CN111815591B - Lung nodule detection method based on CT image - Google Patents
Lung nodule detection method based on CT image Download PDFInfo
- Publication number
- CN111815591B CN111815591B CN202010609059.6A CN202010609059A CN111815591B CN 111815591 B CN111815591 B CN 111815591B CN 202010609059 A CN202010609059 A CN 202010609059A CN 111815591 B CN111815591 B CN 111815591B
- Authority
- CN
- China
- Prior art keywords
- image
- lung nodule
- corrosion
- original
- structural element
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 206010056342 Pulmonary mass Diseases 0.000 title claims abstract description 110
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 230000007797 corrosion Effects 0.000 claims abstract description 47
- 238000005260 corrosion Methods 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000013145 classification model Methods 0.000 claims abstract description 21
- 229910003460 diamond Inorganic materials 0.000 claims description 4
- 239000010432 diamond Substances 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 4
- 239000008186 active pharmaceutical agent Substances 0.000 claims description 3
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 5
- 201000005202 lung cancer Diseases 0.000 description 5
- 208000020816 lung neoplasm Diseases 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 210000004072 lung Anatomy 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000005530 etching Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000003211 malignant effect Effects 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 101100134058 Caenorhabditis elegans nth-1 gene Proteins 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30061—Lung
- G06T2207/30064—Lung nodule
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention relates to a lung nodule detection method based on CT images, and belongs to the technical field of medical image processing. The detection method comprises the following steps: acquiring an original CT image, and binarizing the original CT image; carrying out continuous n times of corrosion on the binarized image by utilizing structural elements, and carrying out convolution treatment on the image after each time of corrosion; performing exclusive OR operation on the images subjected to corrosion and convolution twice to find out the vanishing pixel points in the corrosion process; capturing a lung nodule suspected region picture with a set size on an original CT image by taking the position of a disappeared pixel point as the center; and inputting the lung nodule suspected region pictures into a trained classification model to classify, and finishing the detection of the lung nodule. Before the lung nodule region is positioned, the suspected lung nodule region is firstly found out, the accuracy of positioning the lung nodule region is improved, and the suspected lung nodule region is subjected to binarization, repeated corrosion and convolution operations on an original CT image, so that the whole process is simple.
Description
Technical Field
The invention relates to a lung nodule detection method based on CT images, and belongs to the technical field of medical image processing.
Background
Lung cancer is one of the most common malignant tumors in China, the five-year survival rate of lung cancer in China is only 16.1%, and early diagnosis of lung cancer is an important means for improving the survival rate of patients. However, early lung cancer features, lung nodules, are not apparent and need to be effectively differential diagnosed and treated. The diagnosis process is to quickly determine whether the lung nodule is benign or malignant, and the treatment process is to cut off the malignant nodule as early as possible, so that the diagnosis and treatment of the lung nodule can not only avoid unnecessary excessive treatment, but also prevent and treat the lung cancer.
At present, the diagnosis process of the lung nodule is carried out by matching a CT image with clinical experience, but due to limited medical resources, the diagnosis efficiency is low, so that the detection of the lung nodule by using a neural network is proposed, and a doctor is assisted, so that the human resources can be better saved and the detection efficiency is improved.
In general, the process of performing lung nodule recognition includes: firstly, segmenting a CT image through a neural network to obtain a lung nodule region; then, the lung nodule region is identified through another neural network, and the lung nodules are classified, so that the detection of the lung nodules is completed.
However, in the above detection process, the accuracy of determining the lung nodule region by using the neural network algorithm is low.
Disclosure of Invention
The purpose of the application is to provide a lung nodule detection method based on CT images, which is used for solving the problem of inaccurate existing detection.
In order to achieve the above purpose, the present application proposes a technical solution of a first lung nodule detection method based on CT images, including the following steps:
1) Acquiring an original CT image, and binarizing the original CT image;
2) Carrying out continuous n times of corrosion on the binarized image by utilizing structural elements, and carrying out convolution treatment on the image after each time of corrosion;
3) Performing exclusive OR operation on the images subjected to corrosion and convolution twice to find out the vanishing pixel points in the corrosion process;
4) Capturing a lung nodule suspected region picture with a set size on an original CT image by taking the position of a disappeared pixel point as the center;
5) And inputting the lung nodule suspected region pictures into a trained classification model to classify, and finishing the detection of the lung nodule.
The technical scheme of the lung nodule detection method based on the CT image has the advantages that: after the original CT image is obtained, binarization processing is firstly carried out, the binarization processing is the basis of subsequent mathematical morphology processing, the mathematical morphology processing is corrosion processing, in addition, continuous n times of corrosion processing are needed, each time of corrosion is carried out to eliminate pixel points at the extreme edge of the image, for a nodule area, the pixel points eliminated for the last time are necessarily lung nodule center points, therefore, pixel points disappeared in the corrosion process can be found out by carrying out exclusive OR operation on the images subjected to adjacent corrosion and convolution for two times, the positions of the disappeared pixel points are determined, the positions of suspected lung nodule areas on the original CT image are further determined through the number of times of corrosion and the number of layers of the pixel points subjected to corrosion for each time, finally, the suspected lung nodule areas are intercepted out and classified in a classification model, whether the areas are lung nodule areas or not is determined, and lung nodule detection is completed. Before the lung nodule region is positioned, the suspected lung nodule region is firstly found out, the accuracy of the lung nodule region positioning detection is improved, the suspected lung nodule region is determined by performing binarization, repeated corrosion and convolution operations on an original CT image, and the whole process is simple.
Further, in order to obtain the classification model, the training method of the classification model is as follows: and manually segmenting the marked CT image to obtain a data set, and training the data set.
Further, in order to complete the etching operation, the structural element in the step 2) is a cross structural element, a diamond structural element, a spherical structural element, or a disc structural element of 3*3.
In addition, the application also provides a technical scheme of a second lung nodule detection method based on CT images, which comprises the following steps:
1) Acquiring an original CT image, and binarizing the original CT image;
2) Carrying out continuous n times of corrosion on the binarized image by utilizing structural elements, and carrying out convolution treatment on the image after each time of corrosion;
3) Performing exclusive OR operation on the images subjected to corrosion and convolution twice to find out the vanishing pixel points in the corrosion process;
4) Capturing a lung nodule suspected region picture with a set size on an original CT image by taking the position of a disappeared pixel point as the center;
5) Downsampling the lung nodule suspected region picture to obtain an image pyramid;
6) Inputting the image pyramid in the step 5) into a trained classification model for classification, and performing fusion judgment, thereby completing the detection of the lung nodule.
The technical scheme of the second lung nodule detection method based on CT images has the advantages that: after the original CT image is obtained, binarization processing is firstly carried out, the binarization processing is the basis of subsequent mathematical morphology processing, the mathematical morphology processing is corrosion processing, in addition, continuous n times of corrosion processing are required, each time of corrosion is carried out to eliminate pixel points at the extreme edge of the image, for a nodule area, the pixel points eliminated for the last time are necessarily lung nodule center points, therefore, the pixel points disappeared in the corrosion process can be found out by carrying out exclusive OR operation on the images after adjacent two times of corrosion and convolution, the positions of the disappeared pixel points are determined, the positions of the lung nodule suspected areas on the original CT image are determined through the number of times of corrosion and the number of layers of the pixel points corroded each time, finally, the lung nodule suspected areas are intercepted, downsampling processing is carried out on the lung nodule suspected areas, an image pyramid is obtained, and then classification is carried out in a classification model, whether the areas are lung nodule areas or not is determined, and lung nodule detection is completed. Before the lung nodule region is positioned, the lung nodule suspected region is firstly found out, downsampling treatment is carried out on the lung nodule suspected region, multiscale information of the lung nodule suspected region is obtained, positioning detection accuracy is improved, and the determination of the lung nodule suspected region is carried out only by carrying out binarization, repeated corrosion and convolution operation on an original CT image, so that the whole process is simple.
Further, in order to improve the detection accuracy of the lung nodule, the method for fusion judgment in the step 6) includes: inputting the image pyramid into a trained classification model to obtain the classification probability of each layer of image in the image pyramid, and fusing the obtained classification probability by using a DS evidence theory method to finish the detection of the lung nodule.
Further, in order to obtain the classification model, the training method of the classification model is as follows: manually segmenting the marked CT image, performing downsampling treatment on the segmented image to obtain an image pyramid dataset, and training the image pyramid dataset.
Further, in order to complete the etching operation, the structural element in the step 2) is a cross structural element, a diamond structural element, a spherical structural element, or a disc structural element of 3*3.
Drawings
FIG. 1 is a schematic block diagram of an embodiment 1 of a CT image-based lung nodule detection method of the present invention;
FIG. 2 is a schematic diagram of a suspected region detection algorithm of the present invention;
FIG. 3 is a schematic diagram of the training process of the CNN model of the present invention;
fig. 4 is a schematic block diagram of an embodiment 2 of a lung nodule detection method based on CT images according to the present invention.
