CN117011195A - Human infrared imaging data processing system for assisting traditional Chinese medicine - Google Patents
Human infrared imaging data processing system for assisting traditional Chinese medicine Download PDFInfo
- Publication number
- CN117011195A CN117011195A CN202311280901.6A CN202311280901A CN117011195A CN 117011195 A CN117011195 A CN 117011195A CN 202311280901 A CN202311280901 A CN 202311280901A CN 117011195 A CN117011195 A CN 117011195A
- Authority
- CN
- China
- Prior art keywords
- target
- infrared image
- determining
- target area
- candidate
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- 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
- G06T2207/10048—Infrared 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
- G06T2207/30196—Human being; Person
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radiology & Medical Imaging (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to a human infrared imaging data processing system for assisting traditional Chinese medicine, which comprises the following components: the following steps can be realized by the mutual cooperation among a plurality of modules: acquiring a target human body infrared image and a standard human body infrared image; determining texture noise intensity corresponding to the infrared image of the target human body; dividing the infrared image of the target human body; determining target noise intensity and target filtering variance corresponding to the target human infrared image; screening the optimal filter window size corresponding to each target area from a preset filter window size set; and performing adaptive Gaussian filtering denoising on the infrared image of the target human body according to the target filtering variance and the optimal filtering window size corresponding to each target region. According to the invention, through data processing of the target human body infrared image and the standard human body infrared image, denoising of the target human body infrared image is realized, and denoising effect of the human body infrared image is improved.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a human infrared imaging data processing system for assisting traditional Chinese medicine.
Background
In an infrared imaging image of a human body, due to the influence of an image shooting environment, noise is often randomly generated in the imaging process, the influence of the noise on the image is often obvious, for example, the result of subsequent processing such as image segmentation, feature extraction, edge detection and the like is often influenced to a certain extent, so that the diagnosis of symptoms by traditional Chinese medicine is influenced. Therefore, the collected infrared images of the human body often need to be subjected to denoising treatment. At present, when denoising an image, the following methods are generally adopted: and adopting a window with a preset size and a preset filtering variance to perform Gaussian filtering denoising on the image. Where the filter variance is the variance to which the gaussian filter kernel obeys. The preset filter variance may be a artificially set filter variance.
However, when the gaussian filter denoising is performed on the human infrared image by using a window with a preset size and a preset filter variance, the following technical problems often exist: because the window and the filtering variance involved in Gaussian filtering denoising are often set based on artificial subjective experience, the set result is often inaccurate, and the denoising effect on the human infrared image is often poor.
Disclosure of Invention
The summary of the invention is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. The summary of the invention is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In order to solve the technical problem of poor denoising effect on human infrared images, the invention provides a human infrared imaging data processing system for assisting traditional Chinese medicine.
The invention provides a human infrared imaging data processing system for assisting traditional Chinese medicine, which comprises:
the image acquisition module is used for acquiring a target human body infrared image and a standard human body infrared image;
the texture noise intensity determining module is used for determining the texture noise intensity corresponding to the target human infrared image according to the target human infrared image and the standard human infrared image;
the image segmentation module is used for segmenting the infrared image of the target human body to obtain a target region set;
the noise intensity determining module is used for determining target noise intensity corresponding to the target human body infrared image according to the target region set;
The filtering variance determining module is used for determining a target filtering variance corresponding to the target human infrared image according to the texture noise intensity and the target noise intensity;
the screening module is used for screening the optimal filter window size corresponding to each target area from a preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set;
and the filtering denoising module is used for carrying out adaptive Gaussian filtering denoising on the target human body infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set to obtain a target denoising image.
Optionally, the determining the texture noise intensity corresponding to the target human infrared image according to the target human infrared image and the standard human infrared image includes:
determining the variance of gray values corresponding to all pixel points in the standard human infrared image as standard gray difference;
determining the variances of the gray values corresponding to all pixel points in the target human infrared image as integral gray differences;
and determining the duty ratio of the integral gray level difference in the standard gray level difference as the texture noise intensity corresponding to the infrared image of the target human body.
Optionally, the determining, according to the target area set, the target noise intensity corresponding to the target human infrared image includes:
randomly screening a pixel point from each target area in the target area set to serve as a candidate pixel point corresponding to the target area;
determining a gray value corresponding to a candidate pixel point corresponding to each target area as a candidate gray index corresponding to the target area;
determining the average value of gray values corresponding to all pixel points in each target area as a target representative gray index corresponding to the target area;
determining the difference value between the candidate gray index corresponding to each target area and the target representative gray index as a gray deviation index corresponding to the target area;
determining the variances of gray level deviation indexes corresponding to all target areas in the target area set as target variances;
and determining the ratio of the target variance to a preset variance as the target noise intensity corresponding to the target human infrared image.
Optionally, the determining, according to the texture noise intensity and the target noise intensity, a target filtering variance corresponding to the target human infrared image includes:
And determining the difference value of the texture noise intensity and the target noise intensity as a target filtering variance corresponding to the target human infrared image.
Optionally, the screening the optimal filter window size corresponding to each target area from the preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set includes:
determining a window with a size being a preset filter window size as a candidate window to obtain a candidate window set;
for each candidate window in the target area and the candidate window set, determining a first screening index between the target area and the candidate window according to gradients corresponding to all pixel points in the candidate window corresponding to the pixel points in the target area;
for the target area and each candidate window, determining a second screening index between the target area and the candidate window according to gray values corresponding to all pixel points in the candidate window corresponding to the pixel points in the target area;
for the target area and each candidate window, determining a third screening index between the target area and the candidate window according to a first screening index and a second screening index between the target area and the candidate window, wherein the first screening index and the second screening index are in negative correlation with the third screening index;
Screening a candidate window with the maximum third screening index between the candidate window set and the target area as an optimal candidate window corresponding to the target area;
and determining the preset filter window size corresponding to the optimal candidate window corresponding to the target area as the optimal filter window size corresponding to the target area.
Optionally, the determining, according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point in the target area, a first screening index between the target area and the candidate window includes:
for each pixel point in the target area and the candidate window, determining the gradient complexity degree between the pixel point and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point;
and determining the variance of the gradient complexity degree between all pixel points in the target area and the candidate window as a first screening index between the target area and the candidate window.
Optionally, the determining the gradient complexity between the pixel point and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point includes:
Determining the duty ratio of each gradient in the gradients corresponding to all the pixel points in the candidate window corresponding to the pixel point as the target frequency corresponding to each gradient;
determining the complex entropy of the gradient corresponding to the candidate window corresponding to the pixel point according to the target frequencies corresponding to various gradients in the gradients corresponding to all the pixel points in the candidate window corresponding to the pixel point;
and determining the gradient complex entropy corresponding to the candidate window corresponding to the pixel point as the gradient complexity degree between the pixel point and the candidate window.
Optionally, the determining, according to the gray value corresponding to each pixel point in the candidate window corresponding to the pixel point in the target area, a second screening index between the target area and the candidate window includes:
determining the variance of gray values corresponding to all pixel points in the candidate window corresponding to each pixel point in the target area as the candidate gray difference corresponding to the candidate window corresponding to each pixel point in the target area;
and determining a second screening index between the target area and the candidate window according to the candidate gray level difference corresponding to the candidate window corresponding to each pixel point in the target area, wherein the candidate gray level difference and the second screening index are positively correlated.
Optionally, the segmenting the target human body infrared image to obtain a target region set includes:
and performing super-pixel segmentation on the target human body infrared image, and taking each region obtained after super-pixel segmentation as a target region to obtain a target region set.
Optionally, the performing adaptive gaussian filtering denoising on the target human body infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set includes:
for each target region in the target region set, determining a window with the size being the optimal filter window size corresponding to the target region as the optimal filter window corresponding to the target region;
and carrying out Gaussian filtering denoising on each target region according to the target filtering variance and the optimal filtering window corresponding to each target region in the target region set, so as to realize adaptive Gaussian filtering denoising on the target human infrared image.
The invention has the following beneficial effects:
according to the human body infrared imaging data processing system for assisting traditional Chinese medicine, through data processing of the target human body infrared image and the standard human body infrared image, denoising of the target human body infrared image is achieved, the technical problem that denoising effect of the human body infrared image is poor is solved, and denoising effect of the human body infrared image is improved. Firstly, a target human body infrared image and a standard human body infrared image are acquired, so that the target human body infrared image can be conveniently subjected to denoising treatment. Then, the accuracy of texture noise intensity determination can be improved by comprehensively considering the target human body infrared image and the standard human body infrared image. Then, the infrared image of the target human body is segmented, so that accurate denoising can be conveniently carried out on each target region in the target region set. Continuing, comprehensively considering the target area set, the accuracy of determining the target noise intensity corresponding to the target human body infrared image can be improved. And then, comprehensively considering the texture noise intensity and the target noise intensity, and improving the accuracy of determining the target filtering variance corresponding to the target human infrared image. And then, comprehensively considering the gradient and the gray value corresponding to each pixel point in each target area, the accuracy of the optimal filtering window size screening corresponding to each target area can be improved. Finally, based on the target filtering variance and the optimal filtering window size corresponding to each target area in the target area set, the adaptive Gaussian filtering denoising is carried out on the target human infrared image, so that the denoising of the target human infrared image is realized, and when the optimal filtering window size corresponding to each target area in the target area set is determined, the texture noise intensity, the target noise intensity, the gradient and gray value corresponding to each pixel point in the target area and the like are comprehensively considered, so that the determination of the target filtering variance and the optimal filtering window size corresponding to each target area is objective, the influence of artificial subjective factors is reduced to a certain extent, and the denoising effect of the target human infrared image 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 structural diagram of a human infrared imaging data processing system for assisting traditional Chinese medicine.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. 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 infrared image of human body has certain diagnostic significance in medicine, especially in the aspect of traditional Chinese medicine, the diagnosis of the temperature distribution of different parts of human body in the infrared image of human body can better evaluate and judge the health problem of human body, but the infrared image is a relatively low-frequency image, because the infrared image possibly has a large amount of noise, the temperature distribution is also gradually changed, the infrared image is subjected to Gaussian filtering operation through a window with a preset size and a preset filtering variance, and part of detail information can be smoothed to a certain extent, so that the denoising effect of the infrared image is improved, the invention provides a human body infrared imaging data processing system for assisting the traditional Chinese medicine, which comprises:
the image acquisition module is used for acquiring a target human body infrared image and a standard human body infrared image;
the texture noise intensity determining module is used for determining the texture noise intensity corresponding to the target human infrared image according to the target human infrared image and the standard human infrared image;
the image segmentation module is used for segmenting the infrared image of the target human body to obtain a target region set;
the noise intensity determining module is used for determining target noise intensity corresponding to the target human body infrared image according to the target region set;
The filtering variance determining module is used for determining a target filtering variance corresponding to the infrared image of the target human body according to the texture noise intensity and the target noise intensity;
the screening module is used for screening the optimal filter window size corresponding to each target area from the preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set;
and the filtering denoising module is used for carrying out adaptive Gaussian filtering denoising on the target human body infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set to obtain a target denoising image.
Referring to fig. 1, there is shown a schematic structural diagram of a human infrared imaging data processing system for assisting in Chinese medicine according to the present invention. The human infrared imaging data processing system for assisting traditional Chinese medicine comprises:
the image acquisition module 101 is used for acquiring a target human body infrared image and a standard human body infrared image.
In some embodiments, a target human infrared image and a standard human infrared image may be acquired.
The target human body infrared image may be an infrared image of a photographed human body to be subjected to denoising processing. The standard human infrared image may be an infrared image of a photographed human body without noise.
The target human body infrared image and the standard human body infrared image are acquired, so that the target human body infrared image can be conveniently subjected to subsequent denoising treatment.
As an example, this step may include the steps of:
first, acquiring an infrared image of a target human body.
For example, an infrared image corresponding to a human body may be acquired as a target human body infrared image by an infrared camera.
And secondly, acquiring a standard human body infrared image.
For example, a historical human infrared image without noise may be screened out of the set of historical human infrared images as a standard human infrared image. The historical human body infrared image in the historical human body infrared image set can be an infrared image of a photographed human body.
It should be noted that, because the noise is randomly generated, when the number of the historical human infrared images in the historical human infrared image set is greater, the historical human infrared images without noise are more likely to appear in the historical human infrared image set, and the historical human infrared images without noise are more likely to be screened from the historical human infrared image set.
It should be noted that, with the development of infrared thermal imaging technology in the medical field, the studies of the coverage of the infrared thermal imaging technology, the accuracy of disease diagnosis and efficacy evaluation, the judgment of traditional Chinese medicine symptoms, the evaluation of the efficacy of traditional Chinese medicine treatment, and the like are increasing. The infrared thermal imaging technology has important auxiliary functions in disease diagnosis and efficacy evaluation, and has diagnostic value of traditional inspection methods for certain special diseases. The human body is a natural heating body, infrared rays are continuously radiated to the surrounding through the body surface, the body surface temperature is influenced by various factors, such as human anatomy, tissue metabolism, subcutaneous blood circulation, skin heat conductivity, nerve state and the like, and external factors such as temperature, humidity and the like. The human body has certain regulation capacity, and when the influence of the factors exceeds the regulation capacity, the human body can generate pathological changes to generate certain symptoms. Has important auxiliary effect on the medical diagnosis of traditional Chinese medicine. Because different human organs present different radiation intensities, which often lead to different color distribution in the infrared image, and noise influence often exists in the infrared image, the noise can cause partial information in the image to become fuzzy, so in order to better assist traditional Chinese medicine diagnosis, the human infrared image is often required to be denoised, and the method is mainly used for denoising the human infrared image.
The texture noise intensity determining module 102 is configured to determine the texture noise intensity corresponding to the target human infrared image according to the target human infrared image and the standard human infrared image.
In some embodiments, the texture noise intensity corresponding to the target human infrared image may be determined according to the target human infrared image and the standard human infrared image.
It should be noted that, comprehensively considering the target human infrared image and the standard human infrared image, the accuracy of determining the texture noise intensity can be improved.
As an example, this step may include the steps of:
and determining the variance of gray values corresponding to all pixel points in the standard human infrared image as the standard gray difference.
And secondly, determining the variance of gray values corresponding to all pixel points in the target human body infrared image as the integral gray difference.
And thirdly, determining the duty ratio of the integral gray level difference in the standard gray level difference as the texture noise intensity corresponding to the infrared image of the target human body.
For example, the formula for determining the texture noise intensity corresponding to the infrared image of the target human body may be:
wherein, Is the texture noise intensity corresponding to the infrared image of the target human body. />The integral gray level difference is the variance of gray level values corresponding to all pixel points in the infrared image of the target human body. />The standard gray level difference is the variance of gray level values corresponding to all pixel points in the standard human infrared image.
It should be noted that, in order to prevent the denominator from being 0, for each ratio of the present invention, a preset factor greater than 0 may be added to the denominator included in the ratio, for example, the preset factor may take a value of 0.01.
It should be noted that the number of the substrates,the intensity of the blur noise in the infrared image of the target human body can be characterized. The signal-to-noise ratio is introduced to perform approximate calculation on noise in the infrared image of the target human body, and the concept of the signal-to-noise ratio is the ratio of signal power to noise power in the image, and the power of the noise and the signal is calculated by using gray values in the image. Thus can use +.>To evaluate the intensity of the blur noise in the infrared image of the target human body.
The image segmentation module 103 is configured to segment the infrared image of the target human body to obtain a target region set.
In some embodiments, the target human infrared image may be segmented to obtain a target region set.
It should be noted that, the target human body infrared image is segmented, so that each target region in the target region set can be accurately denoised later.
As an example, the above-mentioned target human body infrared image may be subjected to super-pixel segmentation, and each region obtained after the super-pixel segmentation is taken as a target region, so as to obtain a target region set.
The noise intensity determining module 104 is configured to determine, according to the target region set, a target noise intensity corresponding to the target human infrared image.
In some embodiments, the target noise intensity corresponding to the target human infrared image may be determined according to the target region set.
It should be noted that, by comprehensively considering the target region set, accuracy of determining the target noise intensity corresponding to the target human infrared image can be improved.
As an example, this step may include the steps of:
and a first step of randomly screening a pixel point from each target area in the target area set to be used as a candidate pixel point corresponding to the target area.
For example, the first pixel point in the target area may be used as the candidate pixel point corresponding to the target area.
And secondly, determining the gray value corresponding to the candidate pixel point corresponding to each target area as the candidate gray index corresponding to the target area.
For example, a gray value corresponding to a candidate pixel point corresponding to a target area may be used as a candidate gray index corresponding to the target area.
And thirdly, determining the average value of gray values corresponding to all pixel points in each target area as a target representative gray index corresponding to the target area.
For example, the average value of the gray values corresponding to all the pixel points in the target area may be determined as the target representative gray index corresponding to the target area.
Fourth, determining the difference between the candidate gray scale index corresponding to each target area and the target representative gray scale index as the gray scale deviation index corresponding to the target area.
For example, a difference between the candidate gray index corresponding to the target region and the target representative gray index may be used as the gray scale deviation index corresponding to the target region.
And fifthly, determining the variances of the gray level deviation indexes corresponding to all the target areas in the target area set as target variances.
And sixthly, determining the ratio of the target variance to the preset variance as the target noise intensity corresponding to the target human infrared image.
For example, the formula for determining the target noise intensity corresponding to the target human infrared image may be:
wherein,is the target noise intensity corresponding to the target human infrared image. />Is the firstiWithin the target areajGray values corresponding to the pixel points. />Is the first in the target region setiTargets corresponding to the target areas represent gray indexes; i.e. the firstiAnd the average value of gray values corresponding to all pixel points in each target area. Due to the firstiThe candidate pixel points corresponding to the target areas can be from the firstiOne pixel point randomly screened in each target area, so the firstiWithin the target areajThe pixel point can be the firstiCandidate pixels corresponding to the target area, thus +.>Can be the firstiCandidate gray indexes corresponding to the target areas. />Is the first in the target region setiGray scale deviation indexes corresponding to the target areas.bIs a preset variance. For example,bmay be 1.NIs the number of target regions in the set of target regions. />Is the firstiThe number of pixels in each target area.fIs the average value of gray level deviation indexes corresponding to all target areas in the target area set. />Is the target variance.iIs the sequence number of the target region in the set of target regions. jIs the firstiAnd the serial numbers of the pixel points in the target areas.
It should be noted that the number of the substrates,the noise intensity after removing the texture features in the infrared image of the target human body can be characterized. Assuming that the obtained target area is an area except for the texture features, the target area at this time tends to be uniform, that is, if the texture features in the infrared image of the target human body are removed, the same obtained target area tends to be composed of the same color, that is, the background of the bottom surface of the target area at this time tends to be uniform. The denominator isbMeaning that the region segmented by the super-pixel is assumed to be a region with uniform background, and the gray variance of the background region is close to 0, the gray variance of the background region can be assumed to be 1 for comparison with the texture noise intensity, namelybMay be 1. Thus->The noise intensity after removing the texture features in the infrared image of the target human body can be characterized.
The filtering variance determining module 105 is configured to determine a target filtering variance corresponding to the target human infrared image according to the texture noise intensity and the target noise intensity.
In some embodiments, a target filtering variance corresponding to the target human infrared image may be determined according to the texture noise intensity and the target noise intensity.
It should be noted that, by comprehensively considering the texture noise intensity and the target noise intensity, the accuracy of determining the target filtering variance corresponding to the target human infrared image can be improved.
As an example, a difference between the texture noise intensity and the target noise intensity may be determined as a target filter variance corresponding to the target human infrared image.
For example, the formula for determining the target filter variance corresponding to the target human infrared image may be:
wherein,wis the target filtering variance corresponding to the target human infrared image.Is the texture noise intensity corresponding to the infrared image of the target human body. />Is the target noise intensity corresponding to the target human infrared image.
It should be noted that the number of the substrates,the intensity of the blur noise in the infrared image of the target human body can be characterized, which often includes the complexity of the image texture. />The noise intensity after removing the texture features in the infrared image of the target human body can be characterized. Thus (2)wThe surface texture complexity of the target human infrared image may be characterized. Second, due to->And->All can be regarded as noise variance ratio standard image variance, so that the differences can be put in the same dimension for phase difference, if the denominator is unified, the difference result can be interpreted according to the denominator, and the difference result represents that the background texture variance ratio is equal to the ideal uniform background variance, namely the denominator is 1, and then wWhat is actually expressed is also a variance value, i.e. texture complexity. Since the filtering strength of the Gaussian filter kernel tends to be subject to a certain variance, unlike the mean filtering in that the Gaussian filtering increases the weight of the center pixel, a certain amount of original image information can be retained, while the texture complexity is required, i.ewCan be regarded as noiseThe expectation of the detail information of the filtered image is that the Gaussian filter kernel filtering strength is set by using the texture complexity, so that the Gaussian filter is compliant with the texture strength of the original image when noise is filtered, and the original image information is kept as much as possible.
And the screening module 106 is configured to screen an optimal filter window size corresponding to each target area from a preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set.
In some embodiments, the optimal filter window size corresponding to each target area may be selected from the preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set.
The preset filter window size in the preset filter window size set may be a preset size of a window for performing gaussian filter denoising. The preset filter window size may be an odd number. For example, if the number of preset filter window sizes in the preset filter window size set is 3, the preset filter window sizes in the preset filter window size set may be respectively: 3×3, 5×5, and 7×7. If the number of preset filter window sizes in the preset filter window size set is 4, the preset filter window sizes in the preset filter window size set may be respectively: 3×3, 5×5, 7×7, and 9×9. The gradient corresponding to the pixel point can be detected by a sobel operator.
It should be noted that, the richer the preset filter window size in the preset filter window size set, the more suitable the optimal candidate window corresponding to the target area to be screened is for the target area, so that the better the subsequent denoising effect is. Secondly, the gradient and the gray value corresponding to each pixel point in each target area are comprehensively considered, and the accuracy of the optimal filtering window size screening corresponding to each target area can be improved.
As an example, this step may include the steps of:
the first step, determining a window with the size being the size of a preset filtering window as a candidate window to obtain a candidate window set.
For example, if the preset filter window sizes in the preset filter window size set are respectively: 3 x 3, 5 x 5, and 7 x 7), the candidate window set may include: 3 x 3 windows, 5 x 5 windows, and 7 x 7 windows. The 3×3 window is a window of size 3×3. The 5×5 window is a window of size 5×5. The 7×7 window is a window of size 7×7.
The second step of determining, for each candidate window in the set of candidate windows and the target region, a first screening indicator between the target region and the candidate window according to gradients corresponding to respective pixels in the candidate window corresponding to pixels in the target region may include the sub-steps of:
The first substep, for each pixel point in the target area and the candidate window, of determining the gradient complexity between the pixel point and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point may include the following steps:
first, each gradient of gradients corresponding to all pixels in the candidate window corresponding to the pixel is determined as a target frequency corresponding to each gradient.
The gradients with the same gradient size and gradient direction can be the same gradient.
If a candidate window is a 3×3 window, the candidate window corresponding to the pixel is the 3×3 window corresponding to the pixel, and the pixel is located at the center of the 3×3 window corresponding to the pixel.
And then, determining the complex entropy of the gradient corresponding to the candidate window corresponding to the pixel point according to the target frequencies corresponding to various gradients in the gradients corresponding to all the pixel points in the candidate window corresponding to the pixel point.
And finally, determining the gradient complex entropy corresponding to the candidate window corresponding to the pixel point as the gradient complexity degree between the pixel point and the candidate window.
For example, the formula for determining the gradient complexity between the pixel point and the candidate window may be:
wherein,is the first in the target region setiWithin the target areajThe first pixel point and the candidate window settGradient complexity between candidate windows, i.e. the firstiWithin the target areajThe first pixel point corresponds totGradient complex entropy corresponding to each candidate window. For example, if the firsttThe candidate window is a window with the preset filter window size of 3×3, then the firstjThe first pixel point corresponds totThe candidate window may be the firstjAnd 3×3 windows corresponding to the pixels. />Is the firstiWithin the target areajThe first pixel point corresponds totThe total number of different gradients in the gradients corresponding to all pixel points in the candidate windows; i.e. the firstiWithin the target areajThe first pixel point corresponds totThe number of gradient seeds in the gradients corresponding to all pixel points in each candidate window. />Is the firstiWithin the target areajThe first pixel point corresponds totWithin the candidate windows, the corresponding gradient is the firstpThe number of pixels of the gradient. />Is the firstiWithin the target areajThe first pixel point corresponds totThe number of pixels within each candidate window. />Is based on natural constant->Logarithmic (log). / >Is a logarithmic function.iIs the sequence number of the target region in the set of target regions.jIs the firstiAnd the serial numbers of the pixel points in the target areas.tIs the sequence number of the candidate window in the candidate window set. The candidate windows are in one-to-one correspondence with the preset filter window sizes, and the number of the candidate windows is equal to the number of the preset filter window sizes.pIs the firstiWithin the target areajThe first pixel point corresponds totThe number of gradients in the candidate window, i.e. the firstiWithin the target areajThe first pixel point corresponds totSequence numbers of different gradients within each candidate window.
When the following is performedThe larger the tends to explain the firstiWithin the target areajThe first pixel point corresponds totThe more chaotic the gradient corresponding to each pixel point in each candidate window, the more often the description of the firstiWithin the target areajThe first pixel point corresponds totThe more complex the gradient type distribution corresponding to each pixel point in each candidate window.
And a second substep, determining the variance of the gradient complexity degree between all the pixel points in the target area and the candidate window as a first screening index between the target area and the candidate window.
For example, the firstiTarget area and the firsttThe first screening index between candidate windows may be: first, the iAll pixel points and the first pixel point in each target areatVariance of gradient complexity between candidate windows.
Third, for the target area and each candidate window, determining a second screening indicator between the target area and the candidate window according to the gray values corresponding to the pixels in the candidate window corresponding to the pixels in the target area may include the following substeps:
and a first sub-step of determining the variance of the gray level values corresponding to all the pixel points in the candidate window corresponding to each pixel point in the target area as the candidate gray level difference corresponding to the candidate window corresponding to each pixel point in the target area.
And a second substep, determining a second screening index between the target area and the candidate window according to the candidate gray level difference corresponding to the candidate window corresponding to each pixel point in the target area.
The candidate gray scale difference may be positively correlated with the second screening indicator.
For example, the formula for determining the second screening index correspondence between the target area and the candidate window may be:
wherein,is the first in the target region setiThe target area and the first candidate window set tA second screening indicator between candidate windows. />Is the firstiThe number of pixels in each target area. />Is the firstiWithin the target areajThe first pixel point corresponds totCandidate gray scale differences corresponding to the candidate windows; i.e. the firstiWithin the target areajThe first pixel point corresponds totThe variance of the gray values corresponding to all pixels in each candidate window. />And->And shows positive correlation.iIs the sequence number of the target region in the set of target regions.jIs the firstiAnd the serial numbers of the pixel points in the target areas.tIs the sequence number of the candidate window in the candidate window set.
When the following is performedThe larger the tends to explain the firstiWithin the target areajThe first pixel point corresponds totThe more chaotic the gray scale distribution of each candidate window, the more often the description of the firstiWithin the target areajThe first pixel point corresponds totThe larger the texture information within the candidate window, the more often the description of the firstiWithin the target areajThe first pixel point corresponds totThe more likely each candidate window will contain pixels of different colors. When->The larger the tends to explain the firstiThe first pixel point corresponding to all the pixel points in the target areatThe larger the difference of gray values of the pixel points in the candidate windows, namely the firstiThe first pixel point in the target area corresponds to tThe relatively more texture information in each candidate window. When->The smaller the time, the more description of the firstiThe first pixel point in the target area corresponds totThe less texture information is relatively within a candidate window, i.e. the firstiThe first of the target areastThe temperature distribution in the candidate windows is more uniform, andiall of the firsttThe temperature distribution within each candidate window is relatively similar.
Fourth, for the target area and each candidate window, determining a third screening index between the target area and the candidate window according to the first screening index and the second screening index between the target area and the candidate window.
Wherein, the first screening index and the second screening index may both be negatively correlated with the third screening index.
For example, the formula for determining the third screening index correspondence between the target region and the candidate window may be:
wherein,is the first in the target region setiThe target area and the first candidate window settAnd a third screening indicator between the candidate windows. />Is the first in the target region setiThe target area and the first candidate window settFirst screening criteria between candidate windows, i.e. the firstiAll pixel points and the first pixel point in each target area tVariance of gradient complexity between candidate windows. />Is the first in the target region setiThe target area and the first candidate window settA second screening indicator between candidate windows. />Is a natural constant +.>To the power. />Is an exponential function with a base of natural constant. />And->All are in charge of>And has negative correlation.
It should be noted that the number of the substrates,can represent the firstiThe first pixel point corresponding to all the pixel points in the target areatCandidate windowsDegree of dispersion of intra-oral pixel gradient types. When->The smaller the time, the more description of the firstiEvery pixel point in the target area corresponds to the first pixel pointtThe more uniform the gradient types of each pixel point in each candidate window, the more often the description of the firstiEvery pixel point in the target area corresponds to the first pixel pointtThe closer the complexity of the gradient types of the individual pixels within the candidate windows. When->The larger the tends to explain the firstiThe first pixel point corresponding to all the pixel points in the target areatThe larger the difference of gray values of the pixel points in the candidate windows, namely the firstiThe first pixel point in the target area corresponds totThe relatively more texture information in each candidate window. When->The larger the size, the more often the description is at the firstiThe first pixel point corresponding to all the pixel points in the target area tIn the candidate windows, the first pixel point corresponds totThe more uniform the gradient type distribution of the pixel points in the candidate window is, i.e. the firstiEvery pixel point in the target area corresponds to the first pixel pointtThe gradient types in the candidate windows are relatively few; and (1)iEvery pixel point in the target area corresponds to the first pixel pointtThe smaller gray variance of the pixel points in the candidate window indicates that the window has less texture information, namely the firstiThe more uniform the distribution of the individual target areas within the window; the more representative that the temperature area distribution in the window has less texture information, the more can ensure the first after Gaussian filtering processingiThe more detailed information will not be smoothed out by the target areas, the better the window size at this time.
And fifthly, screening out a candidate window with the maximum third screening index between the candidate window set and the target area as an optimal candidate window corresponding to the target area.
For example, one may select from a set of candidate windowsScreening and the firstiA candidate window with the largest third screening index among the target areas is taken as a first candidate windowiAnd the optimal candidate windows corresponding to the target areas.
And sixthly, determining the preset filter window size corresponding to the optimal candidate window corresponding to the target area as the optimal filter window size corresponding to the target area.
The size of the preset filtering window corresponding to the optimal candidate window is the size of the optimal candidate window.
The filtering denoising module 107 is configured to perform adaptive gaussian filtering denoising on the target human infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set, so as to obtain a target denoised image.
In some embodiments, adaptive gaussian filtering denoising can be performed on the target human infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set, so as to obtain a target denoising image.
The target denoising image may be an image obtained by performing adaptive Gaussian filtering denoising on the target human body infrared image.
The adaptive Gaussian filter denoising is performed on the target human infrared image based on the target filter variance and the optimal filter window size corresponding to each target region in the target region set, so that denoising of the target human infrared image is realized.
As an example, this step may include the steps of:
the first step, for each target region in the target region set, determining a window with a size equal to the optimal filter window size corresponding to the target region as the optimal filter window corresponding to the target region.
For example, the firstiThe optimal filter window corresponding to each target region may be: size is the firstiWindow of optimal filter window size corresponding to each target region.
And secondly, carrying out Gaussian filtering denoising on each target region according to the target filtering variance and the optimal filtering window corresponding to each target region in the target region set, so as to realize adaptive Gaussian filtering denoising on the infrared image of the target human body.
For example, for each target area, the target filtering variance may be taken as the variance obeyed by the gaussian filter kernel, and the optimal filtering window corresponding to the target area is taken as the window in the case of gaussian filtering denoising, so that gaussian filtering denoising is performed on the target area.
It should be noted that, the filtering kernel of the gaussian filtering often depends on the weighted average of surrounding pixel points to replace the value of the center point, but unlike the mean filtering, the weight value of the pixel points around the center point is generally higher as the pixel points are closer to the center point; the farther away from the center point, the lower the weight, the parameters of the Gaussian filter kernel are mainly the size of the window and the variance obeyed by the window, and the target filter variance corresponding to the infrared image of the target human body can be obtained wAs the variance obeyed by the Gaussian filter kernel, the Gaussian filter is carried out on each target area through the optimal filter window corresponding to each target area, so that the texture intensity of the original image is reserved as much as possible, and the subsequent image processing can be more accurate.
The Gaussian filtering belongs to linear filtering as the mean filtering, noise pixel points in an image can be filtered, however, in the filtering process, excessive detail information can be smoothed in the image smoothing process due to the weight in the Gaussian filtering kernel, the infrared image is relatively low in frequency, the excessive detail information is smoothed, the gradual change degree of temperature is more fuzzy, the doctor is not benefited to diagnose the symptoms through the infrared image, the smoothing of the detail information is reduced in the Gaussian filtering process by optimizing the size and the dimension of the filtering kernel, the detail information of a temperature gradual change part is reserved to a great extent, the quality of the image is improved to a certain extent, the follow-up processing operation of the image is facilitated, and the diagnosis of traditional Chinese medicine can be better assisted.
In conclusion, texture noise intensity, target noise intensity, gradient and gray value corresponding to each pixel point in a target area and the like are comprehensively considered, so that the target filtering variance is objectively determined with the optimal filtering window size corresponding to each target area, the influence of artificial subjective factors is reduced to a certain extent, the denoising effect of the target human infrared image is improved, and the diagnosis of symptoms of traditional Chinese medicine can be better assisted. The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention and are intended to be included within the scope of the invention.
Claims (10)
1. A human infrared imaging data processing system for assisting in traditional Chinese medicine, the system comprising:
the image acquisition module is used for acquiring a target human body infrared image and a standard human body infrared image;
the texture noise intensity determining module is used for determining the texture noise intensity corresponding to the target human infrared image according to the target human infrared image and the standard human infrared image;
the image segmentation module is used for segmenting the infrared image of the target human body to obtain a target region set;
the noise intensity determining module is used for determining target noise intensity corresponding to the target human body infrared image according to the target region set;
the filtering variance determining module is used for determining a target filtering variance corresponding to the target human infrared image according to the texture noise intensity and the target noise intensity;
the screening module is used for screening the optimal filter window size corresponding to each target area from a preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target area in the target area set;
and the filtering denoising module is used for carrying out adaptive Gaussian filtering denoising on the target human body infrared image according to the target filtering variance and the optimal filtering window size corresponding to each target region in the target region set to obtain a target denoising image.
2. The system for assisting traditional Chinese medical science in processing infrared imaging data of human body according to claim 1, wherein the determining the texture noise intensity corresponding to the target infrared image according to the target infrared image and the standard infrared image comprises:
determining the variance of gray values corresponding to all pixel points in the standard human infrared image as standard gray difference;
determining the variances of the gray values corresponding to all pixel points in the target human infrared image as integral gray differences;
and determining the duty ratio of the integral gray level difference in the standard gray level difference as the texture noise intensity corresponding to the infrared image of the target human body.
3. The system for assisting traditional Chinese medical science in processing infrared imaging data of human body according to claim 1, wherein the determining the target noise intensity corresponding to the target infrared image of human body according to the target region set comprises:
randomly screening a pixel point from each target area in the target area set to serve as a candidate pixel point corresponding to the target area;
determining a gray value corresponding to a candidate pixel point corresponding to each target area as a candidate gray index corresponding to the target area;
Determining the average value of gray values corresponding to all pixel points in each target area as a target representative gray index corresponding to the target area;
determining the difference value between the candidate gray index corresponding to each target area and the target representative gray index as a gray deviation index corresponding to the target area;
determining the variances of gray level deviation indexes corresponding to all target areas in the target area set as target variances;
and determining the ratio of the target variance to a preset variance as the target noise intensity corresponding to the target human infrared image.
4. The system for assisting traditional Chinese medical science in processing infrared imaging data of human body according to claim 1, wherein the determining the target filtering variance corresponding to the target infrared image of human body according to the texture noise intensity and the target noise intensity comprises:
and determining the difference value of the texture noise intensity and the target noise intensity as a target filtering variance corresponding to the target human infrared image.
5. The infrared imaging data processing system for assisting traditional Chinese medicine according to claim 1, wherein the step of screening the optimal filter window size corresponding to each target region from the preset filter window size set according to the gradient and the gray value corresponding to each pixel point in each target region in the target region set comprises the steps of:
Determining a window with a size being a preset filter window size as a candidate window to obtain a candidate window set;
for each candidate window in the target area and the candidate window set, determining a first screening index between the target area and the candidate window according to gradients corresponding to all pixel points in the candidate window corresponding to the pixel points in the target area;
for the target area and each candidate window, determining a second screening index between the target area and the candidate window according to gray values corresponding to all pixel points in the candidate window corresponding to the pixel points in the target area;
for the target area and each candidate window, determining a third screening index between the target area and the candidate window according to a first screening index and a second screening index between the target area and the candidate window, wherein the first screening index and the second screening index are in negative correlation with the third screening index;
screening a candidate window with the maximum third screening index between the candidate window set and the target area as an optimal candidate window corresponding to the target area;
And determining the preset filter window size corresponding to the optimal candidate window corresponding to the target area as the optimal filter window size corresponding to the target area.
6. The system for assisting traditional Chinese medical science in processing infrared imaging data of human body according to claim 5, wherein the determining the first screening index between the target region and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point in the target region comprises:
for each pixel point in the target area and the candidate window, determining the gradient complexity degree between the pixel point and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point;
and determining the variance of the gradient complexity degree between all pixel points in the target area and the candidate window as a first screening index between the target area and the candidate window.
7. The infrared imaging data processing system for human body assisted traditional Chinese medicine according to claim 6, wherein determining the complexity of the gradient between the pixel point and the candidate window according to the gradient corresponding to each pixel point in the candidate window corresponding to the pixel point comprises:
Determining the duty ratio of each gradient in the gradients corresponding to all the pixel points in the candidate window corresponding to the pixel point as the target frequency corresponding to each gradient;
determining the complex entropy of the gradient corresponding to the candidate window corresponding to the pixel point according to the target frequencies corresponding to various gradients in the gradients corresponding to all the pixel points in the candidate window corresponding to the pixel point;
and determining the gradient complex entropy corresponding to the candidate window corresponding to the pixel point as the gradient complexity degree between the pixel point and the candidate window.
8. The infrared imaging data processing system for human body assisted traditional Chinese medicine according to claim 5, wherein determining a second screening index between the target region and the candidate window according to the gray values corresponding to the pixels in the candidate window corresponding to the pixels in the target region comprises:
determining the variance of gray values corresponding to all pixel points in the candidate window corresponding to each pixel point in the target area as the candidate gray difference corresponding to the candidate window corresponding to each pixel point in the target area;
And determining a second screening index between the target area and the candidate window according to the candidate gray level difference corresponding to the candidate window corresponding to each pixel point in the target area, wherein the candidate gray level difference and the second screening index are positively correlated.
9. The system for assisting traditional Chinese medical science in processing infrared imaging data of human body according to claim 1, wherein the step of dividing the target infrared image of human body to obtain a target region set comprises the steps of:
and performing super-pixel segmentation on the target human body infrared image, and taking each region obtained after super-pixel segmentation as a target region to obtain a target region set.
10. The system for processing infrared imaging data of human body in assistance of traditional Chinese medicine according to claim 1, wherein the adaptive gaussian filtering denoising of the infrared image of the human body according to the target filtering variance and the optimal filtering window size corresponding to each target region in the set of target regions comprises:
for each target region in the target region set, determining a window with the size being the optimal filter window size corresponding to the target region as the optimal filter window corresponding to the target region;
And carrying out Gaussian filtering denoising on each target region according to the target filtering variance and the optimal filtering window corresponding to each target region in the target region set, so as to realize adaptive Gaussian filtering denoising on the target human infrared image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311280901.6A CN117011195B (en) | 2023-10-07 | 2023-10-07 | Human infrared imaging data processing system for assisting traditional Chinese medicine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311280901.6A CN117011195B (en) | 2023-10-07 | 2023-10-07 | Human infrared imaging data processing system for assisting traditional Chinese medicine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117011195A true CN117011195A (en) | 2023-11-07 |
| CN117011195B CN117011195B (en) | 2024-01-23 |
Family
ID=88567599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311280901.6A Active CN117011195B (en) | 2023-10-07 | 2023-10-07 | Human infrared imaging data processing system for assisting traditional Chinese medicine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117011195B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117455800A (en) * | 2023-12-22 | 2024-01-26 | 肥城新查庄地质勘查有限公司 | Coal mine drilling positioning method based on computer vision |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111223068A (en) * | 2019-11-12 | 2020-06-02 | 西安建筑科技大学 | An adaptive non-uniform low-light image enhancement method based on Retinex |
| WO2021217643A1 (en) * | 2020-04-30 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Method and device for infrared image processing, and movable platform |
| CN116205823A (en) * | 2023-05-05 | 2023-06-02 | 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） | A Method of Ultrasound Image Denoising Based on Spatial Filtering |
| CN116758441A (en) * | 2023-06-16 | 2023-09-15 | 河南通航无人机科技有限公司 | Unmanned aerial vehicle cluster intelligent scheduling management system |
-
2023
- 2023-10-07 CN CN202311280901.6A patent/CN117011195B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111223068A (en) * | 2019-11-12 | 2020-06-02 | 西安建筑科技大学 | An adaptive non-uniform low-light image enhancement method based on Retinex |
| WO2021217643A1 (en) * | 2020-04-30 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Method and device for infrared image processing, and movable platform |
| CN116205823A (en) * | 2023-05-05 | 2023-06-02 | 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） | A Method of Ultrasound Image Denoising Based on Spatial Filtering |
| CN116758441A (en) * | 2023-06-16 | 2023-09-15 | 河南通航无人机科技有限公司 | Unmanned aerial vehicle cluster intelligent scheduling management system |
Non-Patent Citations (1)
| Title |
|---|
| LANG WU ETC.: ""Infrared small target detection based on gray intensity descent and local gradient watershed"", 《INFRARED PHYSICS AND TECHNOLOGY》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117455800A (en) * | 2023-12-22 | 2024-01-26 | 肥城新查庄地质勘查有限公司 | Coal mine drilling positioning method based on computer vision |
| CN117455800B (en) * | 2023-12-22 | 2024-03-15 | 肥城新查庄地质勘查有限公司 | Coal mine drilling positioning method based on computer vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117011195B (en) | 2024-01-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107451998B (en) | Fundus image quality control method | |
| CN108272434B (en) | Method and device for processing fundus images | |
| Badsha et al. | A new blood vessel extraction technique using edge enhancement and object classification | |
| EP3070673B1 (en) | Image processing method to diagnose cutaneous lesion, and diagnostic apparatus used for the same method | |
| CN106558031B (en) | A kind of image enchancing method of the colored eyeground figure based on imaging model | |
| Karargyris et al. | Segmenting anatomy in chest x-rays for tuberculosis screening | |
| CN118229538B (en) | Intelligent enhancement method for bone quality CT image | |
| CN109377473B (en) | Method and device for positioning target pulse taking point | |
| CN117011195B (en) | Human infrared imaging data processing system for assisting traditional Chinese medicine | |
| CN109993731A (en) | A kind of eyeground pathological changes analysis method and device | |
| CN119295493A (en) | Tumor medical image processing method and system for tumor ablation treatment system | |
| CN110675325B (en) | Method and device for enhancing fundus image | |
| Jeevakala et al. | A novel segmentation of cochlear nerve using region growing algorithm | |
| Ramasubramanian et al. | A Comprehensive Analysis of Various Delineation method for Exudates in Fundus Images using Miniaturized Pi Board | |
| WO2011047342A1 (en) | Systems and methods for detecting retinal abnormalities | |
| Wu et al. | Hybrid enhancement algorithm for nailfold images with large fields of view | |
| Naidu et al. | Enhancement of X-ray images using various Image Processing Approaches | |
| Idowu et al. | Improved enhancement technique for medical image processing | |
| Vuppu et al. | Early glaucoma detection using LSTM-CNN integrated with multi class SVM | |
| Siddiqi | Filter selection for removing noise from ct scan images using digital image processing algorithm | |
| Jimenez-Martin et al. | Specular reflections removal in colposcopic images based on neural networks: Supervised training with no ground truth previous knowledge | |
| Reddy et al. | Spatial Mutual Relationship Based Retinal Image Contrast Enhancement for Efficient Diagnosis of Diabetic Retinopathy. | |
| Gautam et al. | Implementation of NLM and PNLM for de-noising of MRI images | |
| Lin et al. | Stitching method for panoramic nail fold images based on capillary contour enhancement | |
| Lynn et al. | Segmentation of skin lesion towards melanoma skin cancer classification |
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 |