CN113902676A - Lung nodule image detection method and system based on attention mechanism neural network - Google Patents

Lung nodule image detection method and system based on attention mechanism neural network Download PDF

Info

Publication number
CN113902676A
CN113902676A CN202111049299.6A CN202111049299A CN113902676A CN 113902676 A CN113902676 A CN 113902676A CN 202111049299 A CN202111049299 A CN 202111049299A CN 113902676 A CN113902676 A CN 113902676A
Authority
CN
China
Prior art keywords
neural network
image
attention
lung
lung nodule
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
Application number
CN202111049299.6A
Other languages
Chinese (zh)
Inventor
万洪林
赵莹莹
王嘉鑫
王晓敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Normal University
Original Assignee
Shandong Normal University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong Normal University filed Critical Shandong Normal University
Priority to CN202111049299.6A priority Critical patent/CN113902676A/en
Publication of CN113902676A publication Critical patent/CN113902676A/en
Pending legal-status Critical Current

Links

Images

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
    • G06F18/21—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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20221—Image fusion; Image merging
    • 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
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30061—Lung
    • G06T2207/30064—Lung nodule

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Image Analysis (AREA)

Abstract

The present disclosure provides a pulmonary nodule image detection method and system based on a neural network of an attention mechanism, which includes the following steps: acquiring a lung nodule image; obtaining a classification result of the lung nodule according to the obtained image and a preset neural network detection model; wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network. The trainable 3D attention mechanism module is fused into the neural network, high-level abstract features of data are effectively extracted, the data are directly used for recognition, classification and detection, the automation degree is high, real nodules and non-nodules can be effectively distinguished, and good effects are achieved in the aspects of improving the detection rate and reducing the false positive rate.

Description

Lung nodule image detection method and system based on attention mechanism neural network
Technical Field
The disclosure belongs to the technical field of medical image detection, and particularly relates to a pulmonary nodule image detection method and system based on a neural network of an attention mechanism.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Most early lung cancer patients do not have obvious clinical symptoms and specific biomarkers, so the current primary method for early screening is to examine the lung by radiographic images for suspicious lesions. Early lung cancer is mostly characterized by pulmonary nodules that are small in size, low in contrast, and high in shape heterogeneity. The workload of manual film reading is huge, at least 100 chest CT images of each examined person exist, and the precision scanning is even as high as 600, so that the doctor is easy to fatigue, and the probability of erroneous judgment and missed judgment is increased. Clearly, it is extremely difficult to rely solely on a physician to manually screen lung nodules in CT images.
In recent years, the appearance of large data sets and the rise of deep convolutional neural networks make breakthrough progress on a series of visual tasks, and researchers further draw attention to the attention thinking mode in human vision. The computer vision method based on the trainable attention mechanism can effectively focus on the interested region of the task autonomously, inhibit irrelevant regions, further improve the performance of the model, and gradually become a hot research problem.
The prior art has the characteristics of low detection accuracy and high false positive, and is difficult to be used in actual clinical detection.
Disclosure of Invention
In order to solve the problems, the trainable 3D attention mechanism module is fused into the neural network, high-level abstract features of data are effectively extracted, the data are directly used for recognition, classification and detection, the automation degree is high, real nodules and non-nodules can be effectively distinguished, and better effects are obtained in the aspects of improving the detection rate and reducing the false positive rate.
According to some embodiments, a first aspect of the present disclosure provides a lung nodule image detection method based on a neural network of an attention mechanism, which adopts the following technical solutions:
a lung nodule image detection method based on a neural network of an attention mechanism comprises the following steps:
acquiring a lung nodule image;
obtaining a classification result of the lung nodule according to the obtained image and a preset neural network detection model;
wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network.
As a further technical limitation, the acquired lung nodule image is a preprocessed lung CT image, and the preprocessing includes image enhancement and image parenchymal segmentation.
Further, the image enhancement comprises image texture enhancement and image denoising; the image texture enhancement adopts histogram equalization to enable the texture of the lung CT image to be clearer, and the image denoising adopts a median filtering method to remove noise in the lung CT image.
Further, the neural network detection model consists of a 3D residual error convolution module and an attention mechanism module; the preprocessed lung CT image enters an attention module channel after passing through a 3D residual convolution module, image compression is carried out under the action of an average pooling layer and a maximum pooling layer to generate an average pooling characteristic and a maximum pooling characteristic, and a final attention channel is generated based on the characteristic fusion of a sharing weight layer.
Further, the 3D residual convolution module uses two convolution kernels of 1 × 1 and one convolution kernel of 3 × 3.
As a further technical limitation, the neural network detection model adopts a stochastic gradient descent algorithm to train the model.
As a further technical limitation, the loss function of the neural network detection model includes classification loss and regression loss.
According to some embodiments, a second aspect of the present disclosure provides a pulmonary nodule image detection system based on a neural network of an attention mechanism, which adopts the following technical solutions:
a pulmonary nodule image detection system based on a neural network of an attention mechanism, comprising:
the acquisition module is used for acquiring a lung nodule image;
the detection module is used for obtaining a classification result of the lung nodule according to the acquired image and a preset neural network detection model;
wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network.
According to some embodiments, a third aspect of the present disclosure provides a computer-readable storage medium, which adopts the following technical solutions:
a computer readable storage medium, having stored thereon a program which, when being executed by a processor, carries out the steps of the method for lung nodule image detection based on an attention-based neural network according to the first aspect of the present disclosure.
According to some embodiments, a fourth aspect of the present disclosure provides an electronic device, which adopts the following technical solutions:
an electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, the processor implementing the steps in the method for lung nodule image detection based on attention-driven neural networks according to the first aspect of the present disclosure when executing the program.
Compared with the prior art, the beneficial effect of this disclosure is:
the trainable 3D attention mechanism module is fused into the neural network, high-level abstract features of data are effectively extracted, the data are directly used for identification, classification and detection, the automation degree is high, real nodules and non-nodules can be effectively distinguished, the detection precision is improved, large-scale lung nodules can be detected, small-scale lung nodules can also be detected, and the good effects of improving the detection rate and reducing the false positive rate are achieved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
Fig. 1 is a flowchart of a lung nodule image detection method based on a neural network of an attention mechanism in an embodiment of the disclosure;
FIG. 2 is a block diagram of a pulmonary nodule neural network detection model according to an embodiment of the present disclosure;
FIG. 3 is a block diagram of an attention device in accordance with a first embodiment of the present disclosure;
FIG. 4 is a block diagram of a 3D residual neural network according to an embodiment of the present disclosure;
fig. 5 is a block diagram of a lung nodule image detection system based on a neural network of an attention mechanism in the second embodiment of the disclosure.
The specific implementation mode is as follows:
the present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.
Example one
The embodiment of the disclosure introduces a lung nodule image detection method based on a neural network of an attention mechanism.
Fig. 1 shows a lung nodule image detection method based on a neural network of an attention mechanism, which includes the following steps:
data preprocessing, namely selecting candidate lung nodules from an original scanning data set;
taking a lung nodule region as a center, intercepting 128-by-128 images as input data of a positive sample, dividing a data set, and performing data augmentation on the positive sample data set;
constructing a neural network detection model based on an attention mechanism, and performing model training and parameter adjustment;
and testing the test set of the trained neural network detection model, and evaluating according to the evaluation index to obtain the classification result of the pulmonary nodules.
Specifically, the purpose of data preprocessing is to extract lung regions and remove noise around the lungs, and specifically includes:
(a) in the embodiment, a LIDC data set is adopted, firstly, threshold segmentation is carried out, and a CT value is set to be 480HU to be used for segmenting a CT image so as to obtain a binary image; the values obtained after CT acquisition are X-ray attenuation values in hounsfield units, HU values for water 0, HU values for air-1000, and HU values for other objects calculated as:
Figure BDA0003252132220000061
where μ is a linear attenuation coefficient, related to the X-ray intensity.
Herein, Hounsfield Units (HU) are dimensionless units commonly used in Computed Tomography (CT) for standard, convenient representation of CT values. The Hounsfield unit is obtained by linear transformation of the measured attenuation coefficient; this conversion is based on the density of air, where pure water is defined as a 0Hounsfield unit and air as a-1000 Hounsfield unit. The higher the tissue density is, the stronger the x-ray absorption is, the positive value is, and a bright signal is presented; less dense tissue, less x-ray absorption, negative values, dark signals.
(b) And (3) enhancing the image, adopting a histogram equalization method to make the texture of the image clearer, and adopting a median filtering method to remove the noise of the lung CT image.
(c) Segmenting parenchyma of lung, and carrying out lung contour segmentation on the data; the aim is to focus only on the image of the lung tissue and to reject noise caused by the influence of extraneous parts.
The coarse feature mapping of lung nodules can reduce the accuracy of detection, and thus it is necessary to add a mechanism of attention. The attention mechanism is that under the condition of limited computing capacity, computing resources are allocated to more important tasks, and similar to the human visual attention mechanism, a target area needing attention is obtained by scanning a global image, more attention resources are invested in the area, more information about the target is obtained, other irrelevant information is ignored, and the efficiency and the accuracy of visual information processing are improved. In this embodiment, trainable 3D attention force door modules are leveraged and fused into the network; the network structure increases the high-level abstract feature that the attention mechanism can more effectively extract data, the output feature can be directly used for identification, classification and detection, the degree of automation is high, real nodules and non-nodules can be more effectively distinguished, and a good effect is achieved in the aspects of improving the detection rate and reducing the false positive rate.
The process of the neural network detection model based on the attention mechanism specifically comprises the following steps:
(1) the lung nodule neural network detection model shown in fig. 2 integrates UNet + + codec concept into feature detection, the network forward downsampling part is composed of five 3D residual convolution blocks and an attention mechanism module (shown in fig. 3), a preprocessed image passes through one convolution block and then enters a channel attention module, an input feature map is compressed through an average pooling layer and a maximum pooling layer, and average aggregated feature X is generatedavgAnd maximum Convergence feature XmaxThen X is addedavgAnd XmaxInput deviceTo a shared weight layer of 2 DenseLayers. Fusing the characteristics output by the sharing weight layer, activating by a sigmoid function, and generating a final channel attention Mc∈Rnchannels。
In summary, the general formula of the channel attention is:
Mc＝σ(FC(Max_pool(X))+FC(Avg_pool(X))) (2)
wherein, σ is a sigmoid function, and the function is defined as follows;
Figure BDA0003252132220000081
the spatial attention can be regarded as a supplement to the channel attention, and after the channel attention is obtained, the channel attention and the input feature map are fused to serve as the input feature of the spatial attention module. The inputs to the space attention module XC can be described by the following equation:
Figure BDA0003252132220000082
wherein the content of the first and second substances,
Figure BDA0003252132220000083
representing element-wise multiplication. To fully utilize channel information, three-dimensional feature map X is generated by global average pooling and maximum pooling operationsavgAnd Xmax(ii) a Then they are fused and a three-dimensional space note is generated by a standard convolution layer, which is called Ms. Spatial attention is calculated as:
Mc＝σ(FC(Max_pool(X))+FC(Avg_pool(X))) (5)
the MC and the MS are fused to obtain the output of the whole attention module, which can be expressed by the following formula:
Figure BDA0003252132220000084
after the image is output from the attention module, other convolution blocks are input from the back, each convolution block operation is followed by a maximum pooling layer, the pooling layer reduces the image feature mapping scale by half, the output of the previous layer is subjected to pooling and cropping operation, and the pooling and cropping operation are connected on a channel and used as the input of the next layer; and the image downsampling operation is realized through pooling, the characteristics are extracted, and the subsequent convolution operation parameters are reduced. The deconvolution lifting part of the convolution neural network consists of 3 convolution blocks and an area proposal network output layer.
(2) The 3D residual network structure used in this embodiment is shown in fig. 4, and the convolution block is composed of 3D versions of residual network blocks (two convolution kernels of 1 × 1, one convolution kernel of 3 × 3, each convolution kernel being followed by an activation function and normalization), which reduces the amount of parameter computation by nearly half compared to two convolution kernels of 3 × 3.
The training and testing process of the neural network detection model based on the attention mechanism is as follows:
(a) the network model training is carried out and trained by adopting a random gradient descent algorithm. In the testing stage, the present embodiment adopts a 10-fold cross validation method. In order to better prove the effectiveness of adding the attention module, in the present embodiment, an ablation experiment is performed, as shown in table 1, the first row is that no module is added, the second row is that a multi-scale module is added, and the third row is that the attention module is added, so that the detection effect after the attention module is added is better.
TABLE 1 ablation experiment
Figure BDA0003252132220000091
(b) Loss function
The loss function is divided into two parts, classification loss and regression loss. The multitask penalty function for anchor box i is defined as:
L(pi,ti)＝λLcls(pi,pi*)+piLreg(ti,ti*) (7)
wherein p isiIs pre-pended for the current anchor frame iMeasured as the probability of a nodule, λ is piAnd λ is 0.5.
For LclsConsidering the use of sigmoid as the activation function, the use of the two-class cross entropy loss function is chosen in the present embodiment. It is defined as follows:
Lcls＝-[yi·log(pi)+(1-yi)·log(1-pi)] (8)
in the formula, yiThe truth label of the anchor is followed by the regression loss function. It is defined as follows:
Lreg(ti,ti*)＝H(ti-ti*) (9)
in the formula, tiIs the relative coordinate of the corresponding nodule, tiMarked nodule coordinates, defined as:
Figure BDA0003252132220000101
where (x, y, z, d) is the predicted coordinates and diameter of the nodule in the original space, (x)a,ya,zaAnd d) is the coordinates and dimensions of the anchor frame i.
For H, considering that the L2 loss function will be error squared, the model will be more sensitive to a single sample, and the L1 norm loss function does not exist at the 0 point, which may affect convergence, the specific formula of the L1 loss function selected by this embodiment is as follows:
Figure BDA0003252132220000102
in the embodiment, the trainable 3D attention mechanism module is fused into the neural network, high-level abstract features of data are effectively extracted, the data are directly used for identification, classification and detection, the automation degree is high, real nodules and non-nodules can be effectively distinguished, and good effects are obtained in the aspects of improving the detection rate and reducing the false positive rate.
Example two
The second embodiment of the present disclosure introduces a lung nodule image detection system based on a neural network of an attention mechanism, and adopts the lung nodule image detection method based on the neural network of the attention mechanism introduced in the first embodiment.
A lung nodule image detection system based on attention-based neural networks, as shown in fig. 5, includes:
the acquisition module is used for acquiring a lung nodule image;
the detection module is used for obtaining a classification result of the lung nodule according to the acquired image and a preset neural network detection model;
wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network.
The detailed steps are the same as those of the lung nodule image detection method based on the attention mechanism neural network provided in the first embodiment, and are not described herein again.
EXAMPLE III
The third embodiment of the disclosure provides a computer-readable storage medium.
A computer readable storage medium, on which a program is stored, which when executed by a processor implements the steps in the method for lung nodule image detection based on an attention-based neural network according to one embodiment of the present disclosure.
The detailed steps are the same as those of the lung nodule image detection method based on the attention mechanism neural network provided in the first embodiment, and are not described herein again.
Example four
The fourth embodiment of the disclosure provides an electronic device.
An electronic device includes a memory, a processor, and a program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method for detecting lung nodule images based on attention-based neural network according to an embodiment of the present disclosure.
The detailed steps are the same as those of the lung nodule image detection method based on the attention mechanism neural network provided in the first embodiment, and are not described herein again.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims (10)

1. A lung nodule image detection method based on a neural network of an attention mechanism is characterized by comprising the following steps:
acquiring a lung nodule image;
obtaining a classification result of the lung nodule according to the obtained image and a preset neural network detection model;
wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network.
2. The method as claimed in claim 1, wherein the acquired lung nodule image is a preprocessed lung CT image, and the preprocessing includes image enhancement and image parenchymal segmentation.
3. The method as claimed in claim 2, wherein the image enhancement comprises image texture enhancement and image denoising; the image texture enhancement adopts histogram equalization to enable the texture of the lung CT image to be clearer, and the image denoising adopts a median filtering method to remove noise in the lung CT image.
4. The method for detecting lung nodule images based on attention-based neural network as claimed in claim 3, wherein the neural network detection model is composed of a 3D residual convolution module and an attention-based module; the preprocessed lung CT image enters an attention module channel after passing through a 3D residual convolution module, image compression is carried out under the action of an average pooling layer and a maximum pooling layer to generate an average pooling characteristic and a maximum pooling characteristic, and a final attention channel is generated based on the characteristic fusion of a sharing weight layer.
5. A method for lung nodule image detection based on attention-based neural networks as claimed in claim 4 wherein the 3D residual convolution module uses two 1 x 1 convolution kernels and one 3 x 3 convolution kernel.
6. The method as claimed in claim 1, wherein the neural network detection model is trained by using a stochastic gradient descent algorithm.
7. A method for lung nodule image detection based on neural network of attention mechanism as claimed in claim 1 wherein the loss function of the neural network detection model includes classification loss and regression loss.
8. A pulmonary nodule image detection system based on a neural network of an attention mechanism, comprising:
the acquisition module is used for acquiring a lung nodule image;
the detection module is used for obtaining a classification result of the lung nodule according to the acquired image and a preset neural network detection model;
wherein, the neural network detection model adopts an attention mechanism and a 3D residual neural network.
9. A computer-readable storage medium, on which a program is stored which, when being executed by a processor, carries out the steps of the method for lung nodule image detection based on an attention-mechanism neural network according to any one of claims 1 to 7.
10. An electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps in the method for lung nodule image detection based on attention-based neural network of any of claims 1-7.
CN202111049299.6A 2021-09-08 2021-09-08 Lung nodule image detection method and system based on attention mechanism neural network Pending CN113902676A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111049299.6A CN113902676A (en) 2021-09-08 2021-09-08 Lung nodule image detection method and system based on attention mechanism neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111049299.6A CN113902676A (en) 2021-09-08 2021-09-08 Lung nodule image detection method and system based on attention mechanism neural network

Publications (1)

Publication Number Publication Date
CN113902676A true CN113902676A (en) 2022-01-07

Family

ID=79188855

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111049299.6A Pending CN113902676A (en) 2021-09-08 2021-09-08 Lung nodule image detection method and system based on attention mechanism neural network

Country Status (1)

Country Link
CN (1) CN113902676A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115358976A (en) * 2022-08-10 2022-11-18 北京医准智能科技有限公司 Image identification method, device, equipment and storage medium

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115358976A (en) * 2022-08-10 2022-11-18 北京医准智能科技有限公司 Image identification method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
Gong et al. Automated pulmonary nodule detection in CT images using 3D deep squeeze-and-excitation networks
Wu et al. Predict pneumonia with chest X-ray images based on convolutional deep neural learning networks
Brunetti et al. Computer-assisted frameworks for classification of liver, breast and blood neoplasias via neural networks: A survey based on medical images
CN108464840B (en) Automatic detection method and system for breast lumps
Omonigho et al. Breast cancer: tumor detection in mammogram images using modified alexnet deep convolution neural network
CN110046627B (en) Method and device for identifying mammary gland image
Tao et al. Multi-level ground glass nodule detection and segmentation in CT lung images
Prusty et al. A novel transfer learning technique for detecting breast cancer mammograms using VGG16 bottleneck feature
CN112700461B (en) System for pulmonary nodule detection and characterization class identification
CN112329871B (en) Pulmonary nodule detection method based on self-correction convolution and channel attention mechanism
CN111798424B (en) Medical image-based nodule detection method and device and electronic equipment
CN112508884A (en) Comprehensive detection device and method for cancerous region
Mei et al. YOLO-lung: A practical detector based on imporved YOLOv4 for Pulmonary Nodule Detection
CN110570419A (en) Method and device for acquiring characteristic information and storage medium
CN113269799A (en) Cervical cell segmentation method based on deep learning
Chen [Retracted] Models of Artificial Intelligence‐Assisted Diagnosis of Lung Cancer Pathology Based on Deep Learning Algorithms
Rasheed et al. Use of transfer learning and wavelet transform for breast cancer detection
Manikandan et al. Segmentation and Detection of Pneumothorax using Deep Learning
CN114565786A (en) Tomography image classification device and method based on channel attention mechanism
CN114581698A (en) Target classification method based on space cross attention mechanism feature fusion
CN113902676A (en) Lung nodule image detection method and system based on attention mechanism neural network
Cao et al. 3D convolutional neural networks fusion model for lung nodule detection onclinical CT scans
CN112801992A (en) Pulmonary nodule image classification method based on 3D residual error network
Guo et al. Thyroid nodule ultrasonic imaging segmentation based on a deep learning model and data augmentation
Hou et al. Edge detection of COVID-19 CT image based on GF_SSR, improved multiscale morphology, and adaptive threshold

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