CN116309187B - Intelligent enhancement method for medical image of children pneumonia - Google Patents
Intelligent enhancement method for medical image of children pneumonia Download PDFInfo
- Publication number
- CN116309187B CN116309187B CN202310538317.XA CN202310538317A CN116309187B CN 116309187 B CN116309187 B CN 116309187B CN 202310538317 A CN202310538317 A CN 202310538317A CN 116309187 B CN116309187 B CN 116309187B
- Authority
- CN
- China
- Prior art keywords
- sliding window
- similarity
- image
- label
- regularity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 206010035664 Pneumonia Diseases 0.000 title claims abstract description 31
- 230000003044 adaptive effect Effects 0.000 claims abstract description 16
- 238000002604 ultrasonography Methods 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 8
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 5
- 230000003014 reinforcing effect Effects 0.000 claims description 3
- 230000001502 supplementing effect Effects 0.000 claims description 3
- 210000001519 tissue Anatomy 0.000 description 11
- 238000004364 calculation method Methods 0.000 description 6
- 210000004072 lung Anatomy 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 210000000988 bone and bone Anatomy 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 210000000779 thoracic wall Anatomy 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000000038 chest Anatomy 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007306 functionalization reaction Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000011164 ossification Effects 0.000 description 1
- 210000004224 pleura Anatomy 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration by the use of histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A50/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE in human health protection, e.g. against extreme weather
- Y02A50/30—Against vector-borne diseases, e.g. mosquito-borne, fly-borne, tick-borne or waterborne diseases whose impact is exacerbated by climate change
Abstract
The invention relates to the technical field of image processing, in particular to an intelligent enhancement method for medical images of children pneumonia. According to the method, an initial sliding window is arranged in an ultrasonic image of the pneumonia of the child, an adaptive sliding window is built according to gradient distribution in the initial sliding window, a contrast sliding window corresponding to the adaptive sliding window is built, similarity between two sliding window areas is analyzed, and a similarity label of the corresponding area is given according to the similarity. And setting a regularity operator in a direction perpendicular to the sliding direction of the sliding window, and adjusting the similarity label according to the regularity in the area of the operator to obtain a regular label image. And further obtaining a region to be enhanced, and carrying out targeted enhancement on the region to be enhanced to obtain an enhanced ultrasonic image. According to the invention, through carrying out regularity analysis on the ultrasonic image, irrelevant background information is removed, so that the quality of the enhanced ultrasonic image is higher.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an intelligent enhancement method for medical images of children pneumonia.
Background
Because the chest area of the child is small, the chest wall is thin, and rib ossification is not completed, the ultrasound image of the child lung contains more organ tissue parts, and the information is more abundant compared with the ultrasound image of the adult lung. In the teaching scene, in order to enable medical staff to learn the state of important tissues in the children pneumonia medical image, clear children pneumonia ultrasonic images are required to be used as teaching images for the medical staff to learn.
In the prior art, in order to obtain a clear ultrasonic image, the original ultrasonic image needs to be enhanced, so that the contrast in the image is improved, and detail information is increased. However, because the information in the ultrasound image of the children's pneumonia is rich, a large amount of useless background information exists, if global enhancement is used, not only meaningless calculation amount can be generated, but also the characteristics of the useful information can be covered due to the influence of the background, so that the image quality is affected.
Disclosure of Invention
In order to solve the technical problem that the enhanced image quality is poor due to a large amount of useless background information in the children pneumonia ultrasonic image, the invention aims to provide an intelligent enhancement method for the children pneumonia medical image, which adopts the following technical scheme:
the invention provides an intelligent enhancement method for medical images of children pneumonia, which comprises the following steps:
obtaining a gray image of an ultrasonic image of the children's pneumonia; setting an initial sliding window in the gray level image according to a preset initial size; determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the initial sliding window in the horizontal direction and the gradient distribution of the pixel points in the vertical direction, adjusting the initial sliding window according to the size to be adjusted, obtaining an adaptive sliding window, and constructing a contrast sliding window in a preset sliding direction according to the adaptive sliding window, wherein the contrast sliding window has the same size and is co-limited with the adaptive sliding window;
obtaining the similarity between the corresponding areas of the two sliding windows according to the gray value difference of the same position between the self-adaptive sliding window and the contrast sliding window; according to the similarity, a similarity label is given to the areas corresponding to the self-adaptive sliding window and the contrast sliding window in the gray level image; continuously constructing the initial sliding window for the area without the similarity label and obtaining the similarity label until the similarity label exists at each position on the gray level image, so as to obtain a label image;
traversing the whole label image along the direction perpendicular to the sliding direction by utilizing a regularity operator of a preset shape in the label image; obtaining regional regularity according to the distribution of similarity label categories in a region corresponding to the regularity operator, and adjusting the similarity label of the region center corresponding to the regularity operator according to the regional regularity; obtaining a rule tag image, and screening out a rule area and an area to be enhanced according to the similarity tag in the rule tag image;
and reinforcing the region to be reinforced corresponding to the gray level image to obtain an enhanced ultrasonic image.
Further, the determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the horizontal direction and the gradient distribution in the vertical direction in the initial sliding window includes:
the dimension to be adjusted comprises a length to be adjusted and a width to be adjusted;
taking the left upper corner of the initial sliding window as a starting point;
in the row where the starting point is located, obtaining the gradient of each pixel point in the horizontal direction, selecting the point with the largest gradient as a first cut-off point, and obtaining the width to be adjusted according to the coordinate distance between the starting point and the first cut-off point; if the starting point and the first cut-off point are the same point, the width to be adjusted is 1;
in the column where the starting point is located, obtaining the gradient of each pixel point in the vertical direction, selecting the point with the largest gradient as a second cut-off point, and obtaining the length to be adjusted according to the coordinate distance between the starting point and the second cut-off point; and if the starting point and the second cut-off point are the same point, the length to be adjusted is 1.
Further, the obtaining the similarity between the two sliding window corresponding areas includes:
dividing the gray level value in the gray level image into a preset number of gray levels, and accumulating gray level differences at the same position between the self-adaptive sliding window and the contrast sliding window to obtain an overall difference value; and mapping and normalizing the negative correlation of the integral difference value to obtain the similarity.
Further, the method comprises the step of assigning a similarity label to the areas corresponding to the adaptive sliding window and the contrast sliding window in the gray level image according to the similarity:
the similarity labels comprise similar labels and dissimilar labels;
if the similarity is in a preset similarity interval, giving the similarity labels to all pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window;
and if the similarity is not in the preset similarity interval, giving the dissimilar labels to all the pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window.
Further, the method for acquiring the tag image comprises the following steps:
starting to construct the initial sliding window at the upper left corner of the gray level image and obtaining the similarity label; removing the area with the similarity label in the gray level image every time the giving process of the similarity label is executed to obtain a to-be-processed area, and continuing to construct the initial sliding window from the upper left corner in the to-be-processed area and obtaining the similarity label until the similarity label exists at each position on the gray level image to obtain the label image; if an area which does not contain pixel points exists in the self-adaptive sliding window or the contrast sliding window in the giving process of the similarity label, supplementing 0 to the corresponding area.
Further, the obtaining the region regularity according to the distribution of the similarity tag categories includes:
and counting the number proportion of the similar labels in the area corresponding to the regularity operator, and taking the number proportion as the area regularity.
Further, the adjusting the similarity label of the region center corresponding to the rule operator according to the region rule comprises:
if the regional regularity is smaller than a preset regularity threshold, setting a similarity label of the regional center corresponding to the regularity operator as the dissimilar label;
if the regional regularity is equal to a preset regularity threshold, not changing a similarity label of the regional center corresponding to the regularity operator;
and if the regional regularity is greater than a preset regularity threshold, setting a similarity label of the regional center corresponding to the regularity operator as the similarity label.
Further, the screening the rule area and the area to be enhanced according to the similarity label in the rule label image includes:
and taking the region corresponding to the similar label in the regular label image as a regular region, and taking the region corresponding to the dissimilar label as the region to be enhanced.
Further, the obtaining the enhanced ultrasound image includes:
in the gray level image, removing the area corresponding to the regular area to obtain an image to be enhanced, wherein the image to be enhanced only comprises the area to be enhanced; performing histogram equalization processing on the image to be enhanced to obtain an enhanced image; and splicing the enhanced image with the area corresponding to the regular area in the gray level image to obtain the enhanced ultrasonic image.
The invention has the following beneficial effects:
according to the invention, more tissue information is contained in the ultrasound image of the children pneumonia, so that the size of the initial sliding window is adjusted according to gradient distribution in different directions, and the self-adaptive sliding window contains the regional information of only one tissue as much as possible. And further constructing a comparison sliding window of the self-adaptive sliding window, and judging whether the corresponding areas of the two sliding windows are regular areas or not by comparing the similarity between the two sliding window areas. Further, in the process of giving the similarity labels, errors may occur in the similarity labels due to the special positions of the self-adaptive sliding windows, so that the similarity labels in the label images are adjusted by using a regularity operator in the label images, and the regular label images are obtained. In the regular label image, the regular region is a background region with obvious characteristics and no need of targeted enhancement, and the region to be enhanced is a region needing enhancement, so that the region to be enhanced is targeted enhanced, and the enhanced ultrasonic image is obtained. According to the method, the regular areas are extracted and screened out, so that the pertinence enhancement of the local areas of the ultrasonic image is realized, the quality of the enhanced ultrasonic image is higher, and the efficiency of the enhancement process is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic illustration of a method for intelligently enhancing medical images of pediatric pneumonia according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a gray scale image of an ultrasound image of children's pneumonia according to an embodiment of the present invention;
fig. 3 is a schematic view of an enhanced ultrasound image according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the intelligent enhancement method for the medical image of children pneumonia according to the invention, which is provided by the invention, with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a specific scheme of an intelligent enhancement method for medical images of children pneumonia, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for intelligently enhancing medical images of pneumonia in children according to an embodiment of the invention is shown, where the method includes:
step S1: obtaining a gray image of an ultrasonic image of the children's pneumonia; setting an initial sliding window in the gray level image according to a preset initial size; determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the initial sliding window in the horizontal direction and the gradient distribution in the vertical direction, adjusting the initial sliding window according to the size to be adjusted, obtaining an adaptive sliding window, and constructing a contrast sliding window in the transverse direction according to the adaptive sliding window; the contrast sliding window is the same size and is co-bordered with the adaptive sliding window.
Because the embodiment of the invention is implemented in the medical teaching, the children's pneumonia ultrasonic image can be directly called out in a hospital or college database, and the children's pneumonia ultrasonic image is subjected to gray processing in order to facilitate the subsequent image processing. In the embodiment of the present invention, the graying method adopts weighted graying, and specific operations are technical means well known to those skilled in the art, which are not described herein, and other graying methods may be used to perform graying processing in other embodiments, which are not limited herein.
Referring to fig. 2, a gray scale image diagram of an ultrasound image of children's pneumonia according to an embodiment of the invention is shown. In fig. 2, the upper half is the structure of the rib-and chest-wall-related tissue, i.e., the extraneous background area; the lower half is a more blurred lung symptom area, i.e. the area to be enhanced. As can be seen from the information components in fig. 2, for an irrelevant background composed of tissues such as bones, a more regular approximation exists in a certain direction, such as rib regions, in fig. 2, transverse distribution regions, and obvious region approximation exists in the transverse direction, on the basis of the principle, a sliding window for carrying out similarity analysis along a fixed direction can be constructed in a gray level image of a children pneumonic ultrasonic image, and regions with regular similarity can be segmented by traversing the whole image.
Considering that a background area such as bones and the like in a gray level image has obvious edge texture characteristics, in order to enable the built sliding window area to only contain information of one tissue, an initial sliding window can be built in the gray level image according to a preset initial size, then the size to be adjusted of the initial sliding window is determined according to gradient distribution of pixel points in the horizontal direction and gradient distribution in the vertical direction in the initial sliding window, and then the self-adaptive sliding window is obtained. The self-adaptive sliding window is selected by considering gradient distribution in a region, and the size of the self-adaptive sliding window is controlled according to the gradient distribution, so that no large gradient change exists in the sliding window region, namely no obvious edge texture exists, and the pixel points in the self-adaptive sliding window region are ensured to be the same tissue region as much as possible. In one embodiment of the present invention, the initial size is set to 12×12.
Preferably, in one embodiment of the present invention, determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the horizontal direction and the gradient distribution in the vertical direction in the initial sliding window includes:
the dimension to be adjusted includes a length to be adjusted and a width to be adjusted. Starting from the upper left corner of the initial sliding window. In the row where the starting point is located, the gradient of each pixel point in the horizontal direction is obtained, the point with the largest gradient is selected as a first cut-off point, and the first cut-off point is the point with the largest edge probability in the horizontal direction, so that the width to be adjusted can be obtained according to the coordinate distance between the starting point and the first cut-off point. If the starting point and the first cut-off point are the same point, the width to be adjusted is 1. And in the column where the starting point is located, obtaining the gradient of each pixel point in the vertical direction, and selecting the point with the largest gradient as a second cut-off point, wherein the second cut-off point is the point with the largest edge probability in the vertical direction. And obtaining the length to be adjusted according to the coordinate distance between the starting point and the second cut-off point. If the starting point and the second cut-off point are the same point, the length to be adjusted is 1. And starting from the starting point, constructing the self-adaptive sliding window through the length to be adjusted and the width to be adjusted.
It should be noted that, the gradient acquiring method is a technical means well known to those skilled in the art, and in the embodiment of the present invention, the sobel operator is selected to acquire gradient information of a corresponding pixel, and a specific calculation method is not described herein.
After the self-adaptive sliding window is obtained, region similarity analysis can be started, a comparison sliding window is constructed on the preset sliding direction according to the self-adaptive sliding window, and the comparison sliding window and the self-adaptive sliding window are identical in size and are in common edge. The self-adaptive sliding window slides along the preset sliding direction, and the sliding step length is the dimension in the preset direction. In the embodiment of the invention, because the areas such as ribs and the like in the default ultrasonic image are transversely distributed, the sliding direction is set to be transverse, namely the adaptive sliding window and the contrast sliding window have a broadside sharing edge.
Step S2: obtaining the similarity between the corresponding areas of the two sliding windows according to the gray value difference of the same position between the self-adaptive sliding window and the contrast sliding window; according to the similarity, a similarity label is given to the areas corresponding to the self-adaptive sliding window and the contrast sliding window in the gray level image; and continuously constructing an initial sliding window for the area without the similarity label and obtaining the similarity label until the similarity label exists at each position on the gray level image, so as to obtain a label image.
And (3) after the sliding window is set in the step (S1), performing similarity analysis on the obtained areas corresponding to the two sliding windows, namely obtaining the similarity between the areas corresponding to the two sliding windows according to the gray value difference of the same position between the self-adaptive sliding window and the contrast sliding window. That is, the larger the difference in gray value, the more the two regions are not the same tissue region, and the smaller the similarity between the two regions.
Preferably, considering that the gray values of the pixel points in the same tissue area in the gray image are not completely the same, the change is usually performed in an error interval, in order to reduce the sensitivity of the similarity calculation process, in the embodiment of the present invention, the gray values in the gray image are first divided into a preset number of gray levels, and the gray level differences at the same position between the adaptive sliding window and the contrast sliding window are accumulated to obtain an overall difference value. And carrying out negative correlation mapping and normalization on the overall difference value to obtain the corresponding similarity. In the embodiment of the invention, the number of gray levels is set to 16, and the 16 gray levels are used for replacing pixel values of various positions in the gray image, so that corresponding gray level differences are obtained.
In one embodiment of the present invention, the formula for obtaining the similarity after performing negative correlation mapping and normalization on the overall difference value is expressed as:
wherein, the liquid crystal display device comprises a liquid crystal display device,is the firstThe similarity in the secondary similarity calculation process,is the firstThe overall difference value in the secondary similarity calculation process,the maximum similarity in the process is calculated for all this similarity traversing the entire gray image. In the formula, the maximum value is taken as a denominator, the integral difference value is normalized to be in a value range of 0 to 1, and then the normalized value is subtracted by a constant 1 to realize negative correlation mapping.
Other methods of negative correlation mapping, such as negative power mapping of an exponential function, and normalization, may also be used in other embodiments of the present invention, which are not limited and described herein.
And after the similarity is obtained, a similarity label can be given to the areas corresponding to the self-adaptive sliding window and the contrast sliding window in the gray level image. The method specifically comprises the following steps: if the similarity is in the preset similarity interval, assigning similar labels to all pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window; if the similarity is not in the preset similarity interval, giving dissimilar labels to all pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window. In the embodiment of the invention, the similar interval is set as [0.9,1]; for ease of calculation, the tag value of the similar tag is set to 1 and the tag value of the dissimilar tag is set to 0.
And each time the similarity label is given, the similarity judgment is completed once for the corresponding region, so that the initial sliding window is continuously constructed for the region without the similarity label, the similarity label is obtained until the similarity label exists at each position in the gray level image, the label image is obtained, and the pixel value at each pixel point position in the label image is the corresponding similarity label.
Preferably, in order to ensure the integrity of the label assignment process, a fixed rule needs to be set to limit each assignment of a label value, specifically including: starting to construct an initial sliding window by using the upper left corner of the gray level image and obtaining a similarity label; removing the area with the similarity label in the gray image to obtain a to-be-processed area, continuously constructing an initial sliding window from the upper left corner in the to-be-processed area and obtaining the similarity label until the similarity label exists at each position on the gray image, and obtaining a label image; if the self-adaptive sliding window or the contrast sliding window has a region which does not contain pixel points in the giving process of the similarity label, supplementing 0 to the corresponding region.
Step S3: traversing the whole label image along the direction vertical to the sliding direction by utilizing a regularity operator of a preset shape in the label image; obtaining regional regularity according to the distribution of similarity label categories in a region corresponding to the regularity operator, and adjusting the similarity label of the region center corresponding to the regularity operator according to the regional regularity; and obtaining a rule tag image, and screening out a rule area and an area to be enhanced according to the similarity tag in the rule tag image.
Because the similarity labels in the label image are based on the area analysis in a preset direction, human tissues cannot grow and develop along a fixed direction in the ultrasonic image of the children's pneumonia, so that certain errors exist in the similarity labels in the label image, and in order to eliminate the errors, the whole label image is traversed along a direction perpendicular to the sliding direction by using a regularity operator of the preset shape in the label image. And in the region corresponding to the regularity operator, obtaining the region regularity according to the distribution of the similarity label categories, wherein the larger the region regularity is, the more regular the region corresponding to the regularity operator is, and the more the region accords with the rule organization region. Therefore, the similarity label of the region center corresponding to the regularity operator can be adjusted according to the region regularity to obtain a regular label image. In the embodiment of the invention, the regularity operator is set to be square with the size of 3*3, and traversing is performed by taking each pixel point in the label image as a center.
Preferably, obtaining the region regularity from the distribution of similarity tag categories includes:
and in the region corresponding to the statistical regularity operator, the number of the similar labels is used as the regional regularity according to the number of the similar labels. That is, in the region corresponding to the regularity operator, the more similar labels, the more regular the region, and the greater the region regularity.
Preferably, in one embodiment of the present invention, adjusting the similarity label of the region center corresponding to the regularity operator according to the region regularity includes:
if the regional regularity is smaller than a preset regularity threshold, describing that the region is irregular, and setting a similarity label of the region center corresponding to the regularity operator as a dissimilarity label; if the regional regularity is equal to a preset regularity threshold, the original similarity label is reasonably set, and the similarity label of the regional center corresponding to the regularity operator is not changed; if the regional regularity is greater than a preset regularity threshold, the corresponding region is regular, and a similarity label of the region center corresponding to the regularity operator is set as a similarity label. In the embodiment of the invention, the regularity threshold is set to 0.5.
After the similarity labels are adjusted, the label information in the rule label image has strong referential property, and the rule area and the area to be enhanced can be screened directly according to the similarity labels in the rule label image.
Specifically, an area corresponding to a similar label in the regular label image is used as a regular area, and an area corresponding to a dissimilar label is used as an area to be enhanced. I.e. the regular areas are background areas where no enhancement is needed, while the remaining areas are random sign areas to be enhanced.
Step S4: and reinforcing the corresponding region to be reinforced in the gray level image to obtain the reinforced ultrasonic image.
By the method, the region to be enhanced can be identified in a targeted manner after the region to be enhanced is extracted, influence of irrelevant background information is avoided, the image quality after enhancement is improved, and the calculated amount of the image enhancement process is reduced. Referring to fig. 3, a schematic view of an enhanced ultrasound image according to an embodiment of the invention is shown.
Preferably, obtaining the enhanced ultrasound image comprises:
in the gray level image, removing the area corresponding to the regular area to obtain an image to be enhanced, wherein the image only comprises the area to be enhanced; performing histogram equalization processing on an image to be enhanced to obtain an enhanced image; and splicing the enhanced image with the corresponding area of the regular area in the gray level image to obtain the enhanced ultrasonic image.
It should be noted that, the steps of histogram equalization are generally: counting the gray level histogram of the image, obtaining the gray level frequency probability density, calculating the cumulative distribution probability of the original image, and mapping to obtain an equalization result. The specific process is a technical means well known to those skilled in the art, and will not be described herein.
The enhanced ultrasonic image after the enhancement processing contains information with obvious reference significance, and information such as the number of B lines, the position of the B lines, the area of the B lines, the starting position of abnormal signs from the B lines, the length of pleura lines, the position information, whether intervals exist or not, the number of A lines, the position of the A lines, the interval distance between the A lines and the like can be marked in the image, and the marked information is used for medical teaching.
In summary, in the embodiment of the invention, an initial sliding window is set in an ultrasound image of children pneumonia, an adaptive sliding window is constructed according to gradient distribution in the initial sliding window, a contrast sliding window corresponding to the adaptive sliding window is constructed, similarity between two sliding window areas is analyzed, and a similarity label is given to the corresponding area according to the similarity. And setting a regularity operator in a direction perpendicular to the sliding direction of the sliding window, and adjusting the similarity label according to the regularity in the area of the operator to obtain a regular label image. And further obtaining a region to be enhanced, and carrying out targeted enhancement on the region to be enhanced to obtain an enhanced ultrasonic image. According to the embodiment of the invention, the ultrasound image is subjected to regularity analysis, irrelevant background information is removed, and the quality of the enhanced ultrasound image is higher.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (6)
1. An intelligent enhancement method for medical images of children's pneumonia, which is characterized by comprising the following steps:
obtaining a gray image of an ultrasonic image of the children's pneumonia; setting an initial sliding window in the gray level image according to a preset initial size; determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the initial sliding window in the horizontal direction and the gradient distribution of the pixel points in the vertical direction, adjusting the initial sliding window according to the size to be adjusted, obtaining an adaptive sliding window, and constructing a contrast sliding window in a preset sliding direction according to the adaptive sliding window, wherein the contrast sliding window has the same size and is co-limited with the adaptive sliding window;
obtaining the similarity between the corresponding areas of the two sliding windows according to the gray value difference of the same position between the self-adaptive sliding window and the contrast sliding window; according to the similarity, a similarity label is given to the areas corresponding to the self-adaptive sliding window and the contrast sliding window in the gray level image; continuously constructing the initial sliding window for the area without the similarity label and obtaining the similarity label until the similarity label exists at each position on the gray level image, so as to obtain a label image;
traversing the whole label image along the direction perpendicular to the sliding direction by utilizing a regularity operator of a preset shape in the label image; obtaining regional regularity according to the distribution of similarity label categories in a region corresponding to the regularity operator, and adjusting the similarity label of the region center corresponding to the regularity operator according to the regional regularity; obtaining a rule tag image, and screening out a rule area and an area to be enhanced according to the similarity tag in the rule tag image; the regularity operator is a square window with the size of 3 multiplied by 3;
reinforcing the region to be reinforced corresponding to the gray level image to obtain an enhanced ultrasonic image;
and according to the similarity, giving a similarity label to the areas corresponding to the self-adaptive sliding window and the contrast sliding window in the gray level image:
the similarity labels comprise similar labels and dissimilar labels;
if the similarity is in a preset similarity interval, giving the similarity labels to all pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window;
if the similarity is not in the preset similarity interval, giving the dissimilar labels to all the pixel points in the corresponding areas of the self-adaptive sliding window and the contrast sliding window;
the obtaining the regional regularity according to the distribution of the similarity label categories comprises:
counting the number proportion of the similar labels in the area corresponding to the regularity operator, and taking the number proportion as the area regularity;
the obtaining the similarity between the two sliding window corresponding areas comprises the following steps:
dividing the gray level value in the gray level image into a preset number of gray levels, and accumulating gray level differences at the same position between the self-adaptive sliding window and the contrast sliding window to obtain an overall difference value; and mapping and normalizing the negative correlation of the integral difference value to obtain the similarity.
2. The method for intelligently enhancing medical images of children's pneumonia according to claim 1, wherein determining the size to be adjusted of the initial sliding window according to the gradient distribution of the pixel points in the horizontal direction and the gradient distribution in the vertical direction in the initial sliding window comprises:
the dimension to be adjusted comprises a length to be adjusted and a width to be adjusted;
taking the left upper corner of the initial sliding window as a starting point;
in the row where the starting point is located, obtaining the gradient of each pixel point in the horizontal direction, selecting the point with the largest gradient as a first cut-off point, and obtaining the width to be adjusted according to the coordinate distance between the starting point and the first cut-off point; if the starting point and the first cut-off point are the same point, the width to be adjusted is 1;
in the column where the starting point is located, obtaining the gradient of each pixel point in the vertical direction, selecting the point with the largest gradient as a second cut-off point, and obtaining the length to be adjusted according to the coordinate distance between the starting point and the second cut-off point; and if the starting point and the second cut-off point are the same point, the length to be adjusted is 1.
3. The method for intelligently enhancing medical images of children's pneumonia according to claim 1, wherein the method for acquiring the tag image comprises the following steps:
starting to construct the initial sliding window at the upper left corner of the gray level image and obtaining the similarity label; removing the area with the similarity label in the gray level image every time the giving process of the similarity label is executed to obtain a to-be-processed area, and continuing to construct the initial sliding window from the upper left corner in the to-be-processed area and obtaining the similarity label until the similarity label exists at each position on the gray level image to obtain the label image; if an area which does not contain pixel points exists in the self-adaptive sliding window or the contrast sliding window in the giving process of the similarity label, supplementing 0 to the corresponding area.
4. The intelligent enhancement method for children's pneumonia medical images according to claim 1, wherein the adjusting the similarity label of the region center corresponding to the regularity operator according to the region regularity comprises:
if the regional regularity is smaller than a preset regularity threshold, setting a similarity label of the regional center corresponding to the regularity operator as the dissimilar label;
if the regional regularity is equal to a preset regularity threshold, not changing a similarity label of the regional center corresponding to the regularity operator;
and if the regional regularity is greater than a preset regularity threshold, setting a similarity label of the regional center corresponding to the regularity operator as the similarity label.
5. The method for intelligently enhancing a medical image of childhood pneumonia according to claim 4, wherein said screening out a rule area and an area to be enhanced according to said similarity tag in said rule tag image comprises:
and taking the region corresponding to the similar label in the regular label image as a regular region, and taking the region corresponding to the dissimilar label as the region to be enhanced.
6. The method for intelligently enhancing a medical image of pediatric pneumonia according to claim 1, wherein said obtaining an enhanced ultrasound image comprises:
in the gray level image, removing the area corresponding to the regular area to obtain an image to be enhanced, wherein the image to be enhanced only comprises the area to be enhanced; performing histogram equalization processing on the image to be enhanced to obtain an enhanced image; and splicing the enhanced image with the area corresponding to the regular area in the gray level image to obtain the enhanced ultrasonic image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310538317.XA CN116309187B (en) | 2023-05-15 | 2023-05-15 | Intelligent enhancement method for medical image of children pneumonia |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310538317.XA CN116309187B (en) | 2023-05-15 | 2023-05-15 | Intelligent enhancement method for medical image of children pneumonia |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116309187A CN116309187A (en) | 2023-06-23 |
| CN116309187B true CN116309187B (en) | 2023-07-28 |
Family
ID=86829095
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310538317.XA Active CN116309187B (en) | 2023-05-15 | 2023-05-15 | Intelligent enhancement method for medical image of children pneumonia |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116309187B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116681701B (en) * | 2023-08-02 | 2023-11-03 | 青岛市妇女儿童医院（青岛市妇幼保健院、青岛市残疾儿童医疗康复中心、青岛市新生儿疾病筛查中心） | Children lung ultrasonic image processing method |
| CN117095444B (en) * | 2023-08-29 | 2024-02-13 | 广州市星飞达电子科技有限公司 | Image recognition method based on artificial intelligence |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114419321A (en) * | 2022-03-30 | 2022-04-29 | 珠海市人民医院 | CT image heart segmentation method and system based on artificial intelligence |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1261910C (en) * | 2003-12-08 | 2006-06-28 | 西安理工大学 | Automatic generating method for colour multi-window CT image |
| US20190154822A1 (en) * | 2017-11-21 | 2019-05-23 | The Charles Stark Draper Laboratory, Inc. | System and method for imaging and localization of contrast-enhanced features in the presence of accumulating contrast agent in a body |
| US11638569B2 (en) * | 2018-06-08 | 2023-05-02 | Rutgers, The State University Of New Jersey | Computer vision systems and methods for real-time needle detection, enhancement and localization in ultrasound |
| CN110211058A (en) * | 2019-05-15 | 2019-09-06 | 南京极目大数据技术有限公司 | A kind of data enhancement methods of medical image |
| CN111339828B (en) * | 2020-01-19 | 2021-04-30 | 华中科技大学同济医学院附属协和医院 | Vein developing identification method based on combination of infrared image and ultrasonic Doppler |
| CN113706435A (en) * | 2020-05-09 | 2021-11-26 | 北京康兴顺达科贸有限公司 | Chest enhanced CT image processing method based on traditional image omics |
| CN114648530B (en) * | 2022-05-20 | 2022-08-16 | 潍坊医学院 | CT image processing method |
| CN115100048A (en) * | 2022-06-06 | 2022-09-23 | 青岛海信医疗设备股份有限公司 | Image processing method, device and apparatus |
| CN115797259A (en) * | 2022-11-01 | 2023-03-14 | 郑州大学第一附属医院 | Multi-scale detail enhancement method and system for low-dose lung CT image |
-
2023
- 2023-05-15 CN CN202310538317.XA patent/CN116309187B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114419321A (en) * | 2022-03-30 | 2022-04-29 | 珠海市人民医院 | CT image heart segmentation method and system based on artificial intelligence |
Non-Patent Citations (1)
| Title |
|---|
| 结构纹理分离的对比度和细节增强;尹超 等;《计算机应用研究》(第12期);第318-321页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116309187A (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116309187B (en) | Intelligent enhancement method for medical image of children pneumonia | |
| US11908580B2 (en) | Image classification method, computer-readable storage medium, and computer device | |
| CN115861135B (en) | Image enhancement and recognition method applied to panoramic detection of box body | |
| CN102354397B (en) | Method for reconstructing human facial image super-resolution based on similarity of facial characteristic organs | |
| US9330336B2 (en) | Systems, methods, and media for on-line boosting of a classifier | |
| US8170303B2 (en) | Automatic cardiac view classification of echocardiography | |
| US9152926B2 (en) | Systems, methods, and media for updating a classifier | |
| CN103530882B (en) | Improved image segmentation method based on picture and color texture features | |
| CN110309853B (en) | Medical image clustering method based on variational self-encoder | |
| CN109685804B (en) | Multi-channel head magnetic resonance imaging tissue segmentation method | |
| CN113763442B (en) | Deformable medical image registration method and system | |
| US20190347468A1 (en) | 3d image detection method and apparatus, electronic device, and computer readable medium | |
| CN111462068A (en) | Bolt and nut detection method based on transfer learning | |
| CN112232371A (en) | American license plate recognition method based on YOLOv3 and text recognition | |
| WO2023116632A1 (en) | Video instance segmentation method and apparatus based on spatio-temporal memory information | |
| CN112308822A (en) | Intervertebral disc CT image detection method based on deep convolutional neural network | |
| CN113456031A (en) | Training device and prediction device of brain state prediction model and electronic equipment | |
| CN111402244B (en) | Automatic classification method for standard fetal heart tangent planes | |
| CN112258532B (en) | Positioning and segmentation method for callus in ultrasonic image | |
| CN110111307B (en) | Immune system feedback simulation system and method for immune teaching | |
| CN113469942B (en) | CT image lesion detection method | |
| CN115311491A (en) | Fracture false positive screening method based on course learning and spatial attention | |
| CN113034451A (en) | Chest DR image identification method based on deep learning | |
| CN111507969B (en) | Image fusion quality detection method and device | |
| CN113627231B (en) | Automatic segmentation method for liquid region in retina OCT image based on machine vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |