CN114612710A - Image detection method, image detection device, computer equipment and storage medium - Google Patents
Image detection method, image detection device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN114612710A CN114612710A CN202210199037.6A CN202210199037A CN114612710A CN 114612710 A CN114612710 A CN 114612710A CN 202210199037 A CN202210199037 A CN 202210199037A CN 114612710 A CN114612710 A CN 114612710A
- Authority
- CN
- China
- Prior art keywords
- classification
- detection result
- image
- medical image
- result
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The application relates to an image detection method, an image detection device, computer equipment and a storage medium, wherein a first classification result of a medical image is obtained by inputting the medical image to be processed into a preset first classification model; the first classification result is used for indicating whether a first focus exists in the medical image; inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image; and determining a target detection result of the medical image according to the first classification result and the detection result. According to the method and the device, when the target detection result is determined, various focuses are detected simultaneously, so that the focuses influencing the target detection result are filtered out when the target detection result is determined, and the accuracy of the target detection result is improved.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to an image detection method, an image detection apparatus, a computer device, and a storage medium.
Background
With the development of modern medical technology, artificial intelligence is increasingly applied to medicine.
The prior art provides a lung nodule detection algorithm, which includes identifying suspected nodules in a lung image and classifying each nodule to achieve the purpose of accurately identifying all nodules in the lung image.
However, in this detection algorithm, each nodule is judged and identified only from the nodule level of the lung image, and thus, an identification error is likely to occur, and the true nodule is erroneously determined as a false positive.
Disclosure of Invention
In view of the above, it is necessary to provide an image detection method, an apparatus, a computer device, and a storage medium capable of improving accuracy in view of the above technical problems.
In a first aspect, the present application provides an image detection method, including:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In one embodiment, the first classification model includes a first region classification model and a second region classification model, and the inputting the medical image to be processed into the preset first classification model to obtain the first classification result of the medical image includes:
carrying out image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into a second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the medical image is input into a preset detection model to obtain a detection result of the medical image, and the detection result comprises the following steps:
and inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In one embodiment, determining a target detection result of the medical image according to the first classification result and the detection result includes:
and if the first classification result indicates that the first focus does not exist in the medical image, taking the detection result as a target detection result.
In one embodiment, determining a target detection result of the medical image according to the first classification result and the detection result includes:
if the first classification result indicates that the first region and the second region in the medical image both have the first focus, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result.
In one embodiment, determining a target detection result of the medical image according to the first classification result and the detection result includes:
if the first classification result indicates that the first region or the second region in the medical image has the first focus, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result.
In one embodiment, determining the target detection result according to the second classification result and the detection result includes:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into a second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
In one embodiment, determining the target detection result according to the second classification result and the detection result includes:
if the first classification result indicates that the second region has the first focus and the first region does not have the first focus, inputting a detection result corresponding to the second image into the second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
In a second aspect, the present application also provides an image detection apparatus, comprising:
the classification module is used for inputting the medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
the detection module is used for inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and the determining module is used for determining a target detection result of the medical image according to the first classification result and the detection result.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the following steps when executing the computer program:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In a fifth aspect, the present application further provides a computer program product. The computer program product comprising a computer program which when executed by a processor performs the steps of:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
According to the image detection method, the image detection device, the computer equipment and the storage medium, the medical image to be processed is input into the preset first classification model, so that a first classification result of the medical image is obtained; the first classification result is used for indicating whether a first focus exists in the medical image; inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image; and determining a target detection result of the medical image according to the first classification result and the detection result. According to the method and the device, when the target detection result is determined, various focuses are detected simultaneously, so that the focuses influencing the target detection result are filtered out when the target detection result is determined, and the accuracy of the target detection result is improved.
Drawings
FIG. 1 is a flow diagram illustrating an exemplary image detection method;
FIG. 2 is a flow chart illustrating an image detection method according to another embodiment;
FIG. 3 is a flowchart illustrating a method for determining a target detection result of a medical image according to a first classification result and a detection result by a terminal in one embodiment;
FIG. 4 is a flow diagram of an embodiment of a false positive removal algorithm;
FIG. 5 is a schematic representation of lung nodule false positives;
FIG. 6 is a schematic illustration of left and right lung samples taken;
FIG. 7 is a schematic diagram of an application of the pneumonia classification model;
FIG. 8 is a schematic of pulmonary nodule extraction;
FIG. 9 is a block diagram showing the structure of an image detection apparatus according to an embodiment;
FIG. 10 is a diagram showing an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In the prior art, a focus detection algorithm based on image processing generally performs focus detection according to features of a focus to be identified. However, a large number of false positives often occur in existing lesion detection algorithms, for example, lung nodule detection algorithms are sensitive to inflammatory data, particularly new crown pneumonia, tuberculosis, emphysema, other pneumonia, and the like, so that a large number of false positives are easily detected, because different types of lesions have relevance to the same part, when one lesion is detected, the lesion is easily affected by other types of lesions, and thus false positives occur. However, the occurrence of a large number of false positives can cause interference to the user.
The technical scheme of the application provides an image detection method, the image detection method combines a first classification result and a detection result to determine a target detection result, wherein the first classification result indicates whether a first focus exists in a medical image, and the detection result indicates whether a second focus exists in the medical image.
In an embodiment, as shown in fig. 1, an image detection method is provided, and this embodiment is illustrated by applying this method to a terminal, and it is to be understood that this method may also be applied to a server, and may also be applied to a system including a terminal and a server, and is implemented by interaction between the terminal and the server. In this embodiment, the method includes the steps of:
step 101, inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image.
The medical image to be processed may be, for example, a CT scan image or an image acquired in other ways. Wherein the medical image may be, for example, a lung image, a brain image, a liver image, a uterus and appendages image, etc., and is not limited herein.
The preset first classification model may be a pre-trained neural network model, a machine learning model, or the like. Wherein the first classification model is used for classifying whether the first focus exists in the medical image. The first classification result is used to indicate whether a first lesion is present in the medical image.
Optionally, the training process of the first classification model includes the following steps:
collecting the sample image containing the first focus as a classification positive sample, and collecting the sample image not containing the first focus as a classification negative sample. Wherein, the initial network model of the first classification model selects 3D ResidualNet (Chinese: deep three-dimensional residual error neural network), and adopts Focal Loss as a Loss function. And inputting the classification positive sample and the classification negative sample into the initial network model for training to obtain a final first classification model.
Optionally, the first lesion is a lesion that affects detection of a second lesion. The first lesion may be a plurality of lesions of various types, such as lung, and the first lesion may be, for example, pneumonia, emphysema, pneumothorax, and the like.
Alternatively, when the first lesion is a plurality of lesions, a corresponding first classification model may be set corresponding to each lesion. Or the same first classification model may be used to classify the plurality of lesions.
Step 102, inputting the medical image into a preset detection model to obtain a detection result of the medical image.
Wherein the detection result is used to indicate whether a second lesion is present in the medical image.
In an embodiment of the present application, the detection model is used to detect whether a second lesion exists in the medical image, where the second lesion is a lesion to be detected in the present application, and the first lesion and the second lesion correspond to a same organ.
Optionally, the training process of the detection model includes the following steps:
and collecting an image containing a second focus as a training sample, wherein the initial Network model of the detection model selects 3D FPN (English: Feature Pyramid Network; Chinese: Feature Pyramid Network). And inputting the training sample into the initial network model for training to obtain a final detection model.
And 103, determining a target detection result of the medical image according to the first classification result and the detection result.
In the embodiment of the application, the first classification result and the detection result can be fused to obtain the target detection result. The mode of fusing the first classification result and the detection result may be: and carrying out false positive removal on the detection result according to the first classification result.
In the technical scheme of the application, the first classification model is used for classifying whether the first focus exists in the medical image, so that factors which may have adverse effects on the detection of the second focus can be screened out based on the first classification result. Whether the detection result is adversely affected or not can be determined by fusing the first classification result and the detection result, and factors which have adverse effects on the detection of the second focus can be removed by fusing the first classification result and the detection result, so that the accuracy of the target detection result is improved.
On the basis of the foregoing embodiment, the first classification model includes a first region classification model and a second region classification model, and the embodiment of the present application further provides another image detection method, as shown in fig. 2:
step 201, performing image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image, and processing the first image and the second image respectively to obtain a first classification result corresponding to the first image and a first classification result corresponding to the second image.
In the embodiment of the application, when the medical image is a lung image, the first region is a region where a left lung is located, the first image is a left lung image, the second region is a region where a right lung is located, and the second image is a right lung image.
When the medical image is a brain image, the first region is a region where a left brain is located, the first image is a left brain image, the second region is a region where a right brain is located, and the second image is a right brain image.
When the medical image is an abdominal image, the first region may be, for example, a region where a spleen is located, the first image is an image corresponding to the spleen, the second region is a region where a liver is located, and the second image is an image corresponding to the liver.
It should be noted that the first region and the second region in the medical image may be specified manually based on the actual needs of the user, and are not limited to the above example.
The process of respectively processing the first image and the second image to obtain a first classification result corresponding to the first image and a first classification result corresponding to the second image comprises the following contents:
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model; and inputting the second image into the second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model.
The first region classification model is obtained by training based on a sample image corresponding to the first region, and the second region classification model is obtained by training based on a sample image corresponding to the second region. The training process of the first region classification model and the second region classification model is not limited.
Optionally, the first region classification model and the second region classification model may be the same model or may be two models. When the two models are the same model, the number of the models is reduced, thereby saving the video memory and reducing the loading time of the models.
In an optional implementation manner, before the medical image is input to the first classification model, the image segmentation model segments the medical image to obtain a first image and a second image, and then the first image and the second image are input to the first classification model for classification to obtain a first classification result corresponding to the first image and a first classification result corresponding to the second image.
Step 202, inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In the embodiment of the application, after the detection model detects the medical image, the detection result may be output, then, the image segmentation model may perform image segmentation on the medical image to obtain a segmentation result of the first image and a segmentation result of the second image, the detection model determines a detection result corresponding to the first image by comparing the segmentation result and the detection result of the first image, and the detection result corresponding to the second image is obtained by comparing the segmentation result and the detection result of the second image.
The detection result corresponding to the first image is used for indicating whether a second focus exists in the first image, and the detection result corresponding to the second image is used for indicating whether a second focus exists in the second image.
And step 203, determining a target detection result of the medical image according to the first classification result and the detection result.
In this embodiment of the application, the first classification result includes a first classification result corresponding to the first image and a first classification result corresponding to the second image, and the detection result includes a detection result corresponding to the first image and a detection result corresponding to the second image. In the embodiment of the application, the target detection result is determined according to the first classification result corresponding to the first image, the first classification result corresponding to the second image, the detection result corresponding to the first image and the detection result corresponding to the second image.
In the embodiment of the application, a first image corresponding to a first region in a medical image and a second image corresponding to a second region in the medical image are obtained by segmenting the medical image, then a first classification result corresponding to the first image, a detection result corresponding to the first image, a first classification result corresponding to the second image and a detection result corresponding to the second image are obtained respectively, and the medical image is subdivided, so that information contained in the first classification result and the detection result is more specific and detailed, and the accuracy of a target detection result is improved.
In an embodiment of the application, the detection result is determined to be a target detection result if the first classification result indicates that the first lesion is not present in the medical image.
Optionally, the first classification result indicates that there is no first lesion representation in the medical image: the first classification result corresponding to the first image indicates that the first lesion does not exist in the first image, and the first classification result corresponding to the second image indicates that the first lesion does not exist in the second image. In this case, the detection result corresponding to the first image and the detection result corresponding to the second image are determined as the target detection result.
The following description will be given by taking the medical image as a lung image, the first region being a region where the left lung is located, the second region being a region where the right lung is located, the first lesion being pneumonia, and the second lesion being a lung nodule as an example:
if the first classification result indicates that the first focus does not exist in the medical image, i.e. the pneumonia does not exist in the left lung and the right lung, the lung nodules in the left lung and the right lung in the detection result can be used as the target detection result. Therefore, the algorithm can be ensured to only perform false positive removal on the detection result of the infection pneumonia, on one hand, the false positive removal is not performed if no pneumonia focus exists, the calculation resource is saved, on the other hand, the detection result without pneumonia can not be accidentally injured, and the higher accuracy is ensured.
Since the first focus does not exist in the medical image, the visible detection result is not influenced by the first focus, so that the detection result can be considered to be credible.
In an embodiment of the application, if the first classification result indicates that the first focus exists in both the first region and the second region in the medical image, the detection result is input into a preset second classification model, a second classification result output by the second classification model is obtained, and a target detection result is determined according to the second classification result.
The first classification result indicating that the first region and the second region in the medical image both have the first lesion may indicate that the first classification result corresponding to the first image indicates that the first lesion exists in the first image, and the first classification result corresponding to the second image indicates that the first lesion also exists in the second image.
The second classification model may be a false positive removal model, the second classification model may perform reclassification on the input detection result, classify the attribute of the second lesion in the detection result through the second classification model, and determine attribute information of the second lesion, where the attribute information of the second lesion may be used to indicate whether the second lesion is affected by the first lesion, and finally determine the second classification result by determining the attribute of each second lesion, and determine the target detection result according to the second classification result.
Optionally, in this embodiment of the application, if the first classification result indicates that the first region and the second region in the medical image both have the first lesion, the detection result is input into a preset second classification model, so as to obtain a second classification result output by the second classification model, where the detection result may be, for example, a detection result corresponding to the first image and a detection result corresponding to the second image; and determining a target detection result according to the second classification result. Alternatively, the second classification result may be directly determined as the target monitoring result.
The following description still takes the medical image as a lung image, the first region is a region where the left lung is located, the second region is a region where the right lung is located, the first focus is pneumonia, and the second focus is a pulmonary nodule as an example:
if both the left lung and the right lung have pneumonia foci, respectively inputting the detection result of the left lung and the detection result of the right lung into a second classification model, judging the attribute of each nodule in the detection result by the second classification model, namely judging whether each nodule is a pneumonia area, if so, indicating that the detection result is a false positive detection result generated by being influenced by pneumonia (the first focus), and if so, judging that the nodule is false positive, and removing the nodule judged to be false positive to obtain a final target detection result. Therefore, the algorithm can be ensured to only remove false positive from the detection result of the infection pneumonia, and the detection result of the infection without the pneumonia can not be accidentally injured, so that higher accuracy is ensured.
On the basis of the above embodiment, as shown in fig. 3, the method for determining the target detection result of the medical image by the terminal according to the first classification result and the detection result further includes the following steps:
step 301, if the first classification result indicates that the first region or the second region in the medical image has the first lesion, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model.
And step 302, determining a target detection result according to the second classification result and the detection result.
Wherein, the second classification model can be a false positive removing model which is used for reclassifying the detection result of the medical image.
In an embodiment of the present application, the first classification result indicates that the first lesion exists in the first region or the second region in the medical image, and may include the following cases:
the first method comprises the following steps: the first classification result indicates that the first region has a first lesion and the second region does not have the first lesion.
And the second method comprises the following steps: the first classification result indicates that the second region has the first lesion and the first region does not have the first lesion.
For the first case, the first region has a first lesion, the second region does not have a first lesion, and the detection result corresponding to the first image corresponding to the first region is adversely affected by the first lesion, so that it is necessary to perform false positive removal on the detection result, while the detection result corresponding to the second image corresponding to the second region is not affected by the first lesion, so that false positive removal is not necessary, in this case, inputting the detection result of the medical image into a preset second classification model means inputting the detection result corresponding to the first image into the second classification model to obtain a second classification result output by the second classification model; and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
The following description will still take the medical image as a lung image, the first region is the region where the left lung is located, the second region is the region where the right lung is located, the first focus is pneumonia, and the second focus is a lung nodule as an example:
if pneumonia exists in the left lung and pneumonia does not exist in the right lung, the detection result corresponding to the left lung needs to be input into the second classification model, and the second classification model is used for judging the attribute information of each nodule so as to determine whether the nodule is a pneumonia region. If the area is a pneumonia area, it indicates that the nodule is false positive, and the detection result corresponding to the right lung (the detection result corresponding to the second image) and the detection result corresponding to the left lung excluding the nodule determined as false positive are set as target detection results.
In the second case, the second region has the first lesion, the first region does not have the first lesion, and the detection result corresponding to the second image corresponding to the second region is adversely affected by the first lesion, so that it is necessary to perform false positive removal on the detection result, and the detection result corresponding to the first image corresponding to the first region is not affected by the first lesion, so that false positive removal is not necessary, in which case, inputting the detection result of the medical image into the preset second classification model means inputting the detection result corresponding to the second image into the second classification model to obtain the second classification result output by the second classification model; and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
In the detection, when the first classification result indicates that the first region or the second region in the medical image has the first focus, the second classification model is used for performing false positive processing on the medical image to obtain the second classification result. Thus, the second classification result is the detection result of the false positive nodule. And finally, determining a target detection result according to the second classification result and a detection result corresponding to the region without removing the false positive, so that the target detection result is not influenced by the first focus to cause the false positive, and the accuracy of the target detection result is improved.
The following description will still take the medical image as a lung image, the first region is the region where the left lung is located, the second region is the region where the right lung is located, the first focus is pneumonia, and the second focus is a lung nodule as an example:
if pneumonia exists in the right lung and pneumonia does not exist in the left lung, the detection result of the right lung needs to be input into a second classification model, the second classification model is used for judging the attribute of each nodule in the detection result, namely whether each nodule is a pneumonia area or not, if yes, the detection result is a false positive detection result generated under the influence of pneumonia (a first focus), in this case, the nodule is judged to be false positive, the nodule is removed from being judged to be false positive, and the detection result corresponding to the left lung (the detection result corresponding to the first image) and the detection result corresponding to the right lung without the nodule judged to be false positive are used as target detection results.
When the detection is carried out, the target detection result is determined by combining the detection result corresponding to the left lung and the detection result corresponding to the right lung which is judged to be the false positive nodule, so that the target detection result is not influenced by the first focus to cause false positive, and the accuracy of the target detection result is improved.
The technical solution of the present application is described below with reference to specific embodiments, as shown in fig. 4, which shows a flowchart of a false positive removing method, specifically including the following steps:
step 401, acquiring a medical image.
The medical image may be a chest CT image.
Step 402, performing lung segmentation on the medical image to obtain a first image (left lung) and a second image (right lung).
And 403, inputting the medical image into the pneumonia classification model to obtain a pneumonia classification result, wherein the pneumonia classification result comprises a left pneumonia classification result and a right pneumonia classification result.
Step 404, inputting the medical image into the lung nodule detection model to obtain an initial lung nodule detection result, obtaining an initial left lung nodule detection result by combining the initial lung nodule detection result and the first image corresponding to the left lung, and obtaining an initial right lung nodule detection result by combining the initial lung nodule detection result and the second image corresponding to the right lung.
And step 405, combining the pneumonia classification result and the pulmonary nodule detection result to obtain a target detection result.
And if the pneumonia does not exist in the two lungs, determining the pulmonary nodule detection results of the left lung and the right lung as target detection results.
If only the right lung has pneumonia, the initial right lung nodule detection result is input into a false positive removal model, and the attribute of each nodule is judged, wherein the attribute can be used for indicating whether the nodule is a pneumonia area, if the nodule is a pneumonia area, the nodule is determined to be false positive, and a false positive removal result is obtained, and as shown in fig. 5, the area covered by a line frame in fig. 5 is a false positive nodule. And taking the false positive removed right pulmonary nodule detection result and the initial left pulmonary nodule detection result as target detection results.
And if only the left lung has pneumonia, inputting the initial left lung nodule detection result into a false positive removal model, judging the attribute of each nodule, wherein the attribute can be used for indicating whether the nodule is a pneumonia region, if so, determining that the nodule is false positive to obtain a false positive removal result, and taking the false positive removal left lung nodule detection result and the initial right lung nodule detection result as target detection results.
If pneumonia exists in the left lung and the right lung, the pulmonary nodule detection results of the left lung and the right lung are input into a false positive removing model, and the attribute of each nodule is judged. And removing the left and right pulmonary nodules in the pneumonia area to obtain a target detection result.
The algorithm provided by the embodiment of the application only removes false positives from lung detection results of the infected pneumonia, but does not accidentally injure detection results of the non-pneumonia infection, and ensures higher accuracy.
The pneumonia classification model is explained below:
and collecting lung images of infection and inflammation as a classification positive sample, collecting normal lung images as a classification negative sample, adopting a 3D ResidualNet (deep three-dimensional residual error neural network) as an initial network model, and adopting a Focal local as a Loss function to train the initial network model to obtain the pneumonia classification model.
The acquisition process of the classified positive samples and the classified negative samples comprises the following contents:
acquiring a whole lung image, and performing image recognition on the whole lung image to obtain three-dimensional bounding boxes (english) of the left and right lungs, as shown by a line frame in fig. 6.
Then, a lung region is extracted based on the three-dimensional bounding box of the lung. The specific method is based on the bounding box expanding according to a certain proportion, for example, expanding 0.1 times along the xyz direction (ensuring to contain all the left lung or right lung region), then cutting out the lung structure and resampling to a fixed size, for example, 96 × 96 × 96. As shown in fig. 6, fig. 6 shows a schematic diagram of the extraction of a left lung sample and a right lung sample.
And regarding the left lung sample and the right lung sample, the cut lung image of the infection inflammation is used as a classification positive sample, and the cut normal lung image is used as a classification negative sample. Optionally, in the application, data enhancement can be performed on the classification positive sample and the classification negative sample in a translation, scaling, rotation, inversion and other modes, and the universality of the trained pneumonia classification model can be improved through the data enhancement.
Optionally, the classification positive sample takes a single lung as a unit, that is, the cut left lung image of the infection and inflammation is used as the classification positive sample to train the left pneumonia classification model, and the cut right lung image of the infection and inflammation is used as the classification positive sample to train the right pneumonia classification model. Or the left lung and the right lung can be cut out to obtain a sample and train the left and right lung pneumonia classification models at the same time.
As shown in fig. 7, the medical image is subjected to image segmentation, boundary boxes of the left lung and the right lung are respectively extracted according to the segmentation result, so as to cut a first image corresponding to the left lung and a second image corresponding to the right lung, and then the first image and the second image are respectively input into a trained pneumonia classification model to judge whether the left lung and the right lung are infected with pneumonia, and obtain a judgment result, wherein the judgment result can be represented by 0 or 1, the input image in fig. 7 is the left lung and the right lung, and outputs 1 and 0 represent that pneumonia exists in the input left lung and pneumonia does not exist in the right lung.
The following explains the classification model for removing false positives:
the input of the false positive removal classification model is a nodule detection result, the training samples comprise classification positive samples and classification negative samples, the classification positive samples refer to the training samples containing the first focus, the classification negative samples refer to the training samples not containing the first focus, inflammatory nodules are extracted in a fixed _ box mode to serve as the classification positive samples, non-pneumonia nodules serve as the classification negative samples, the initial network model is trained, and the lung nodule false positive removal classification model is obtained. In this case, as shown in fig. 8, when extracting the nodule region, the nodule region may be clipped out by expanding 2.5 times based on the bounding box of the nodule.
In the embodiment of the application, whether pneumonia (a first focus) exists in a medical image is identified through a pneumonia classification model, and then false positive removing processing is carried out on a pulmonary nodule detection result according to a pneumonia classification result, so that nodules in a final target detection result do not contain inflammatory nodules caused by the first focus, and the accuracy of the target detection result is improved.
It should be understood that, although the steps in the flowcharts related to the embodiments as described above are sequentially displayed as indicated by arrows, the steps are not necessarily performed sequentially as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the embodiments described above may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the execution order of the steps or stages is not necessarily sequential, but may be rotated or alternated with other steps or at least a part of the steps or stages in other steps.
Based on the same inventive concept, the embodiment of the present application further provides an image detection apparatus for implementing the image detection method. The implementation scheme for solving the problem provided by the apparatus is similar to the implementation scheme described in the above method, so specific limitations in one or more embodiments of the image detection apparatus provided below can be referred to the limitations of the image detection method in the foregoing, and details are not described here.
In one embodiment, as shown in fig. 9, there is provided an image detection apparatus including: a classification module 901, a detection module 902, and a determination module 903, wherein:
the classification module 901 is configured to input a medical image to be processed into a preset first classification model, so as to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
the detection module 902 is configured to input the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
a determining module 903, configured to determine a target detection result of the medical image according to the first classification result and the detection result.
In one embodiment, the first classification model includes a first region classification model and a second region classification model, and the classification module 901 is specifically configured to:
carrying out image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into a second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the medical image is input into a preset detection model to obtain a detection result of the medical image, and the detection result comprises the following steps:
and inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In one embodiment, the classification module 901 is specifically configured to:
and if the first classification result indicates that the first focus does not exist in the medical image, taking the detection result as a target detection result.
In one embodiment, the classification module 901 is specifically configured to:
if the first classification result indicates that the first region and the second region in the medical image both have the first focus, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result.
In one embodiment, the determining module 903 is specifically configured to:
if the first classification result indicates that the first region or the second region in the medical image has the first focus, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result.
In one embodiment, the determining module 903 is specifically configured to:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into a second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
In one embodiment, the determining module 903 is specifically configured to:
if the first classification result indicates that the second region has the first focus and the first region does not have the first focus, inputting a detection result corresponding to the second image into the second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
The modules in the image detection device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 10. The computer device includes a processor, a memory, and a network interface 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, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used for storing the trained data of the first classification model, the second classification model, the image segmentation model and the like. 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 an image detection method.
Those skilled in the art will appreciate that the architecture shown in fig. 10 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain 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 a computer program stored therein, the processor implementing the following steps when executing the computer program:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In one embodiment, the first classification model comprises a first region classification model and a second region classification model, and the processor when executing the computer program further performs the steps of:
carrying out image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into a second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the medical image is input into a preset detection model to obtain a detection result of the medical image, and the detection result comprises the following steps:
and inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In one embodiment, the processor when executing the computer program further performs the steps of:
and if the first classification result indicates that the first focus does not exist in the medical image, taking the detection result as a target detection result.
In one embodiment, the processor when executing the computer program further performs the steps of:
if the first classification result indicates that the first region and the second region in the medical image both have the first focus, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
if the first classification result indicates that the first region or the second region in the medical image has the first focus, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into a second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
if the first classification result indicates that the second region has the first focus and the first region does not have the first focus, inputting a detection result corresponding to the second image into the second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, performs the steps of:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In an embodiment, the first classification model comprises a first region classification model and a second region classification model, the computer program, when executed by the processor, further realizing the steps of:
carrying out image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into a second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the medical image is input into a preset detection model to obtain a detection result of the medical image, and the detection result comprises the following steps:
and inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and if the first classification result indicates that the first focus does not exist in the medical image, taking the detection result as a target detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region and the second region in the medical image both have the first focus, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region or the second region in the medical image has the first focus, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into a second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the second region has the first focus and the first region does not have the first focus, inputting a detection result corresponding to the second image into the second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, performs the steps of:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
In an embodiment, the first classification model comprises a first region classification model and a second region classification model, the computer program, when executed by the processor, further realizing the steps of:
carrying out image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into a first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into a second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the medical image is input into a preset detection model to obtain a detection result of the medical image, and the detection result comprises the following steps:
and inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and if the first classification result indicates that the first focus does not exist in the medical image, taking the detection result as a target detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region and the second region in the medical image both have the first focus, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region or the second region in the medical image has the first focus, inputting the detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into a second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the second image.
In one embodiment, the computer program when executed by the processor further performs the steps of:
if the first classification result indicates that the second region has the first focus and the first region does not have the first focus, inputting a detection result corresponding to the second image into the second classification model to obtain a second classification result output by the second classification model;
and determining a target detection result according to the second classification result and the detection result corresponding to the first image.
It should be noted that the medical images referred to in the present application are all images authorized by the user or fully authorized by each party.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by hardware instructions of a computer program, which may be stored in a non-volatile computer-readable storage medium, and when executed, may include the processes of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high-density embedded nonvolatile Memory, resistive Random Access Memory (ReRAM), Magnetic Random Access Memory (MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (PCM), graphene Memory, and the like. Volatile Memory can include Random Access Memory (RAM), external cache Memory, and the like. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others. The databases referred to in various embodiments provided herein may include at least one of relational and non-relational databases. The non-relational database may include, but is not limited to, a block chain based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing based data processing logic devices, etc., without limitation.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.
Claims (10)
1. An image detection method, characterized in that the method comprises:
inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and determining a target detection result of the medical image according to the first classification result and the detection result.
2. The method according to claim 1, wherein the first classification model comprises a first region classification model and a second region classification model, and the inputting the medical image to be processed into the preset first classification model to obtain the first classification result of the medical image comprises:
performing image segmentation on the medical image to obtain a first image corresponding to a first region in the medical image and a second image corresponding to a second region in the medical image;
inputting the first image into the first region classification model to obtain a first classification result corresponding to the first image output by the first region classification model;
inputting the second image into the second region classification model to obtain a first classification result corresponding to the second image output by the second region classification model;
correspondingly, the step of inputting the medical image into a preset detection model to obtain a detection result of the medical image comprises:
inputting the medical image into a preset detection model to obtain a detection result of the medical image, and determining a detection result corresponding to the first image and a detection result corresponding to the second image according to the detection result.
3. The method of claim 1, wherein determining the target detection result of the medical image according to the first classification result and the detection result comprises:
if the first classification result indicates that the first lesion does not exist in the medical image, taking the detection result as the target detection result.
4. The method of claim 1, wherein determining the target detection result of the medical image according to the first classification result and the detection result comprises:
if the first classification result indicates that the first focus exists in both the first region and the second region in the medical image, inputting the detection result into a preset second classification model to obtain a second classification result output by the second classification model;
and determining the target detection result according to the second classification result.
5. The method of claim 2, wherein determining the target detection result of the medical image according to the first classification result and the detection result comprises:
if the first classification result indicates that the first focus exists in a first region or a second region in the medical image, inputting a detection result of the medical image into a preset second classification model to obtain a second classification result output by the second classification model;
and determining the target detection result according to the second classification result and the detection result.
6. The method of claim 5, wherein the determining the target detection result according to the second classification result and the detection result comprises:
if the first classification result indicates that the first region has the first focus and the second region does not have the first focus, inputting a detection result corresponding to the first image into the second classification model to obtain the second classification result output by the second classification model;
and determining the target detection result according to the second classification result and the detection result corresponding to the second image.
7. The method of claim 5, wherein the determining the target detection result according to the second classification result and the detection result comprises:
if the first classification result indicates that the first focus exists in the second region and the first focus does not exist in the first region, inputting a detection result corresponding to the second image into the second classification model to obtain the second classification result output by the second classification model;
and determining the target detection result according to the second classification result and the detection result corresponding to the first image.
8. An image detection apparatus, characterized in that the apparatus comprises:
the classification module is used for inputting a medical image to be processed into a preset first classification model to obtain a first classification result of the medical image; the first classification result is used for indicating whether a first focus exists in the medical image;
the detection module is used for inputting the medical image into a preset detection model to obtain a detection result of the medical image; the detection result is used for indicating whether a second focus exists in the medical image;
and the determining module is used for determining a target detection result of the medical image according to the first classification result and the detection result.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210199037.6A CN114612710A (en) | 2022-03-01 | 2022-03-01 | Image detection method, image detection device, computer equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210199037.6A CN114612710A (en) | 2022-03-01 | 2022-03-01 | Image detection method, image detection device, computer equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114612710A true CN114612710A (en) | 2022-06-10 |
Family
ID=81860740
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210199037.6A Pending CN114612710A (en) | 2022-03-01 | 2022-03-01 | Image detection method, image detection device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114612710A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115131345A (en) * | 2022-08-29 | 2022-09-30 | 杭州堃博生物科技有限公司 | CT image-based focus detection method and device and computer-readable storage medium |
-
2022
- 2022-03-01 CN CN202210199037.6A patent/CN114612710A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115131345A (en) * | 2022-08-29 | 2022-09-30 | 杭州堃博生物科技有限公司 | CT image-based focus detection method and device and computer-readable storage medium |
CN115131345B (en) * | 2022-08-29 | 2023-02-03 | 杭州堃博生物科技有限公司 | CT image-based focus detection method and device and computer-readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108765369B (en) | Method, apparatus, computer device and storage medium for detecting lung nodule | |
Su et al. | Region segmentation in histopathological breast cancer images using deep convolutional neural network | |
JP6710135B2 (en) | Cell image automatic analysis method and system | |
CN110852983A (en) | Method for detecting defects in semiconductor device | |
KR101938992B1 (en) | Explainable computeraided diagnosis system and the method thereof | |
CN113272827A (en) | Validation of classification decisions in convolutional neural networks | |
CN111008643B (en) | Picture classification method and device based on semi-supervised learning and computer equipment | |
US20180307940A1 (en) | A method and a device for image matching | |
CN116168351B (en) | Inspection method and device for power equipment | |
JP2008225692A (en) | Image retrieval method, device and program | |
US11328402B2 (en) | Method and system of image based anomaly localization for vehicles through generative contextualized adversarial network | |
CN110570390A (en) | image detection method and device | |
US20180047158A1 (en) | Chest radiograph (cxr) image analysis | |
EP3712845B1 (en) | Image identifying method and image identifying device | |
CN114612710A (en) | Image detection method, image detection device, computer equipment and storage medium | |
CN113160199B (en) | Image recognition method and device, computer equipment and storage medium | |
Dai et al. | Detection and segmentation of image anomalies based on unsupervised defect reparation | |
CN111401102A (en) | Deep learning model training method and device, electronic equipment and storage medium | |
CN114511523A (en) | Gastric cancer molecular subtype classification method and device based on self-supervision learning | |
CN113506288A (en) | Lung nodule detection method and device based on transform attention mechanism | |
CN110991408B (en) | Method and device for segmenting white matter high signal based on deep learning method | |
CN112287905A (en) | Vehicle damage identification method, device, equipment and storage medium | |
CN116958679A (en) | Target detection method based on weak supervision and related equipment | |
CN110852384A (en) | Medical image quality detection method, device and storage medium | |
CN112862785B (en) | CTA image data identification method, device and storage 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 |