Detailed Description
Lung nodule detection method based on CT image example 1:
the lung nodule detection method based on CT images is mainly characterized in that based on the problem of large calculation amount of existing suspected region detection, the original CT images are subjected to binarization processing, structural elements are utilized to continuously corrode the binarized images for multiple times, convolution operation is carried out on the corroded images after each corrosion operation to obtain a plurality of corroded and convolved images, exclusive OR operation is carried out on the images after adjacent corrosion and convolution for two times, the pixel points disappeared in the corrosion process can be obtained, and on the original CT images, pictures with set sizes are cut by taking the positions of the disappeared pixel points as the centers, and the pictures are lung nodule suspected region pictures.
Specifically, the lung nodule detection method based on the CT image is shown in fig. 1, and comprises the following steps:
1) An original CT image of the lung is acquired.
2) And (3) finding out a lung nodule suspected region picture (image) of the original CT image in the step 1) through a suspected region detection algorithm.
The suspected region detection algorithm is shown in fig. 2:
a. performing binarization operation on the original CT image;
binarization is a process of setting the gray value of a pixel point on an image to 0 or 255, that is, displaying a clear black-and-white effect on the whole image.
b. Continuously performing n times of corrosion operation (n is generally 3-10 times) on the binarized image by using structural elements, wherein each time of corrosion can eliminate the edge of 1 pixel, and performing convolution operation after each time of corrosion operation to obtain a plurality of corroded and convolved images (feature map);
the cross-shaped structure element of 3*3 as the structure element in step b, corrosion is a process of eliminating boundary points, and making the boundary shrink inwards, can be used to eliminate small and meaningless objects. The specific process of corrosion is as follows: scanning each pixel in the binarized image by using the cross-shaped structural element of 3*3, performing AND operation by using the cross-shaped structural element and the binarized image covered by the cross-shaped structural element, and if the cross-shaped structural element and the binarized image are both 1, setting the pixel as 1; otherwise, the binary image is 0, and finally, the binary image is reduced by one circle; and the purpose of the convolution operation is to extract image features, finding isolated blocks of pixels;
c. performing exclusive or operation (the exclusive or operation is to remove the same part in the two pictures and keep different parts) on the pictures (such as the nth-1 corroded and convolved pictures and the nth corroded and convolved pictures) after adjacent corrosion and convolution for two times, finding out the vanishing pixel point in the corrosion process and taking the pixel point as the center to intercept a suspected Region (ROI) with a set size;
for example, a cross-shaped structural element of 3*3 is adopted, each time the edge 1 layer of pixel point is corroded, the position of the lung nodule suspected region is determined through the n step, on the original CT image, the pixel point corroded last time is taken as the center, and a rectangular region with the pixel size (i.e. the set size) of (n×1) is cut off.
3) Inputting the lung nodule suspected region picture obtained in the step 2) into a trained classification model (CNN model), finishing classification of the picture, determining whether the lung nodule suspected region is a lung nodule region or not, and further detecting the lung nodule suspected region.
The training method of the classification model is shown in fig. 3:
a. acquiring a plurality of CT images, and labeling the CT images as lung nodule areas (i.e. true nodules) and non-lung nodule areas (i.e. false nodules);
b. and manually segmenting the marked lung nodule area and the non-lung nodule area to obtain a series of images of the lung nodule area and the non-lung nodule area, segmenting to form a training set, and performing model training.
In the above embodiment, the structural element is a cross-shaped structural element of 3*3, and as other embodiments, the structural element may be a diamond-shaped structural element, a sphere-shaped structural element, a disc-shaped structural element, or the like with other dimensions, which is not limited in the present invention.
When the lung nodule suspected region is determined, only binarization, repeated corrosion and convolution operations are carried out on the original CT image, the whole process is simple, the calculation resources are reduced, the loss of the calculation resources is reduced, and the detection efficiency is improved.
Lung nodule detection method based on CT image example 2:
the difference between the lung nodule detection method based on the CT image and the lung nodule detection method based on the CT image of the embodiment 1 is that in the embodiment, when a lung nodule suspected region picture is obtained, downsampling is performed on the lung nodule suspected region picture to obtain an image pyramid of the picture, meanwhile, a corresponding classification model is an image pyramid training model, the image pyramid of the lung nodule suspected region is input into the image pyramid training model to complete classification detection, the image pyramid can obtain multi-scale features of the image, and detection accuracy is further improved.
Specifically, as shown in fig. 4, the lung nodule detection method based on CT images includes the following steps:
1) An original CT image of the lung is acquired.
2) And (3) finding out a lung nodule suspected region picture of the original CT image in the step (1) through a suspected region detection algorithm.
