CN111882538A - Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information - Google Patents
Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information Download PDFInfo
- Publication number
- CN111882538A CN111882538A CN202010738037.XA CN202010738037A CN111882538A CN 111882538 A CN111882538 A CN 111882538A CN 202010738037 A CN202010738037 A CN 202010738037A CN 111882538 A CN111882538 A CN 111882538A
- Authority
- CN
- China
- Prior art keywords
- lung
- image
- module
- program
- patient
- 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.)
- Withdrawn
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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- 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
-
- 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
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Radiology & Medical Imaging (AREA)
- Library & Information Science (AREA)
- Databases & Information Systems (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention belongs to the technical field of medical diagnosis, and discloses a processing system and an information processing method for distinguishing pulmonary tuberculosis from tumor information, wherein the processing system for distinguishing pulmonary tuberculosis from tumor information comprises: the lung diagnosis system comprises a patient information acquisition module, a lung image acquisition module, a central control module, a lung image enhancement module, an image segmentation module, an image feature extraction module, an image retrieval module, a comparison module, a disease analysis module, a treatment scheme compiling module, a diagnosis report generating module and a display module. According to the invention, interference factors such as human trunk and bed board are automatically removed through the image segmentation module, so that the lung parenchyma image can be rapidly and accurately extracted to better assist a doctor; meanwhile, the efficiency of searching the similar lung images by a doctor user is greatly improved through the image retrieval module, corresponding retrieval feature vectors are obtained according to different types of focuses contained in the images to be retrieved, retrieval of similar sample lung images is carried out based on the retrieval feature vectors, and therefore the accuracy of lung image retrieval is improved.
Description
Technical Field
The invention belongs to the technical field of medical diagnosis, and particularly relates to a processing system and an information processing method for distinguishing pulmonary tuberculosis information from tumor information.
Background
Lung tumors refer to tumors that occur in the lung parenchyma and the lung interstitium. Classified by their origin as primary and secondary (metastatic); classified as benign or malignant according to their biological properties; they are classified by their tissue morphology into epithelial tumors, soft tissue tumors and mesothelioma. Tuberculosis is a chronic infectious disease caused by mycobacterium tuberculosis, can invade many organs, and is the most common pulmonary tuberculosis infection. The bacteria expeller is an important source of infection. The disease does not always occur after the human body is infected with tubercle bacillus, and clinical onset can be caused when the resistance is reduced or the cell-mediated allergic reaction is increased. If the diagnosis can be made in time and the treatment is reasonable, most patients can be cured clinically. Tuberculosis belongs to the genus Mycobacterium of the family Mycobacteriaceae of the order Actinomycetales, and is a pathogenic acid-fast bacterium. Mainly divided into human, cattle, bird, rat, etc. The main pathogenicity of human is human type bacteria, and bovine type bacteria have less infection. The drug resistance of tubercle bacillus to the drug can be formed by the development of the innate drug-resistant bacteria in the flora, and the drug resistance to the drug can be rapidly generated due to the independent use of an antituberculous drug in a human body, namely the drug-resistant bacteria are obtained. Drug-resistant bacteria can cause difficulty in treatment and affect the curative effect. However, the existing lung image segmentation speed is low, and the segmentation accuracy is poor; meanwhile, the lung images of similar focus cases are searched in a manual mode, and the most similar lung images obtained from a large number of historical lung images are low in operation efficiency and accuracy.
In summary, the problems of the prior art are as follows: the existing lung image segmentation speed is low, and the segmentation accuracy is poor; meanwhile, the lung images of similar focus cases are searched in a manual mode, and the most similar lung images obtained from a large number of historical lung images are low in operation efficiency and accuracy.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a processing system and an information processing method for distinguishing tuberculosis from tumor information.
The invention is realized in such a way that a method for processing information for distinguishing tuberculosis from tumor comprises the following steps:
acquiring identity, age, work, address, disease state information and lung image data;
step two, utilizing an image enhancement program to enhance the lung image collected in the step one; using a two-layer feedforward neural network as a deep neural network to calculate the expected enhancement value, the parameter enhancement value of the network will be updated and learned by returning from feedback in the environment, randomly sampling a batch of quadruples from lung image data, and the model minimizes the cost function by periodically updating the parameters:
Lt(θt)＝Es,a[(yt-Q(s,a；θt))2]；
wherein y ist＝r+γmaxa'Q(s',a'；θt-1) The target optimal enhancement is obtained by summing the current return r and the optimal enhancement of the subsequent step;
is expected to be about the sampled quadruple (s, a, s', r);
parameters of the pulmonary image data deep neural network update learning by narrowing the difference before target optimal enhancement by enhancement of the pulmonary image data deep neural network prediction using gradient descent of the cost function:
step three, segmenting the lung parenchyma image by utilizing a segmentation program; labeling the lung contour and the target area, and performing numerical clipping and normalization processing on the image data; training and learning by utilizing a first neural network to obtain a lung contour segmentation model; carrying out gray level histogram statistics on the lung image to be segmented to obtain a gray level histogram;
step four, performing Gaussian smoothing on the gray level histogram to obtain a smoothed gray level histogram;
the step of determining all extreme points in the gray level histogram comprises: determining all extreme points in the smoothed gray level histogram;
the step of determining the maximum point in the gray histogram at which the gray value is closest to 0 includes: determining a maximum value point of which the gray value in the smoothed gray histogram is closest to 0;
determining all extreme points in the gray level histogram, wherein the extreme points comprise a maximum point and a minimum point; determining a maximum value point with the gray value closest to 0 in the gray histogram, and determining the gray value of a first minimum value point positioned on the right side of the maximum value point as a preset gray threshold;
carrying out binarization processing on the lung image to be segmented according to a preset gray threshold value to obtain a first binarized image;
performing negation processing on the first binarized image to obtain a second binarized image; filling holes in the lung region in the second binary image, and performing negation processing on the second binary image after hole filling is completed to obtain a third binary image; removing an interference region in the third binary image to obtain a first mask with a lung region inside; filling the first mask with a lung region to obtain a second mask;
seventhly, performing subtraction operation on the second mask and the first mask to obtain a lung region mask; multiplying the lung region mask and the lung image to be segmented through a segmentation model to obtain a lung parenchymal image;
step eight, extracting the lung image characteristic elements of the patient from the lung parenchyma image obtained in the step seven by using an extraction program; retrieving the lung disease image by using a retrieval procedure; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program; analyzing the type of the lung disease of the patient according to the comparison result by utilizing an analysis program; writing a treatment scheme according to the analysis result by using a writing program;
step nine, generating a diagnosis report of the lung disease of the patient by using a report generation program;
step ten, displaying the collected patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by using a display.
Further, the step six of filling the lung region of the first mask to obtain a second mask includes:
the lung region in the first mask is filled to white to obtain a second mask.
Further, the step six of hole filling of the lung region in the second binarized image includes:
and filling holes in the lung region in the second binary image by adopting morphological closed operation processing of expansion and corrosion.
Further, the image retrieval method in the eight steps is as follows:
1) extracting a lung nodule mixed symptom region in a lung image, and intercepting each single symptom region; extracting semantic features expressing lung nodule symptom information by adopting a parameter sharing-based Convolutional Neural Network (CNN);
2) acquiring the type of a focus contained in the lung image to be retrieved, and acquiring a corresponding retrieval feature vector from the lung image to be retrieved according to the type of the focus;
3) in a sample image library corresponding to the type of the focus, acquiring at least one sample lung image with the highest similarity to the lung image to be retrieved according to the semantic features and the retrieval feature vectors; the sample image library comprises sample lung images belonging to the same lesion type and corresponding sample feature vectors obtained from the sample lung images according to the same lesion type.
Further, the acquiring the type of the lesion included in the lung image to be retrieved includes:
2.1) inputting the lung image to be retrieved into a pre-constructed focus type identification model, and acquiring an identification result of the lung image to be retrieved, wherein the lung image to be retrieved contains the type of the focus belonging to the corresponding focus;
2.2) the pre-constructed focus type identification model comprises an input layer, a convolution layer, a down-sampling layer, a full-connection layer and an output layer; the lesion type identification model inputs the lung image to be retrieved and outputs an identification result of the type of the lesion containing the lesion belonging to the corresponding lesion in the lung image to be retrieved.
Further, the obtaining, in the sample image library corresponding to the type of the lesion, at least one sample lung image with the highest similarity to the lung image to be retrieved according to the semantic features and the retrieval feature vector includes:
acquiring a sample image library corresponding to the type of the focus according to the type of the focus;
and calculating the similarity between the retrieval feature vector and the sample feature vector of each corresponding dimension of each sample lung image in the sample image library, and multiplying the similarity by the weight value of each dimension and summing to obtain at least one sample lung image with the highest summation value.
Further, the type of lesion includes at least one of: pulmonary nodules, pneumonia, bronchiectasis, emphysema, and tuberculosis;
the search feature vector and the sample feature vector corresponding to the lung nodule comprise at least one of: benign and malignant feature vectors, burr feature vectors, calcification feature vectors, vacuole feature vectors and edge definition feature vectors;
the retrieval feature vector and the sample feature vector corresponding to pneumonia comprise at least one of the following: lung texture feature vectors, lung blood vessel thickening degree feature vectors and bronchial thickening degree feature vectors;
the retrieval feature vector and the sample feature vector corresponding to the bronchus extension include at least one of: the characteristic vector of the shape change of the bronchus and whether the grape-shaped characteristic vector exists or not;
the retrieval feature vector and the sample feature vector corresponding to the emphysema comprise at least one of the following: low-density shadow feature vectors and lung structure destruction degree feature vectors in the lung area;
the retrieval feature vector and the sample feature vector corresponding to the tuberculosis comprise at least one of the following: the feature vector of the speckled shadow, the feature vector of the tuberculosphere, the feature vector of pleurisy and the feature vector of the node tuberculosis in the chest.
Another object of the present invention is to provide a processing system for distinguishing pulmonary tuberculosis from tumor information, comprising:
the patient information acquisition module is connected with the central control module and is used for acquiring information such as patient identity, age, work, address, disease state and the like;
the lung image acquisition module is connected with the central control module and is used for acquiring the lung image data of the patient through medical imaging equipment;
the central control module is connected with the patient information acquisition module, the lung image enhancement module, the image segmentation module, the image feature extraction module, the image retrieval module, the comparison module, the disease analysis module, the treatment scheme compiling module, the diagnosis report generating module and the display module and is used for controlling each module to normally work through the host;
the lung image enhancement module is connected with the central control module and is used for enhancing the lung image of the patient through an image enhancement program;
the image segmentation module is connected with the central control module and is used for segmenting the lung parenchyma image through a segmentation program;
the image feature extraction module is connected with the central control module and used for extracting image feature elements of the lung of the patient through an extraction program;
the image retrieval module is connected with the central control module and is used for retrieving the lung disease images through a retrieval program;
the comparison module is connected with the central control module and is used for comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor through a comparison program;
the disease analysis module is connected with the central control module and is used for analyzing the lung disease type of the patient according to the comparison result through an analysis program;
the treatment plan compiling module is connected with the central control module and used for compiling a treatment plan according to the analysis result through a compiling program;
the diagnosis report generation module is connected with the central control module and is used for generating a diagnosis report of the lung disease of the patient through a report generation program;
and the display module is connected with the central control module and is used for displaying the acquired patient information, the lung images, the retrieval results, the comparison results, the analysis results, the treatment schemes and the diagnosis reports through the display.
It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:
firstly, acquiring information such as patient identity, age, work, address, disease state and the like through a patient information acquisition module; acquiring lung image data of a patient by using medical imaging equipment through a lung image acquisition module;
secondly, the central control module utilizes an image enhancement program to enhance the lung image of the patient through the lung image enhancement module; segmenting the lung parenchyma image by an image segmentation module through a segmentation program; extracting the image characteristic elements of the lung of the patient by an image characteristic extraction module by utilizing an extraction program; retrieving the lung disease image by an image retrieval module by utilizing a retrieval program; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program through a comparison module; analyzing the lung disease type of the patient according to the comparison result by using an analysis program through a disease analysis module; compiling a treatment plan according to the analysis result by utilizing a compiling program through a treatment plan compiling module;
then, generating a diagnosis report of the lung disease of the patient by a diagnosis report generating module by using a report generating program;
and finally, the display module is used for displaying the acquired patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by utilizing the display.
It is another object of the present invention to provide a computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:
firstly, acquiring information such as patient identity, age, work, address, disease state and the like through a patient information acquisition module; acquiring lung image data of a patient by using medical imaging equipment through a lung image acquisition module;
secondly, the central control module utilizes an image enhancement program to enhance the lung image of the patient through the lung image enhancement module; segmenting the lung parenchyma image by an image segmentation module through a segmentation program; extracting the image characteristic elements of the lung of the patient by an image characteristic extraction module by utilizing an extraction program; retrieving the lung disease image by an image retrieval module by utilizing a retrieval program; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program through a comparison module; analyzing the lung disease type of the patient according to the comparison result by using an analysis program through a disease analysis module; compiling a treatment plan according to the analysis result by utilizing a compiling program through a treatment plan compiling module;
then, generating a diagnosis report of the lung disease of the patient by a diagnosis report generating module by using a report generating program;
and finally, the display module is used for displaying the acquired patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by utilizing the display.
The invention has the advantages and positive effects that: according to the invention, interference factors such as human trunk and bed board are automatically removed through the image segmentation module, so that the lung parenchyma image can be rapidly and accurately extracted to better assist a doctor; meanwhile, the efficiency of searching the similar lung images by a doctor user is greatly improved through the image retrieval module, corresponding retrieval feature vectors are obtained according to different types of focuses contained in the images to be retrieved, retrieval of similar sample lung images is carried out based on the retrieval feature vectors, and therefore the accuracy of lung image retrieval is further improved.
The invention utilizes an image enhancement program to enhance the collected lung image; the method can meet the application of later-stage related programs, and compared with the prior art, the method for obtaining the image information is clearer.
Drawings
FIG. 1 is a block diagram of a processing system for distinguishing tuberculosis from tumor information according to an embodiment of the present invention.
In fig. 1: 1. a patient information acquisition module; 2. a lung image acquisition module; 3. a central control module; 4. a lung image enhancement module; 5. an image segmentation module; 6. an image feature extraction module; 7. an image retrieval module; 8. a comparison module; 9. a disease analysis module; 10. a treatment plan compiling module; 11. a diagnostic report generation module; 12. and a display module.
Fig. 2 is a flowchart of an image segmentation module segmentation method according to an embodiment of the present invention.
Fig. 3 is a flowchart of a method for filling a lung region of the first mask to obtain a second mask according to an embodiment of the present invention.
Fig. 4 is a flowchart of an image retrieval module retrieval method according to an embodiment of the present invention.
Fig. 5 is a flowchart of a method for obtaining a type of lesion included in a lung image to be retrieved according to an embodiment of the present invention.
Detailed Description
In order to further understand the contents, features and effects of the present invention, the following embodiments are illustrated and described in detail with reference to the accompanying drawings.
The structure of the present invention will be described in detail below with reference to the accompanying drawings.
As shown in fig. 1, a processing system for distinguishing tuberculosis from tumor information according to an embodiment of the present invention includes: the system comprises a patient information acquisition module 1, a lung image acquisition module 2, a central control module 3, a lung image enhancement module 4, an image segmentation module 5, an image feature extraction module 6, an image retrieval module 7, a comparison module 8, a disease analysis module 9, a treatment scheme compiling module 10, a diagnosis report generating module 11 and a display module 12.
The patient information acquisition module 1 is connected with the central control module 3 and is used for acquiring information such as patient identity, age, work, address, disease state and the like;
the lung image acquisition module 2 is connected with the central control module 3 and is used for acquiring the lung image data of the patient through medical imaging equipment;
the central control module 3 is connected with the patient information acquisition module 1, the lung image acquisition module 2, the lung image enhancement module 4, the image segmentation module 5, the image feature extraction module 6, the image retrieval module 7, the comparison module 8, the disease analysis module 9, the treatment scheme compiling module 10, the diagnosis report generating module 11 and the display module 12 and is used for controlling the normal work of each module through a host;
the lung image enhancement module 4 is connected with the central control module 3 and is used for enhancing the lung image of the patient through an image enhancement program;
the image segmentation module 5 is connected with the central control module 3 and is used for segmenting the lung parenchyma image through a segmentation program;
the image feature extraction module 6 is connected with the central control module 3 and is used for extracting image feature elements of the lung of the patient through an extraction program;
the image retrieval module 7 is connected with the central control module 3 and is used for retrieving the lung disease images through a retrieval program;
the comparison module 8 is connected with the central control module 3 and is used for comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor through a comparison program;
the disease analysis module 9 is connected with the central control module 3 and is used for analyzing the lung disease type of the patient according to the comparison result through an analysis program;
the treatment plan compiling module 10 is connected with the central control module 3 and is used for compiling a treatment plan according to the analysis result through a compiling program;
a diagnosis report generating module 11 connected with the central control module 3 and used for generating a diagnosis report of the lung disease of the patient through a report generating program;
and the display module 12 is connected with the central control module 3 and is used for displaying the acquired patient information, the lung images, the retrieval results, the comparison results, the analysis results, the treatment schemes and the diagnosis reports through a display.
As shown in fig. 2, the image segmentation module 5 provided by the present invention comprises the following steps:
s101, labeling a lung contour and a target area, and performing numerical clipping and normalization processing on image data; training and learning by utilizing a first neural network to obtain a lung contour segmentation model; according to a preset gray threshold value, carrying out binarization processing on the lung image to be segmented to obtain a first binarized image;
s102, performing inversion processing on the first binarized image to obtain a second binarized image; filling holes in the lung region in the second binary image, and performing negation processing on the second binary image after hole filling is completed to obtain a third binary image; removing an interference region in the third binary image to obtain a first mask with a lung region inside; filling the first mask with a lung region to obtain a second mask;
s103, performing subtraction operation on the second mask and the first mask to obtain a lung region mask; and multiplying the lung region mask and the lung image to be segmented through a segmentation model to obtain a lung parenchymal image.
Before the binarization processing is carried out on the lung image to be segmented according to the preset gray threshold, the method provided by the invention comprises the following steps:
carrying out gray level histogram statistics on the lung image to be segmented to obtain a gray level histogram;
determining all extreme points in the gray level histogram, wherein the extreme points comprise a maximum value point and a minimum value point; and
determining a maximum value point with the gray value closest to 0 in the gray histogram, and determining the gray value of a first minimum value point positioned on the right side of the maximum value point as a preset gray threshold.
Before determining all extreme points in the gray histogram, the method comprises the following steps:
performing Gaussian smoothing on the gray level histogram to obtain a smoothed gray level histogram;
the step of determining all extreme points in the gray level histogram comprises: determining all extreme points in the smoothed gray level histogram;
the step of determining the maximum point in the gray histogram at which the gray value is closest to 0 includes: and determining the maximum value point of which the gray value in the smoothed gray histogram is closest to 0.
As shown in fig. 3, the step of filling the lung region of the first mask to obtain the second mask provided by the present invention comprises:
s201, filling the lung region in the first mask to be white to obtain a second mask.
The step of filling the hole in the lung region in the second binary image provided by the invention comprises the following steps:
and filling holes in the lung region in the second binary image by adopting morphological closed operation processing of expansion and corrosion.
As shown in fig. 4, the image retrieval module 7 provided by the present invention has the following retrieval method:
s301, extracting a lung nodule mixed symptom region in a lung image, and intercepting each single symptom region; extracting semantic features expressing lung nodule symptom information by adopting a parameter sharing-based Convolutional Neural Network (CNN);
s302, acquiring the type of a focus contained in the lung image to be retrieved, and acquiring a corresponding retrieval feature vector from the lung image to be retrieved according to the type of the focus;
s303, in a sample image library corresponding to the type of the focus, acquiring at least one sample lung image with the highest similarity to the lung image to be retrieved according to the semantic features and the retrieval feature vectors; the sample image library comprises sample lung images belonging to the same lesion type and corresponding sample feature vectors obtained from the sample lung images according to the same lesion type.
As shown in fig. 5, the acquiring of the lung image to be retrieved, which includes the type of lesion, includes:
s401, inputting the lung image to be retrieved into a pre-constructed focus type identification model, and acquiring an identification result of the lung image to be retrieved, wherein the identification result comprises the type of the focus belonging to the corresponding focus;
s402, constructing a focus type recognition model in advance, wherein the focus type recognition model comprises an input layer, a convolution layer, a down-sampling layer, a full-connection layer and an output layer; the lesion type identification model inputs the lung image to be retrieved and outputs an identification result of the type of the lesion containing the lesion belonging to the corresponding lesion in the lung image to be retrieved.
The invention provides a method for acquiring at least one sample lung image with highest similarity with a lung image to be retrieved according to semantic features and retrieval feature vectors in a sample image library corresponding to the type of a focus, which comprises the following steps:
acquiring a sample image library corresponding to the type of the focus according to the type of the focus;
and calculating the similarity between the retrieval feature vector and the sample feature vector of each corresponding dimension of each sample lung image in the sample image library, and multiplying the similarity by the weight value of each dimension and summing to obtain at least one sample lung image with the highest summation value.
The types of lesions provided by the invention include at least one of: pulmonary nodules, pneumonia, bronchiectasis, emphysema, and tuberculosis;
correspondingly, the search feature vector and the sample feature vector corresponding to the lung nodule comprise at least one of: benign and malignant feature vectors, burr feature vectors, calcification feature vectors, vacuole feature vectors and edge definition feature vectors;
the retrieval feature vector and the sample feature vector corresponding to pneumonia comprise at least one of the following: lung texture feature vectors, lung blood vessel thickening degree feature vectors and bronchial thickening degree feature vectors;
the retrieval feature vector and the sample feature vector corresponding to the bronchus extension include at least one of: the characteristic vector of the shape change of the bronchus and whether the grape-shaped characteristic vector exists or not;
the retrieval feature vector and the sample feature vector corresponding to the emphysema comprise at least one of the following: low-density shadow feature vectors and lung structure destruction degree feature vectors in the lung area;
the retrieval feature vector and the sample feature vector corresponding to the tuberculosis comprise at least one of the following: the feature vector of the speckled shadow, the feature vector of the tuberculosphere, the feature vector of pleurisy and the feature vector of the node tuberculosis in the chest.
The information processing method of the present invention includes:
firstly, information such as patient identity, age, work, address, disease state and the like is acquired through a patient information acquisition module 1; acquiring lung image data of a patient by using medical imaging equipment through a lung image acquisition module 2;
secondly, the central control module 3 utilizes an image enhancement program to enhance the lung image of the patient through the lung image enhancement module 4; segmenting the lung parenchyma image by an image segmentation module 5 by utilizing a segmentation program; extracting the image characteristic elements of the lung of the patient by an image characteristic extraction module 6 by using an extraction program; retrieving the lung disease image by an image retrieval module 7 by using a retrieval program; comparing the extracted characteristic elements with the retrieved characteristics of the tuberculosis and the tumor by using a comparison program through a comparison module 8; analyzing the lung disease type of the patient according to the comparison result by using a disease analysis module 9; writing a treatment plan according to the analysis result by using a writing program through the treatment plan writing module 10;
then, generating a diagnosis report of the lung disease of the patient by a diagnosis report generating module 11 by using a report generating program;
finally, the display module 12 is used to display the collected patient information, lung images, search results, comparison results, analysis results, treatment plans and diagnosis reports.
The invention utilizes an image enhancement program to enhance the collected lung image; using a two-layer feedforward neural network as a deep neural network to calculate the expected enhancement value, the parameter enhancement value of the network will be updated and learned by returning from feedback in the environment, randomly sampling a batch of quadruples from lung image data, and the model minimizes the cost function by periodically updating the parameters:
Lt(θt)＝Es,a[(yt-Q(s,a；θt))2]；
wherein y ist＝r+γmaxa'Q(s',a'；θt-1) The target optimal enhancement is obtained by summing the current return r and the optimal enhancement of the subsequent step;
is expected to be about the sampled quadruple (s, a, s', r);
parameters of the pulmonary image data deep neural network update learning by narrowing the difference before target optimal enhancement by enhancement of the pulmonary image data deep neural network prediction using gradient descent of the cost function:
experiments prove that the enhancement method of the invention only needs about 8 times of running time to obtain enhanced images while improving the accuracy. More importantly, the accuracy is improved to 15% on the most challenging benchmark dataset.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent changes and modifications made to the above embodiment according to the technical spirit of the present invention are within the scope of the technical solution of the present invention.
Claims (10)
1. A method for processing information for distinguishing pulmonary tuberculosis from tumor is characterized by comprising the following steps:
acquiring identity, age, work, address, disease state information and lung image data;
step two, utilizing an image enhancement program to enhance the lung image collected in the step one; using a two-layer feedforward neural network as a deep neural network to calculate the expected enhancement value, the parameter enhancement value of the network will be updated and learned by returning from feedback in the environment, randomly sampling a batch of quadruples from lung image data, and the model minimizes the cost function by periodically updating the parameters:
Lt(θt)＝Es,a[(yt-Q(s,a；θt))2]；
wherein y ist＝r+γmaxa'Q(s',a'；θt-1) The target optimal enhancement is obtained by summing the current return r and the optimal enhancement of the subsequent step;
is expected to be about the sampled quadruple (s, a, s', r);
parameters of the pulmonary image data deep neural network update learning by narrowing the difference before target optimal enhancement by enhancement of the pulmonary image data deep neural network prediction using gradient descent of the cost function:
step three, segmenting the lung parenchyma image by utilizing a segmentation program; labeling the lung contour and the target area, and performing numerical clipping and normalization processing on the image data; training and learning by utilizing a first neural network to obtain a lung contour segmentation model; carrying out gray level histogram statistics on the lung image to be segmented to obtain a gray level histogram;
step four, performing Gaussian smoothing on the gray level histogram to obtain a smoothed gray level histogram;
the step of determining all extreme points in the gray level histogram comprises: determining all extreme points in the smoothed gray level histogram;
the step of determining the maximum point in the gray histogram at which the gray value is closest to 0 includes: determining a maximum value point of which the gray value in the smoothed gray histogram is closest to 0;
determining all extreme points in the gray level histogram, wherein the extreme points comprise a maximum point and a minimum point; determining a maximum value point with the gray value closest to 0 in the gray histogram, and determining the gray value of a first minimum value point positioned on the right side of the maximum value point as a preset gray threshold;
carrying out binarization processing on the lung image to be segmented according to a preset gray threshold value to obtain a first binarized image;
performing negation processing on the first binarized image to obtain a second binarized image; filling holes in the lung region in the second binary image, and performing negation processing on the second binary image after hole filling is completed to obtain a third binary image; removing an interference region in the third binary image to obtain a first mask with a lung region inside; filling the first mask with a lung region to obtain a second mask;
seventhly, performing subtraction operation on the second mask and the first mask to obtain a lung region mask; multiplying the lung region mask and the lung image to be segmented through a segmentation model to obtain a lung parenchymal image;
step eight, extracting the lung image characteristic elements of the patient from the lung parenchyma image obtained in the step seven by using an extraction program; retrieving the lung disease image by using a retrieval procedure; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program; analyzing the type of the lung disease of the patient according to the comparison result by utilizing an analysis program; writing a treatment scheme according to the analysis result by using a writing program;
step nine, generating a diagnosis report of the lung disease of the patient by using a report generation program;
step ten, displaying the collected patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by using a display.
2. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 1, wherein the step six of filling the first mask with lung regions to obtain a second mask comprises:
the lung region in the first mask is filled to white to obtain a second mask.
3. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 1, wherein the step six of hole filling the lung area in the second binarized image comprises:
and filling holes in the lung region in the second binary image by adopting morphological closed operation processing of expansion and corrosion.
4. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 1, wherein the image retrieval method of the step eight comprises the following steps:
1) extracting a lung nodule mixed symptom region in a lung image, and intercepting each single symptom region; extracting semantic features expressing lung nodule symptom information by adopting a parameter sharing-based Convolutional Neural Network (CNN);
2) acquiring the type of a focus contained in the lung image to be retrieved, and acquiring a corresponding retrieval feature vector from the lung image to be retrieved according to the type of the focus;
3) in a sample image library corresponding to the type of the focus, acquiring at least one sample lung image with the highest similarity to the lung image to be retrieved according to the semantic features and the retrieval feature vectors; the sample image library comprises sample lung images belonging to the same lesion type and corresponding sample feature vectors obtained from the sample lung images according to the same lesion type.
5. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 4, wherein the obtaining of the pulmonary image to be retrieved including the type of lesion comprises:
2.1) inputting the lung image to be retrieved into a pre-constructed focus type identification model, and acquiring an identification result of the lung image to be retrieved, wherein the lung image to be retrieved contains the type of the focus belonging to the corresponding focus;
2.2) the pre-constructed focus type identification model comprises an input layer, a convolution layer, a down-sampling layer, a full-connection layer and an output layer; the lesion type identification model inputs the lung image to be retrieved and outputs an identification result of the type of the lesion containing the lesion belonging to the corresponding lesion in the lung image to be retrieved.
6. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 4, wherein the obtaining at least one sample lung image with the highest similarity to the lung image to be retrieved according to the semantic features and the retrieval feature vector in the sample image library corresponding to the type of the lesion comprises:
acquiring a sample image library corresponding to the type of the focus according to the type of the focus;
and calculating the similarity between the retrieval feature vector and the sample feature vector of each corresponding dimension of each sample lung image in the sample image library, and multiplying the similarity by the weight value of each dimension and summing to obtain at least one sample lung image with the highest summation value.
7. The information processing method for distinguishing pulmonary tuberculosis from tumor according to claim 4, wherein the type of the lesion includes at least one of: pulmonary nodules, pneumonia, bronchiectasis, emphysema, and tuberculosis;
the search feature vector and the sample feature vector corresponding to the lung nodule comprise at least one of: benign and malignant feature vectors, burr feature vectors, calcification feature vectors, vacuole feature vectors and edge definition feature vectors;
the retrieval feature vector and the sample feature vector corresponding to pneumonia comprise at least one of the following: lung texture feature vectors, lung blood vessel thickening degree feature vectors and bronchial thickening degree feature vectors;
the retrieval feature vector and the sample feature vector corresponding to the bronchus extension include at least one of: the characteristic vector of the shape change of the bronchus and whether the grape-shaped characteristic vector exists or not;
the retrieval feature vector and the sample feature vector corresponding to the emphysema comprise at least one of the following: low-density shadow feature vectors and lung structure destruction degree feature vectors in the lung area;
the retrieval feature vector and the sample feature vector corresponding to the tuberculosis comprise at least one of the following: the feature vector of the speckled shadow, the feature vector of the tuberculosphere, the feature vector of pleurisy and the feature vector of the node tuberculosis in the chest.
8. A system for distinguishing tuberculosis from tumor information, the system comprising:
the patient information acquisition module is connected with the central control module and is used for acquiring information such as patient identity, age, work, address, disease state and the like;
the lung image acquisition module is connected with the central control module and is used for acquiring the lung image data of the patient through medical imaging equipment;
the central control module is connected with the patient information acquisition module, the lung image enhancement module, the image segmentation module, the image feature extraction module, the image retrieval module, the comparison module, the disease analysis module, the treatment scheme compiling module, the diagnosis report generating module and the display module and is used for controlling each module to normally work through the host;
the lung image enhancement module is connected with the central control module and is used for enhancing the lung image of the patient through an image enhancement program;
the image segmentation module is connected with the central control module and is used for segmenting the lung parenchyma image through a segmentation program;
the image feature extraction module is connected with the central control module and used for extracting image feature elements of the lung of the patient through an extraction program;
the image retrieval module is connected with the central control module and is used for retrieving the lung disease images through a retrieval program;
the comparison module is connected with the central control module and is used for comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor through a comparison program;
the disease analysis module is connected with the central control module and is used for analyzing the lung disease type of the patient according to the comparison result through an analysis program;
the treatment plan compiling module is connected with the central control module and used for compiling a treatment plan according to the analysis result through a compiling program;
the diagnosis report generation module is connected with the central control module and is used for generating a diagnosis report of the lung disease of the patient through a report generation program;
and the display module is connected with the central control module and is used for displaying the acquired patient information, the lung images, the retrieval results, the comparison results, the analysis results, the treatment schemes and the diagnosis reports through the display.
9. A computer device, characterized in that the computer device comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of:
firstly, acquiring information such as patient identity, age, work, address, disease state and the like through a patient information acquisition module; acquiring lung image data of a patient by using medical imaging equipment through a lung image acquisition module;
secondly, the central control module utilizes an image enhancement program to enhance the lung image of the patient through the lung image enhancement module; segmenting the lung parenchyma image by an image segmentation module through a segmentation program; extracting the image characteristic elements of the lung of the patient by an image characteristic extraction module by utilizing an extraction program; retrieving the lung disease image by an image retrieval module by utilizing a retrieval program; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program through a comparison module; analyzing the lung disease type of the patient according to the comparison result by using an analysis program through a disease analysis module; compiling a treatment plan according to the analysis result by utilizing a compiling program through a treatment plan compiling module;
then, generating a diagnosis report of the lung disease of the patient by a diagnosis report generating module by using a report generating program;
and finally, the display module is used for displaying the acquired patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by utilizing the display.
10. A computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:
firstly, acquiring information such as patient identity, age, work, address, disease state and the like through a patient information acquisition module; acquiring lung image data of a patient by using medical imaging equipment through a lung image acquisition module;
secondly, the central control module utilizes an image enhancement program to enhance the lung image of the patient through the lung image enhancement module; segmenting the lung parenchyma image by an image segmentation module through a segmentation program; extracting the image characteristic elements of the lung of the patient by an image characteristic extraction module by utilizing an extraction program; retrieving the lung disease image by an image retrieval module by utilizing a retrieval program; comparing the extracted characteristic elements with the characteristics of the searched tuberculosis and tumor by using a comparison program through a comparison module; analyzing the lung disease type of the patient according to the comparison result by using an analysis program through a disease analysis module; compiling a treatment plan according to the analysis result by utilizing a compiling program through a treatment plan compiling module;
then, generating a diagnosis report of the lung disease of the patient by a diagnosis report generating module by using a report generating program;
and finally, the display module is used for displaying the acquired patient information, the lung image, the retrieval result, the comparison result, the analysis result, the treatment scheme and the diagnosis report by utilizing the display.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010738037.XA CN111882538A (en) | 2020-07-28 | 2020-07-28 | Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010738037.XA CN111882538A (en) | 2020-07-28 | 2020-07-28 | Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111882538A true CN111882538A (en) | 2020-11-03 |
Family
ID=73200369
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010738037.XA Withdrawn CN111882538A (en) | 2020-07-28 | 2020-07-28 | Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111882538A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112669314A (en) * | 2021-01-18 | 2021-04-16 | 四川大学 | Lung cancer full-period intelligent management image data platform |
| CN113160153A (en) * | 2021-04-06 | 2021-07-23 | 宁波大学医学院附属医院 | Lung nodule screening method and system based on deep learning technology |
| CN113409924A (en) * | 2021-07-01 | 2021-09-17 | 上海市第一人民医院 | Artificial intelligence-based lung examination image auxiliary marking method and system |
| CN113450899A (en) * | 2021-06-22 | 2021-09-28 | 上海市第一人民医院 | Intelligent diagnosis guiding method based on artificial intelligence cardiopulmonary examination images |
| CN113628187A (en) * | 2021-08-09 | 2021-11-09 | 中南大学湘雅医院 | Automatic migratable artificial intelligence medical image recognition system and recognition method |
| CN117974631A (en) * | 2024-03-27 | 2024-05-03 | 天津市肿瘤医院（天津医科大学肿瘤医院） | Method, system and medium for identifying lung tumor lesion characteristics based on CT image |
-
2020
- 2020-07-28 CN CN202010738037.XA patent/CN111882538A/en not_active Withdrawn
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112669314A (en) * | 2021-01-18 | 2021-04-16 | 四川大学 | Lung cancer full-period intelligent management image data platform |
| CN112669314B (en) * | 2021-01-18 | 2022-02-11 | 四川大学 | Lung cancer full-period intelligent management image data platform |
| CN113160153A (en) * | 2021-04-06 | 2021-07-23 | 宁波大学医学院附属医院 | Lung nodule screening method and system based on deep learning technology |
| CN113450899A (en) * | 2021-06-22 | 2021-09-28 | 上海市第一人民医院 | Intelligent diagnosis guiding method based on artificial intelligence cardiopulmonary examination images |
| CN113409924A (en) * | 2021-07-01 | 2021-09-17 | 上海市第一人民医院 | Artificial intelligence-based lung examination image auxiliary marking method and system |
| CN113628187A (en) * | 2021-08-09 | 2021-11-09 | 中南大学湘雅医院 | Automatic migratable artificial intelligence medical image recognition system and recognition method |
| CN117974631A (en) * | 2024-03-27 | 2024-05-03 | 天津市肿瘤医院（天津医科大学肿瘤医院） | Method, system and medium for identifying lung tumor lesion characteristics based on CT image |
| CN117974631B (en) * | 2024-03-27 | 2024-07-19 | 天津市肿瘤医院（天津医科大学肿瘤医院） | Method, system and medium for identifying lung tumor lesion characteristics based on CT image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yousef et al. | A holistic overview of deep learning approach in medical imaging | |
| CN111882538A (en) | Processing system and information processing method for distinguishing pulmonary tuberculosis information and tumor information | |
| US11423540B2 (en) | Segmentation of anatomical regions and lesions | |
| Sweetlin et al. | Feature selection using ant colony optimization with tandem-run recruitment to diagnose bronchitis from CT scan images | |
| Nithila et al. | Automatic detection of solitary pulmonary nodules using swarm intelligence optimized neural networks on CT images | |
| CN111028224A (en) | Data labeling method, model training device, image processing method, image processing device and storage medium | |
| Perez et al. | Automated lung cancer diagnosis using three-dimensional convolutional neural networks | |
| Jebarani et al. | A novel hybrid K-means and GMM machine learning model for breast cancer detection | |
| Shao et al. | A detection approach for solitary pulmonary nodules based on CT images | |
| CN111709485B (en) | Medical image processing method, device and computer equipment | |
| Orozco et al. | Lung nodule classification in CT thorax images using support vector machines | |
| Sadad et al. | Internet of medical things embedding deep learning with data augmentation for mammogram density classification | |
| Javaid et al. | Multi-organ segmentation of chest CT images in radiation oncology: comparison of standard and dilated UNet | |
| Chunran et al. | Automatic detection and segmentation of lung nodule on ct images | |
| CN112288752B (en) | Full-automatic coronary calcified focus segmentation method based on chest flat scan CT | |
| Kaya | Feature fusion-based ensemble CNN learning optimization for automated detection of pediatric pneumonia | |
| JP2024043567A (en) | Training method, training device, electronic apparatus, storage medium, and pathology image classification system for pathology image feature extractor based upon feature separation | |
| Zhang et al. | LungSeek: 3D Selective Kernel residual network for pulmonary nodule diagnosis | |
| Anter et al. | Deep federated machine learning-based optimization methods for liver tumor diagnosis: A review | |
| Kavitha et al. | Volumetric analysis framework for accurate segmentation and classification (VAF-ASC) of lung tumor from CT images | |
| Sameer et al. | Brain tumor segmentation and classification approach for MR images based on convolutional neural networks | |
| Elizabeth et al. | Computer‐aided diagnosis system for the detection of bronchiectasis in chest computed tomography images | |
| Khattab et al. | Compound W‐Net with Fully Accumulative Residual Connections for Liver Segmentation Using CT Images | |
| Jesi et al. | HRSHO: A hybrid rain optimized spotted hyena optimizer for efficient feature selection in CNN-based sinusitis classification | |
| KN et al. | Comparison of 3-segmentation techniques for intraventricular and intracerebral hemorrhages in unenhanced computed tomography scans |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201103 |