixel correction using adversarial networks | |
| CN103222876B (en) | Medical image-processing apparatus, image diagnosing system, computer system and medical image processing method | |
| CN111340825B (en) | Method and system for generating mediastinum lymph node segmentation model | |
| US20210312629A1 (en) | Methods, systems and apparatus for processing medical chest images | |
| CN103514597A (en) | Image processing device | |
| CN111540025B (en) | Predicting images for image processing | |
| US10303965B2 (en) | Defective pixel identification using machine learning | |
| CN104125841A (en) | Control method and control system | |
| CN111080583B (en) | Medical image detection method, computer device, and readable storage medium | |
| CN110458837B (en) | Image post-processing method and device, electronic equipment and storage medium | |
| US20220101034A1 (en) | Method and system for segmenting interventional device in image | |
| CN111179366A (en) | Low-dose image reconstruction method and system based on anatomical difference prior | |
| CN113362272A (en) | Medical image segmentation with uncertainty estimation | |
| US20210398329A1 (en) | Artificial intelligence (ai)-based standardized uptake vaule (suv) correction and variation assessment for positron emission tomography (pet) | |
| CN111128348B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| CN110751187A (en) | Training method of abnormal area image generation network and related product | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| CN110473226B (en) | Training method of image processing network, computer device and readable storage medium | |
| CN113192031B (en) | Vascular analysis method, vascular analysis device, vascular analysis computer device, and vascular analysis storage medium | |
| CN110992312B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| CN111260636B (en) | Model training method and device, image processing method and device, and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |