CN113077464A - Medical image processing method, medical image identification method and device - Google Patents
Medical image processing method, medical image identification method and device Download PDFInfo
- Publication number
- CN113077464A CN113077464A CN202110488048.1A CN202110488048A CN113077464A CN 113077464 A CN113077464 A CN 113077464A CN 202110488048 A CN202110488048 A CN 202110488048A CN 113077464 A CN113077464 A CN 113077464A
- Authority
- CN
- China
- Prior art keywords
- medical image
- image
- feature
- gray
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 82
- 238000003672 processing method Methods 0.000 title claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 65
- 238000001914 filtration Methods 0.000 claims abstract description 44
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 39
- 238000009792 diffusion process Methods 0.000 claims abstract description 31
- 238000007637 random forest analysis Methods 0.000 claims abstract description 28
- 230000011218 segmentation Effects 0.000 claims abstract description 28
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 27
- 230000008569 process Effects 0.000 claims description 20
- 238000005070 sampling Methods 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000003902 lesion Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 7
- 238000010586 diagram Methods 0.000 claims description 6
- 238000011176 pooling Methods 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000011524 similarity measure Methods 0.000 claims description 4
- 230000002093 peripheral effect Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 description 4
- 230000002708 enhancing effect Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006996 mental state Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image processing, and discloses a medical image processing method and a medical image identification method, which comprise the following steps: acquiring a medical image, and preprocessing image graying, gray stretching and gray enhancement on the medical image; filtering and denoising the preprocessed medical image by using a diffusion filtering method; carrying out feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest to obtain feature points in the medical image; according to the position of the feature point of the medical image, the self-adaptive segmentation algorithm is utilized to carry out self-adaptive segmentation processing on the medical image to obtain a plurality of medical image blocks; and inputting the medical image block into a medical image recognition model, and recognizing the focus information in the medical image block by using the medical image recognition model. The invention also provides a medical image processing and identifying device. The invention realizes the identification processing of medical images.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a medical image processing method, a medical image recognition method, and a medical image recognition device.
Background
Medical images can provide important clinical diagnostic information, and the identification of a large number of medical conditions is dependent on the identification of the medical images. In recent years, as the performance of equipment is continuously improved, a large amount of medical image data is generated, richer characteristic information is provided, and how to accurately and effectively identify useful information from a large amount of medical images has important significance for disease diagnosis.
Currently, in clinical diagnosis and treatment, the positioning of important feature points of medical images is still mainly performed by manual operation. The manual operation not only needs to consume huge workload and take longer time, but also the accuracy is inevitably influenced by subjective factors such as human factors and doctor states; furthermore, since there are unavoidable but not negligible differences in clinical practice experience and individual mental states among different doctors, the positioning accuracy is also directly affected.
In view of this, how to realize more efficient identification and positioning of feature points in medical images becomes an urgent problem to be solved by those skilled in the art.
Disclosure of Invention
The invention provides a medical image processing method and a medical image identification method, wherein a diffusion filtering method is used for filtering and denoising a preprocessed medical image to obtain a denoised medical image, a random forest-based image feature point positioning algorithm is used for positioning feature points of the medical image to obtain feature points in the medical image, a self-adaptive segmentation algorithm is used for self-adaptively segmenting the medical image according to the feature point positions of the medical image to obtain a plurality of medical image blocks, the medical image blocks are input into a medical image identification model, and the medical image identification model is used for identifying focus information in the medical image blocks.
In order to achieve the above object, the present invention provides a medical image processing method and a medical image recognition method, including:
acquiring a medical image, and performing image graying, gray stretching and gray enhancement pretreatment on the medical image to obtain a pretreated medical image;
filtering and denoising the preprocessed medical image by using a diffusion filtering method to obtain a denoised medical image;
carrying out feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest to obtain feature points in the medical image;
according to the position of the feature point of the medical image, the self-adaptive segmentation algorithm is utilized to carry out self-adaptive segmentation processing on the medical image to obtain a plurality of medical image blocks;
and inputting the medical image block into a medical image recognition model, and recognizing the focus information in the medical image block by using the medical image recognition model.
Optionally, the preprocessing of image graying, grayscale stretching and grayscale enhancement on the medical image includes:
1) solving the maximum value of three components of each pixel in the medical image, and setting the maximum value as the gray value of the pixel point to obtain the gray image of the image, wherein the formula of the gray processing is as follows:
G(i,j)＝max{R(i,j),G(i,j),B(i,j)}
wherein:
(i, j) is a pixel point in the medical image;
r (i, j), G (i, j) and B (i, j) are respectively the values of the pixel point (i, j) in R, G, B three color channels;
g (i, j) is the gray value of the pixel point (i, j);
2) for the gray-scale image, stretching the gray-scale of the image by using a piecewise linear transformation, wherein the formula of the gray-scale stretching is as follows:
wherein:
f (x, y) is a gray scale map;
MAXf(x,y),MINf(x,y) Respectively the maximum gray value and the minimum gray value of the gray map;
3) carrying out gray level enhancement processing on the image after gray level stretching by using a gray level enhancement formula, wherein the gray level enhancement formula is as follows:
wherein:
x is a gray value before gray enhancement, and y is a gray value after gray enhancement;
a and b are parameters, a is 0, and b is 100.
Optionally, the performing, by using a diffusion filtering method, filtering and denoising the preprocessed medical image includes:
filtering and denoising the preprocessed medical image by using a diffusion filtering method, wherein the formula of the diffusion filtering method is as follows:
wherein:
t represents the filtering moment, and if t is 0, the image is represented as an initial medical image;
div represents the divergence;
k denotes a filter parameter, in one embodiment of the invention, when
Indicating that the edge gradient of the current image area is large, which causes diffusion to become unobvious, and then enhancing the image of the area by using a diffusion filtering method formula; when in use
Indicating that the edge gradient of the current image region is small and the diffusion becomes very smooth around x, the diffusion filtering method formula is used for the regionAnd carrying out smoothing processing on the image.
Optionally, the performing feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest includes:
1) randomly selecting a plurality of sampling points from the medical image, and calculating the Haar characteristic of each sampling point:
2) when a first-layer random forest model is trained, the input value is the three-dimensional coordinate offset of the appearance characteristic of a sampling point, and the value of the three-dimensional coordinate offset is defined as the distance value between the sampling point and a target characteristic point; processing each training image by using a first-layer random forest model and generating a deviation vector q ═ q of each feature pointx,qy]Then the feature distance of different feature points is
3) Jointly extracting context features among the feature points from the distance graphs of all the feature points, wherein the extracting process of the context features comprises the following steps: taking the position of a current feature point pixel as a circle center, uniformly dividing the peripheral area of the current feature point pixel by 8 rays, respectively taking different radius values, taking a sampling point on the rays to calculate a Haar feature, and taking the current pixel and the Haar feature in the 3 × 3 neighborhood around the current pixel as the context feature of the current feature point pixel; taking the context characteristics of the image as the input of a second layer of random forest model, and training to obtain two layers of random forest models;
and inputting the medical image to be processed stored in the database into a two-layer random forest model, and outputting the positions of the feature points in the image.
Optionally, the performing, by using an adaptive segmentation algorithm, an adaptive segmentation process on the medical image includes:
1) constructing a characteristic similarity matrix W according to the information of the characteristic points in the medical image:
wherein:
Fithe color feature vector information of the feature point i in the color feature space is the color feature vector average value of all pixel points in the neighborhood of the feature point i;
Xithe position feature vector information of the feature point i in the coordinate space is the position vector mean value of all pixel points in the neighborhood of the feature point i;
σ represents the variance of the color feature vector;
ria pixel area radius representing the feature point i, r representing a radius threshold, is set to 1.5 pixels;
2) constructing an undirected weighted graph G (V, E) from the medical image, wherein the similarity measure between pixel points in the image is the weight of an edge of the graph G connected with a vertex;
3) to D-1/2WD-1/2Decomposing the eigenvalues to generate corresponding eigenvectors, where D is a diagonal matrix and D is a diagonal matrixi＝∑iw (i, j) is the sum of the relationship between the node i and all the nodes in V, the nodes are sorted on the basis, and the feature vectors corresponding to the N larger feature values are selected for clustering, so that a clustering result, namely a final image segmentation result, is obtained.
Optionally, the identifying the lesion information in the medical image block by using the medical image identification model includes:
inputting a medical image block into a medical image recognition model, wherein the starting position of a forward flow is a convolutional layer, processing is carried out through 4 three-dimensional residual error network modules in the subsequent operation process, then down sampling is carried out on a maximum pooling layer, the kernel size of the pooling layer is 2 multiplied by 2 pixels, the kernel sizes of all convolution kernels in the forward flow are 3 multiplied by 3 pixels, and the filling is 1; in the operation process, the step length of each transposed convolution layer is set to be 2, the size of an inner core is 2, and a blank pixel in the convolution process is filled by using a nearest neighbor interpolation method; each residual error network combination unit sends the characteristics to the next residual error network; after the residual error network operation is finished, outputting a feature map of the medical image block;
and inputting the image characteristic diagram into a full connection layer, and outputting the lesion classification information of the medical image block by the full connection layer, wherein the lesion classification information comprises the position, the category and the name of a lesion.
Further, to achieve the above object, the present invention also provides a medical image processing identification apparatus, comprising:
medical image acquisition means for acquiring a medical image;
the data processor is used for preprocessing the medical image by carrying out image graying, gray stretching and gray enhancement to obtain a preprocessed medical image, and carrying out filtering and noise reduction processing on the preprocessed medical image by using a diffusion filtering method to obtain a noise-reduced medical image;
the medical image recognition device is used for positioning the feature points of a medical image by using an image feature point positioning algorithm based on random forests to obtain the feature points in the medical image, performing self-adaptive segmentation processing on the medical image by using a self-adaptive segmentation algorithm according to the feature point positions of the medical image to obtain a plurality of medical image blocks, inputting the medical image blocks into a medical image recognition model, and recognizing focus information in the medical image blocks by using the medical image recognition model.
Furthermore, to achieve the above object, the present invention also provides a computer readable storage medium having stored thereon medical image processing identification program instructions executable by one or more processors to implement the steps of the image processing implementation method as described above.
Compared with the prior art, the invention provides a medical image processing method and a medical image identification method, and the technology has the following advantages:
firstly, the invention uses a diffusion filtering method to carry out filtering and noise reduction processing on a preprocessed medical image, and the formula of the diffusion filtering method is as follows:
wherein: t represents the filtering moment, and if t is 0, the image is represented as an initial medical image; div represents the divergence;
representing a medical image gradient; k denotes a filter parameter when
Indicating that the edge gradient of the current image area is large, which causes diffusion to become unobvious, and then enhancing the image of the area by using a diffusion filtering method formula; when in use
Indicating that the edge gradient of the current image area is small, so that the diffusion becomes very smooth near x, the image of the area is smoothed by using the diffusion filtering method formula. Compared with the traditional algorithm, the algorithm can select different image areas and then automatically judge whether the area is subjected to smoothing processing or enhancement processing, so that the medical image edge information can be better kept while noise is filtered.
Meanwhile, the invention provides a self-adaptive segmentation algorithm, so that the self-adaptive segmentation algorithm is utilized to carry out self-adaptive segmentation processing on the medical image, and the flow of the self-adaptive segmentation algorithm is as follows: constructing a characteristic similarity matrix W according to the information of the characteristic points in the medical image:
wherein: fiThe color feature vector information of the feature point i in the color feature space is the color feature vector average value of all pixel points in the neighborhood of the feature point i; xiThe position feature vector information of the feature point i in the coordinate space is the position vector mean value of all pixel points in the neighborhood of the feature point i; σ represents the variance of the color feature vector; r isiRepresents a feature point iR represents a radius threshold, which is set to 1.5 pixels; constructing an undirected weighted graph G (V, E) from the medical image, wherein the similarity measure between pixel points in the image is the weight of an edge of the graph G connected with a vertex; to D-1/2WD-1/2Decomposing the eigenvalues to generate corresponding eigenvectors, where D is a diagonal matrix and D is a diagonal matrixi＝∑iw (i, j) is the sum of the relationship between the node i and all the nodes in V, the nodes are sorted on the basis, and the feature vectors corresponding to the N larger feature values are selected for clustering, so that a clustering result, namely a final image segmentation result, is obtained. Compared with the traditional algorithm, the algorithm of the invention has the advantages that
The similarity between the characteristic points is calculated by the coordinate distance of the regions, the regions are wrapped by ellipses which are generated by preprocessing and are externally connected with the minimum area, the half of the longer half shaft and the half of the shorter half shaft are used as the radius of the regions, the threshold value is set under the condition, and the condition that the traditional algorithm uses the half of the longer half shaft and the half of the shorter half shaft is avoided
Calculating the similarity of the feature points when
And when sigma is stable, the similarity between the characteristic points i and j is equal to the similarity between the characteristic points m and n, even if the coordinate distance between the characteristic points i and j is far greater than the coordinate distance between the characteristic points m and n, and the pixel regions between the characteristic points m and n are closely adjacent, the algorithm of the invention normalizes the range of the threshold value according to the similarity between the pixel regions, thereby eliminating the defect of manual parameter setting and converting the similarity calculation into the self-adaptive processing.
Drawings
Fig. 1 is a schematic flow chart of a medical image processing method and a medical image recognition method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a medical image processing and identifying apparatus according to an embodiment of the present invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The method comprises the steps of conducting filtering denoising processing on a preprocessed medical image by using a diffusion filtering method to obtain a denoised medical image, conducting feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest to obtain feature points in the medical image, conducting self-adaptive segmentation processing on the medical image by using a self-adaptive segmentation algorithm according to the feature point positions of the medical image to obtain a plurality of medical image blocks, inputting the medical image blocks into a medical image recognition model, and recognizing focus information in the medical image blocks by using the medical image recognition model. Referring to fig. 1, a schematic diagram of a medical image processing method and a medical image recognition method according to an embodiment of the present invention is shown.
In this embodiment, the medical image processing method and the medical image recognition method include:
and S1, acquiring the medical image, and performing preprocessing of image graying, gray stretching and gray enhancement on the medical image to obtain the preprocessed medical image.
Firstly, the invention carries out image graying, grayscale stretching and grayscale enhancement pretreatment on a medical image, wherein the pretreatment process of the image graying, the grayscale stretching and the grayscale enhancement comprises the following steps:
1) solving the maximum value of three components of each pixel in the medical image, and setting the maximum value as the gray value of the pixel point to obtain the gray image of the image, wherein the formula of the gray processing is as follows:
G(i,j)＝max{R(i,j),G(i,j),B(i,j)}
wherein:
(i, j) is a pixel point in the medical image;
r (i, j), G (i, j) and B (i, j) are respectively the values of the pixel point (i, j) in R, G, B three color channels;
g (i, j) is the gray value of the pixel point (i, j);
2) for the gray-scale image, stretching the gray-scale of the image by using a piecewise linear transformation, wherein the formula of the gray-scale stretching is as follows:
wherein:
f (x, y) is a gray scale map;
MAXf(x,y),MINf(x,y)respectively the maximum gray value and the minimum gray value of the gray map;
3) carrying out gray level enhancement processing on the image after gray level stretching by using a gray level enhancement formula, wherein the gray level enhancement formula is as follows:
wherein:
x is a gray value before gray enhancement, and y is a gray value after gray enhancement;
a and b are parameters, a is 0, and b is 100.
And S2, filtering and denoising the preprocessed medical image by using a diffusion filtering method to obtain a denoised medical image.
Furthermore, the invention utilizes a diffusion filtering method to carry out filtering and noise reduction processing on the preprocessed medical image, and the formula of the diffusion filtering method is as follows:
wherein:
t represents the filtering moment, and if t is 0, the image is represented as an initial medical image;
div represents the divergence;
k denotes a filter parameter, in one embodiment of the invention, when
Indicating that the edge gradient of the current image area is large, which causes diffusion to become unobvious, and then enhancing the image of the area by using a diffusion filtering method formula; when in use
Indicating that the edge gradient of the current image area is small, so that the diffusion becomes very smooth near x, the image of the area is smoothed by using the diffusion filtering method formula.
S3, positioning the feature points of the medical image by using an image feature point positioning algorithm based on random forests to obtain the feature points in the medical image.
Furthermore, the invention utilizes an image feature point positioning algorithm based on random forest to position the feature points of the medical image, and the training process of the random forest model comprises the following steps:
1) randomly selecting a plurality of sampling points from the medical image, and calculating the Haar characteristic of each sampling point:
2) when a first-layer random forest model is trained, the input value is the three-dimensional coordinate offset of the appearance characteristic of a sampling point, and the value of the three-dimensional coordinate offset is defined as the distance value between the sampling point and a target characteristic point; processing each training image by using a first-layer random forest model and generating a deviation vector q ═ q of each feature pointx,qy]Then the feature distance of different feature points is
3) Jointly extracting context features among the feature points from the distance graphs of all the feature points, wherein the extracting process of the context features comprises the following steps: taking the position of a current feature point pixel as a circle center, uniformly dividing the peripheral area of the current feature point pixel by 8 rays, respectively taking different radius values, taking a sampling point on the rays to calculate a Haar feature, and taking the current pixel and the Haar feature in the 3 × 3 neighborhood around the current pixel as the context feature of the current feature point pixel; and taking the context characteristics of the image as the input of the second layer of random forest model, and training to obtain a two-layer random forest model.
And inputting the medical image to be processed stored in the database into a two-layer random forest model, and outputting the positions of the feature points in the image.
And S4, performing self-adaptive segmentation processing on the medical image by using a self-adaptive segmentation algorithm according to the position of the feature point of the medical image to obtain a plurality of medical image blocks.
Further, according to the position of the feature point of the medical image, the invention uses the self-adaptive segmentation algorithm to perform self-adaptive segmentation processing on the medical image, and the flow of the self-adaptive segmentation algorithm is as follows:
1) constructing a characteristic similarity matrix W according to the information of the characteristic points in the medical image:
wherein:
Fithe color feature vector information of the feature point i in the color feature space is the color feature vector average value of all pixel points in the neighborhood of the feature point i;
Xithe position feature vector information of the feature point i in the coordinate space is the position vector mean value of all pixel points in the neighborhood of the feature point i;
σ represents the variance of the color feature vector;
ria pixel area radius representing the feature point i, r representing a radius threshold, is set to 1.5 pixels;
2) constructing an undirected weighted graph G (V, E) from the medical image, wherein the similarity measure between pixel points in the image is the weight of an edge of the graph G connected with a vertex;
3) to D-1/2WD-1/2Decomposing the eigenvalues to generate corresponding eigenvectors, where D is a diagonal matrix and D is a diagonal matrixi＝∑iw (i, j) is the sum of the relationship between the node i and all the nodes in V, the nodes are sorted on the basis, and the feature vectors corresponding to the N larger feature values are selected for clustering, so that a clustering result, namely a final image segmentation result, is obtained.
And S5, inputting the medical image blocks into the medical image recognition model, and recognizing the focus information in the medical image blocks by using the medical image recognition model.
Furthermore, the medical image block is input into the medical image identification model, and the medical image identification model is utilized to identify the focus information in the medical image block;
the process of the medical image identification model for identifying the focus information of the medical image block comprises the following steps:
inputting a medical image block into a medical image recognition model, wherein the starting position of a forward flow is a convolutional layer, processing is carried out through 4 three-dimensional residual error network modules in the subsequent operation process, then down sampling is carried out on a maximum pooling layer, the kernel size of the pooling layer is 2 multiplied by 2 pixels, the kernel sizes of all convolution kernels in the forward flow are 3 multiplied by 3 pixels, and the filling is 1; in the operation process, the step length of each transposed convolution layer is set to be 2, the size of an inner core is 2, and a blank pixel in the convolution process is filled by using a nearest neighbor interpolation method; each residual error network combination unit sends the characteristics to the next residual error network; after the residual error network operation is finished, outputting a feature map of the medical image block;
and inputting the image characteristic diagram into a full connection layer, and outputting the lesion classification information of the medical image block by the full connection layer, wherein the lesion classification information comprises the position, the category and the name of a lesion.
The following describes embodiments of the present invention through an algorithmic experiment and tests of the inventive treatment method. The hardware test environment of the algorithm of the invention is as follows: inter (R) core (TM) i7-6700KCPU with software Matlab2018 a; the comparison method is a VGG 16-based medical image identification method and an RNN-based medical image identification method.
In the algorithm experiment, the data set is 10G of medical image data. In the experiment, the medical image data is input into the algorithm model, and the accuracy of medical image identification is used as an evaluation index of algorithm feasibility, wherein the higher the accuracy of medical image identification is, the higher the effectiveness and the feasibility of the algorithm are.
According to the experimental result, the medical image identification accuracy of the medical image identification method based on VGG16 is 85.33%, the medical image identification accuracy of the medical image identification method based on RNN is 83.22%, and the medical image identification accuracy of the medical image identification method is 88.72%.
The invention also provides a medical image processing and identifying device. Referring to fig. 2, a schematic diagram of an internal structure of a medical image processing and identifying apparatus according to an embodiment of the present invention is provided.
In the present embodiment, the medical image processing and recognition apparatus 1 includes at least a medical image acquisition apparatus 11, a data processor 12, a medical image recognition apparatus 13, a communication bus 14, and a network interface 15.
The medical image acquiring apparatus 11 may be a Personal Computer (PC), a terminal device such as a smart phone, a tablet Computer, or a mobile Computer, or may be a server.
The data processor 12 includes at least one type of readable storage medium including flash memory, hard disks, multi-media cards, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disks, optical disks, and the like. The data processor 12 may in some embodiments be an internal storage unit of the medical image processing recognition apparatus 1, for example a hard disk of the medical image processing recognition apparatus 1. The data processor 12 may also be an external storage device of the medical image processing identification apparatus 1 in other embodiments, such as a plug-in hard disk provided on the medical image processing identification apparatus 1, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like. Further, the data processor 12 may also include both an internal storage unit and an external storage device of the medical image processing identification apparatus 1. The data processor 12 may be used not only to store application software installed in the medical image processing recognition apparatus 1 and various types of data, but also to temporarily store data that has been output or is to be output.
The medical image recognition device 13 may be, in some embodiments, a Central Processing Unit (CPU), controller, microcontroller, microprocessor or other data Processing chip for executing program code stored in the data processor 12 or Processing data, such as medical image Processing recognition program instructions 16.
The communication bus 14 is used to enable connection communication between these components.
The network interface 15 may optionally comprise a standard wired interface, a wireless interface (e.g. WI-FI interface), typically used for establishing a communication connection between the apparatus 1 and other electronic devices.
Optionally, the medical image processing and recognition apparatus 1 may further comprise a user interface, the user interface may comprise a Display (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface may further comprise a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the medical image processing identification apparatus 1 and for displaying a visual user interface.
Fig. 2 only shows the medical image processing identification apparatus 1 with the components 11-15, and it will be understood by those skilled in the art that the structure shown in fig. 1 does not constitute a limitation of the medical image processing identification apparatus 1, and may comprise fewer or more components than those shown, or some components may be combined, or a different arrangement of components.
In the embodiment of the medical image processing identification apparatus 1 shown in fig. 2, the data processor 12 has stored therein medical image processing identification program instructions 16; the steps of the medical image recognition apparatus 13 executing the medical image processing recognition program instructions 16 stored in the data processor 12 are the same as the medical image processing method, the implementation method of the medical image recognition method, and are not described here.
Furthermore, an embodiment of the present invention also provides a computer-readable storage medium having stored thereon medical image processing identification program instructions executable by one or more processors to implement the following:
acquiring a medical image, and performing image graying, gray stretching and gray enhancement pretreatment on the medical image to obtain a pretreated medical image;
filtering and denoising the preprocessed medical image by using a diffusion filtering method to obtain a denoised medical image;
carrying out feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest to obtain feature points in the medical image;
according to the position of the feature point of the medical image, the self-adaptive segmentation algorithm is utilized to carry out self-adaptive segmentation processing on the medical image to obtain a plurality of medical image blocks;
and inputting the medical image block into a medical image recognition model, and recognizing the focus information in the medical image block by using the medical image recognition model.
It should be noted that the above-mentioned numbers of the embodiments of the present invention are merely for description, and do not represent the merits of the embodiments. And the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (8)
1. A medical image processing method and a medical image recognition method are characterized in that the method comprises the following steps:
acquiring a medical image, and performing image graying, gray stretching and gray enhancement pretreatment on the medical image to obtain a pretreated medical image;
filtering and denoising the preprocessed medical image by using a diffusion filtering method to obtain a denoised medical image;
carrying out feature point positioning on the medical image by using an image feature point positioning algorithm based on a random forest to obtain feature points in the medical image;
according to the position of the feature point of the medical image, the self-adaptive segmentation algorithm is utilized to carry out self-adaptive segmentation processing on the medical image to obtain a plurality of medical image blocks;
and inputting the medical image block into a medical image recognition model, and recognizing the focus information in the medical image block by using the medical image recognition model.
2. The medical image processing method and the medical image recognition method according to claim 1, wherein the preprocessing of image graying, grayscale stretching and grayscale enhancement on the medical image comprises:
1) solving the maximum value of three components of each pixel in the medical image, and setting the maximum value as the gray value of the pixel point to obtain the gray image of the image, wherein the formula of the gray processing is as follows:
G(i，j)＝max{R(i，j)，G(i，j)，B(i，j)}
wherein:
(i, j) is a pixel point in the medical image;
r (i, j), G (i, j) and B (i, j) are respectively the values of the pixel point (i, j) in R, G, B three color channels;
g (i, j) is the gray value of the pixel point (i, j);
2) for the gray-scale image, stretching the gray-scale of the image by using a piecewise linear transformation, wherein the formula of the gray-scale stretching is as follows:
wherein:
f (x, y) is a gray scale map;
MAXf(x，y)，MINf(x，y)respectively the maximum gray value and the minimum gray value of the gray map;
3) carrying out gray level enhancement processing on the image after gray level stretching by using a gray level enhancement formula, wherein the gray level enhancement formula is as follows:
wherein:
x is a gray value before gray enhancement, and y is a gray value after gray enhancement;
a and b are parameters, a is 0, and b is 100.
3. A medical image processing method and a medical image recognition method according to claim 2, wherein the filtering and denoising processing of the preprocessed medical image by using the diffusion filtering method includes:
filtering and denoising the preprocessed medical image by using a diffusion filtering method, wherein the formula of the diffusion filtering method is as follows:
wherein:
t represents the filtering moment, and if t is 0, the image is represented as an initial medical image;
div represents the divergence;
k denotes a filter parameter.
4. A medical image processing method and a medical image recognition method as claimed in claim 3, wherein the feature point locating of the medical image by using the image feature point locating algorithm based on the random forest comprises:
1) randomly selecting a plurality of sampling points from the medical image, and calculating the Haar characteristic of each sampling point:
2) in the first trainingWhen a layer is random forest models, the input value is the three-dimensional coordinate offset of the appearance characteristic of the sampling point, and the value of the three-dimensional coordinate offset is defined as the distance value between the sampling point and the target characteristic point; processing each training image by using a first-layer random forest model and generating a deviation vector q ═ q of each feature pointx，qy]Then the feature distance of different feature points is
3) Jointly extracting context features among the feature points from the distance graphs of all the feature points, wherein the extracting process of the context features comprises the following steps: taking the position of a current feature point pixel as a circle center, uniformly dividing the peripheral area of the current feature point pixel by 8 rays, respectively taking different radius values, taking a sampling point on the rays to calculate a Haar feature, and taking the current pixel and the Haar feature in the 3 × 3 neighborhood around the current pixel as the context feature of the current feature point pixel; taking the context characteristics of the image as the input of a second layer of random forest model, and training to obtain two layers of random forest models;
and inputting the medical image to be processed stored in the database into a two-layer random forest model, and outputting the positions of the feature points in the image.
5. The medical image processing method and the medical image identification method according to claim 4, wherein the adaptive segmentation processing of the medical image by the adaptive segmentation algorithm comprises:
1) constructing a characteristic similarity matrix W according to the information of the characteristic points in the medical image:
wherein:
Fithe color feature vector information of the feature point i in the color feature space is the color feature vector average value of all pixel points in the neighborhood of the feature point i;
Xiis a characteristic pointi, position feature vector information of the coordinate space, namely the position vector mean value of all pixel points in the neighborhood of the feature point i;
σ represents the variance of the color feature vector;
ria pixel area radius representing the feature point i, r representing a radius threshold, is set to 1.5 pixels;
2) constructing an undirected weighted graph G (V, E) from the medical image, wherein the similarity measure between pixel points in the image is the weight of an edge of the graph G connected with a vertex;
3) to D-1/2WD-1/2Decomposing the eigenvalues to generate corresponding eigenvectors, where D is a diagonal matrix and D is a diagonal matrixi＝∑iw (i, j) is the sum of the relationship between the node i and all the nodes in V, the nodes are sorted on the basis, and the feature vectors corresponding to the N larger feature values are selected for clustering, so that a clustering result, namely a final image segmentation result, is obtained.
6. The medical image processing method and the medical image identification method according to claim 5, wherein the identifying the lesion information in the medical image block by using the medical image identification model comprises:
inputting a medical image block into a medical image recognition model, wherein the starting position of a forward flow is a convolutional layer, processing is carried out through 4 three-dimensional residual error network modules in the subsequent operation process, then down sampling is carried out on a maximum pooling layer, the kernel size of the pooling layer is 2 multiplied by 2 pixels, the kernel sizes of all convolution kernels in the forward flow are 3 multiplied by 3 pixels, and the filling is 1; in the operation process, the step length of each transposed convolution layer is set to be 2, the size of an inner core is 2, and a blank pixel in the convolution process is filled by using a nearest neighbor interpolation method; each residual error network combination unit sends the characteristics to the next residual error network; after the residual error network operation is finished, outputting a feature map of the medical image block;
and inputting the image characteristic diagram into a full connection layer, and outputting the lesion classification information of the medical image block by the full connection layer, wherein the lesion classification information comprises the position, the category and the name of a lesion.
7. A medical image processing identification apparatus, characterized in that the apparatus comprises:
medical image acquisition means for acquiring a medical image;
the data processor is used for preprocessing the medical image by carrying out image graying, gray stretching and gray enhancement to obtain a preprocessed medical image, and carrying out filtering and noise reduction processing on the preprocessed medical image by using a diffusion filtering method to obtain a noise-reduced medical image;
the medical image recognition device is used for positioning the feature points of a medical image by using an image feature point positioning algorithm based on random forests to obtain the feature points in the medical image, performing self-adaptive segmentation processing on the medical image by using a self-adaptive segmentation algorithm according to the feature point positions of the medical image to obtain a plurality of medical image blocks, inputting the medical image blocks into a medical image recognition model, and recognizing focus information in the medical image blocks by using the medical image recognition model.
8. A computer readable storage medium having stored thereon medical image processing identification program instructions executable by one or more processors to implement steps of an implementation method of image processing as described above.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110488048.1A CN113077464A (en) | 2021-05-06 | 2021-05-06 | Medical image processing method, medical image identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110488048.1A CN113077464A (en) | 2021-05-06 | 2021-05-06 | Medical image processing method, medical image identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113077464A true CN113077464A (en) | 2021-07-06 |
Family
ID=76616616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110488048.1A Pending CN113077464A (en) | 2021-05-06 | 2021-05-06 | Medical image processing method, medical image identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113077464A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114742836A (en) * | 2022-06-13 | 2022-07-12 | 浙江太美医疗科技股份有限公司 | Medical image processing method and device and computer equipment |
| CN117541484A (en) * | 2024-01-09 | 2024-02-09 | 巨野县龙祥面粉有限公司 | Image enhancement method for detecting bran star of flour |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845098A (en) * | 2017-11-14 | 2018-03-27 | 南京理工大学 | Liver cancer image full-automatic partition method based on random forest and fuzzy clustering |
| CN109214428A (en) * | 2018-08-13 | 2019-01-15 | 平安科技（深圳）有限公司 | Image partition method, device, computer equipment and computer storage medium |
| CN109920002A (en) * | 2019-05-15 | 2019-06-21 | 南京邮电大学 | Characteristic point positioning method in cephalometry image based on three-dimensional random forest model |
| WO2020151307A1 (en) * | 2019-01-23 | 2020-07-30 | 平安科技（深圳）有限公司 | Automatic lesion recognition method and device, and computer-readable storage medium |
| CN111951283A (en) * | 2020-08-12 | 2020-11-17 | 汪礼君 | Medical image identification method and system based on deep learning |
| CN112263224A (en) * | 2020-12-24 | 2021-01-26 | 南京信息工程大学 | Medical information processing method based on FPGA edge calculation |
| CN112396564A (en) * | 2020-11-19 | 2021-02-23 | 汪金玲 | Product packaging quality detection method and system based on deep learning |
| CN112750120A (en) * | 2021-01-20 | 2021-05-04 | 唐美荣 | Medical image acquisition management method and system |
-
2021
- 2021-05-06 CN CN202110488048.1A patent/CN113077464A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845098A (en) * | 2017-11-14 | 2018-03-27 | 南京理工大学 | Liver cancer image full-automatic partition method based on random forest and fuzzy clustering |
| CN109214428A (en) * | 2018-08-13 | 2019-01-15 | 平安科技（深圳）有限公司 | Image partition method, device, computer equipment and computer storage medium |
| WO2020151307A1 (en) * | 2019-01-23 | 2020-07-30 | 平安科技（深圳）有限公司 | Automatic lesion recognition method and device, and computer-readable storage medium |
| CN109920002A (en) * | 2019-05-15 | 2019-06-21 | 南京邮电大学 | Characteristic point positioning method in cephalometry image based on three-dimensional random forest model |
| CN111951283A (en) * | 2020-08-12 | 2020-11-17 | 汪礼君 | Medical image identification method and system based on deep learning |
| CN112396564A (en) * | 2020-11-19 | 2021-02-23 | 汪金玲 | Product packaging quality detection method and system based on deep learning |
| CN112263224A (en) * | 2020-12-24 | 2021-01-26 | 南京信息工程大学 | Medical information processing method based on FPGA edge calculation |
| CN112750120A (en) * | 2021-01-20 | 2021-05-04 | 唐美荣 | Medical image acquisition management method and system |
Non-Patent Citations (1)
| Title |
|---|
| 秦臻等: "基于双层回归森林模型的头影测量图像结构特征点自动定位", 应用科学学报, vol. 37, no. 4, pages 1 - 9 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114742836A (en) * | 2022-06-13 | 2022-07-12 | 浙江太美医疗科技股份有限公司 | Medical image processing method and device and computer equipment |
| CN114742836B (en) * | 2022-06-13 | 2022-09-09 | 浙江太美医疗科技股份有限公司 | Medical image processing method and device and computer equipment |
| CN117541484A (en) * | 2024-01-09 | 2024-02-09 | 巨野县龙祥面粉有限公司 | Image enhancement method for detecting bran star of flour |
| CN117541484B (en) * | 2024-01-09 | 2024-05-14 | 巨野县龙祥面粉有限公司 | Image enhancement method for detecting bran star of flour |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110516544B (en) | Face recognition method and device based on deep learning and computer readable storage medium | |
| CN113159147B (en) | Image recognition method and device based on neural network and electronic equipment | |
| CN110264443B (en) | Fundus image lesion labeling method, device and medium based on feature visualization | |
| CN108596258A (en) | A kind of image classification method based on convolutional neural networks random pool | |
| CN110110600B (en) | Eye OCT image focus identification method, device and storage medium | |
| CN108875734B (en) | Liver canceration positioning method, device and storage medium | |
| WO2021136368A1 (en) | Method and apparatus for automatically detecting pectoralis major region in molybdenum target image | |
| CN113077464A (en) | Medical image processing method, medical image identification method and device | |
| CN110570394B (en) | Medical image segmentation method, device, equipment and storage medium | |
| CN112464803A (en) | Image comparison method and device | |
| CN111666813B (en) | Subcutaneous sweat gland extraction method of three-dimensional convolutional neural network based on non-local information | |
| CN110245621A (en) | Face identification device and image processing method, Feature Selection Model, storage medium | |
| CN111951283A (en) | Medical image identification method and system based on deep learning | |
| CN115620384A (en) | Model training method, fundus image prediction method and device | |
| CN111881789A (en) | Skin color identification method and device, computing equipment and computer storage medium | |
| CN111968087B (en) | Plant disease area detection method | |
| CN114170440A (en) | Method and device for determining image feature points, computer equipment and storage medium | |
| CN111598144B (en) | Training method and device for image recognition model | |
| CN111368602A (en) | Face image blurring degree evaluation method and device, readable storage medium and equipment | |
| CN113435377A (en) | Medical palm vein image acquisition monitoring method and system | |
| CN115294426B (en) | Method, device and equipment for tracking interventional medical equipment and storage medium | |
| CN113780555B (en) | Model training method, device, equipment and storage medium based on data enhancement | |
| CN113158960A (en) | Medical image recognition model construction and recognition method and device | |
| CN114841974A (en) | Nondestructive testing method and system for internal structure of fruit, electronic equipment and medium | |
| CN112330652A (en) | Chromosome recognition method and device based on deep learning and computer equipment |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210706 |