The specific implementation process of step 2) is described in the above embodiment 1 of the lung nodule detection method based on CT images, and will not be described here.
3) And (3) downsampling the lung nodule suspected region picture obtained in the step (2) to obtain an image pyramid of the suspected region.
The image pyramid is a combination of a series of images of different sizes of the same picture, the lowest image has the largest size, and the top image has the smallest size, and is like a pyramid from the space downwards. And downsampling the image with the largest size at the lowest position through filtering operation, such as Gaussian filtering, and obtaining the image pyramid after processing.
4) Inputting the image pyramid of the suspected region into a trained classification model to obtain the classification probability of each layer of image in the image pyramid, fusing the obtained classification probabilities through a Dempster synthesis rule in a DS evidence theory method to obtain a final recognition probability, and determining whether the suspected region of the lung nodule is the lung nodule region or not to finish the detection of the lung nodule.
The synthesis rule is as follows:
wherein A is 1 The value of the layer 1 image is taken; a is that 2 The value of the layer 2 image is taken; a is that n The value of the nth layer image is taken; the value can be lung nodule or non-lung nodule; m is m 1 (A 1 ) A as layer 1 image 1 Classification probability; m is m 2 (A 2 ) A as layer 2 image 2 Classification probability; m is m n (A n ) Is the firstA of n-layer image n Classification probability; n is the number of layers of the image pyramid;the final recognition probability is obtained, and when A is a lung nodule, the final recognition probability is obtained; when a is a non-lung nodule, a probability is identified for the final non-lung nodule.
The classification model is a CNN model of an image pyramid, and the training method comprises the following steps:
a. acquiring a plurality of CT images of the lung 2D, marking the CT images as a lung nodule area and a non-lung nodule area, and then manually dividing the marked lung nodule area and non-lung nodule area to obtain a series of images of the lung nodule area and the non-lung nodule area;
b. downsampling each image in a series of images of the lung nodule area and the non-lung nodule area to obtain a corresponding image pyramid, and obtaining multi-scale information of the images;
c. and taking all the marked image pyramids as training sets to train the classification model.
On the basis of improving the detection efficiency, the invention fully utilizes the multi-scale information of the CT image and improves the detection precision.
Claims (5)
1. A lung nodule detection method based on CT images, comprising the steps of:
1) Acquiring an original CT image, and binarizing the original CT image;
2) Carrying out continuous n times of corrosion on the binarized image by utilizing structural elements, and carrying out convolution treatment on the image after each time of corrosion; the structural element is a cross structural element, a diamond structural element, a spherical structural element or a disc structural element of 3*3;
3) Performing exclusive OR operation on the images subjected to corrosion and convolution twice to find out the vanishing pixel points in the corrosion process;
4) Capturing a lung nodule suspected region picture with a set size on an original CT image by taking the position of a disappeared pixel point as the center;
5) And inputting the lung nodule suspected region pictures into a trained classification model to classify, and finishing the detection of the lung nodule.
2. The method for detecting lung nodules based on CT images according to claim 1, wherein the training method of the classification model is: and manually segmenting the marked CT image to obtain a data set, and training the data set.
3. A lung nodule detection method based on CT images, comprising the steps of:
1) Acquiring an original CT image, and binarizing the original CT image;
2) Carrying out continuous n times of corrosion on the binarized image by utilizing structural elements, and carrying out convolution treatment on the image after each time of corrosion; the structural element is a cross structural element, a diamond structural element, a spherical structural element or a disc structural element of 3*3;
3) Performing exclusive OR operation on the images subjected to corrosion and convolution twice to find out the vanishing pixel points in the corrosion process;
4) Capturing a lung nodule suspected region picture with a set size on an original CT image by taking the position of a disappeared pixel point as the center;
5) Downsampling the lung nodule suspected region picture to obtain an image pyramid;
6) Inputting the image pyramid in the step 5) into a trained classification model for classification, and performing fusion judgment, thereby completing the detection of the lung nodule.
4. The method for detecting lung nodules based on CT images according to claim 3, wherein the method for fusion judgment in step 6) comprises: inputting the image pyramid into a trained classification model to obtain the classification probability of each layer of image in the image pyramid, and fusing the obtained classification probability by using a DS evidence theory method to finish the detection of the lung nodule.
5. The method for detecting lung nodules based on CT images according to claim 3 or 4, wherein the training method of the classification model is: manually segmenting the marked CT image, performing downsampling treatment on the segmented image to obtain an image pyramid dataset, and training the image pyramid dataset.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010609059.6A CN111815591B (en) | 2020-06-29 | 2020-06-29 | Lung nodule detection method based on CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010609059.6A CN111815591B (en) | 2020-06-29 | 2020-06-29 | Lung nodule detection method based on CT image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111815591A CN111815591A (en) | 2020-10-23 |
| CN111815591B true CN111815591B (en) | 2024-02-27 |
Family
ID=72856584
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010609059.6A Active CN111815591B (en) | 2020-06-29 | 2020-06-29 | Lung nodule detection method based on CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111815591B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113177955B (en) * | 2021-05-10 | 2022-08-05 | 电子科技大学成都学院 | Lung cancer image lesion area dividing method based on improved image segmentation algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107016665A (en) * | 2017-02-16 | 2017-08-04 | 浙江大学 | A kind of CT pulmonary nodule detection methods based on depth convolutional neural networks |
| CN108198179A (en) * | 2018-01-03 | 2018-06-22 | 华南理工大学 | A kind of CT medical image pulmonary nodule detection methods for generating confrontation network improvement |
| WO2019238104A1 (en) * | 2018-06-16 | 2019-12-19 | 深圳市前海安测信息技术有限公司 | Computer apparatus and method for implementing classification detection of pulmonary nodule images |
-
2020
- 2020-06-29 CN CN202010609059.6A patent/CN111815591B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107016665A (en) * | 2017-02-16 | 2017-08-04 | 浙江大学 | A kind of CT pulmonary nodule detection methods based on depth convolutional neural networks |
| CN108198179A (en) * | 2018-01-03 | 2018-06-22 | 华南理工大学 | A kind of CT medical image pulmonary nodule detection methods for generating confrontation network improvement |
| WO2019238104A1 (en) * | 2018-06-16 | 2019-12-19 | 深圳市前海安测信息技术有限公司 | Computer apparatus and method for implementing classification detection of pulmonary nodule images |
Non-Patent Citations (2)
| Title |
|---|
| 郑光远 ; 刘峡壁 ; 韩光辉 ; .医学影像计算机辅助检测与诊断系统综述.软件学报.2018,(05),全文. * |
| 魏颖 ; 徐心和 ; 贾同 ; 赵大哲 ; .基于多尺度形态学滤波的CT图像疑似肺结节提取.东北大学学报(自然科学版).2008,(03),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111815591A (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022100034A1 (en) | Detection method for malignant region of thyroid cell pathological section based on deep learning | |
| CN109636824B (en) | Multi-target counting method based on image recognition technology | |
| Aminzadeh et al. | Automatic thresholding for defect detection by background histogram mode extents | |
| Bandyopadhyay | Pre-processing of mammogram images | |
| CN111415352B (en) | Cancer metastasis panoramic pathological section analysis method based on deep cascade network | |
| CN111369530A (en) | CT image pulmonary nodule rapid screening method based on deep learning | |
| CN112699845A (en) | Online non-contact palm vein region-of-interest extraction method | |
| CN104036521A (en) | Novel retina eye fundus image segmenting method | |
| CN110175556B (en) | Remote sensing image cloud detection method based on Sobel operator | |
| Sarwar et al. | Segmentation of cervical cells for automated screening of cervical cancer: a review | |
| Unni et al. | Tumour detection in double threshold segmented mammograms using optimized GLCM features fed SVM | |
| Khordehchi et al. | Automatic lung nodule detection based on statistical region merging and support vector machines | |
| CN111105427B (en) | Lung image segmentation method and system based on connected region analysis | |
| Lu et al. | Curve-structure segmentation from depth maps: A cnn-based approach and its application to exploring cultural heritage objects | |
| Fu et al. | Automatic detection of lung nodules using 3D deep convolutional neural networks | |
| CN111815591B (en) | Lung nodule detection method based on CT image | |
| CN114972272A (en) | Grad-CAM-based segmentation method for new coronary pneumonia lesions | |
| Malek et al. | Seed point selection for seed-based region growing in segmenting microcalcifications | |
| CN112819834B (en) | Method and device for classifying stomach pathological images based on artificial intelligence | |
| CN112801992A (en) | Pulmonary nodule image classification method based on 3D residual error network | |
| Rajkumar et al. | Automated mammogram segmentation using seed point identification and modified region growing algorithm | |
| CN111612738B (en) | Chrysanthemum pattern geometric information extraction method based on detection of straight lines and line segments | |
| CN114529570A (en) | Image segmentation method, image identification method, user certificate subsidizing method and system | |
| CN113643308A (en) | Lung image segmentation method and device, storage medium and computer equipment | |
| CN114155205A (en) | Mammary nodule boundary definition judging device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |