KR101812406B1 - The method and system for diagnosing skin disease - Google Patents
The method and system for diagnosing skin disease Download PDFInfo
- Publication number
- KR101812406B1 KR101812406B1 KR1020160031509A KR20160031509A KR101812406B1 KR 101812406 B1 KR101812406 B1 KR 101812406B1 KR 1020160031509 A KR1020160031509 A KR 1020160031509A KR 20160031509 A KR20160031509 A KR 20160031509A KR 101812406 B1 KR101812406 B1 KR 101812406B1
- Authority
- KR
- South Korea
- Prior art keywords
- skin disease
- lesion
- image
- skin
- unit
- Prior art date
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/44—Detecting, measuring or recording for evaluating the integumentary system, e.g. skin, hair or nails
- A61B5/441—Skin evaluation, e.g. for skin disorder diagnosis
- A61B5/444—Evaluating skin marks, e.g. mole, nevi, tumour, scar
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/44—Detecting, measuring or recording for evaluating the integumentary system, e.g. skin, hair or nails
- A61B5/441—Skin evaluation, e.g. for skin disorder diagnosis
- A61B5/445—Evaluating skin irritation or skin trauma, e.g. rash, eczema, wound, bed sore
Abstract
The present invention relates to a technique for diagnosing a skin disease, and more particularly, to a method and system for diagnosing a skin disease by analyzing characteristics of a skin disease lesion by detecting a skin disease lesion based on an image of the skin disease will be.
A method for diagnosing a skin disease according to an embodiment of the present invention includes: obtaining a skin disease image; Processing the skin disease image to remove noise from the obtained skin disease image and preserve the outline information of the skin disease lesion; Detecting a skin disease lesion from the treated skin disease image; Analyzing the characteristics of the skin lesion based on the detected skin lesion; And classifying and diagnosing the skin disease based on the characteristics of the analyzed skin disease lesion.
Description
The present invention relates to a technique for diagnosing a skin disease, and more particularly, to a method and system for diagnosing a skin disease by analyzing characteristics of a skin disease lesion by detecting a skin disease lesion based on an image of the skin disease will be.
Recently, interest in skin health care has increased, and devices and medical devices for skin diagnosis or skin care have been developed.
With regard to techniques for diagnosing skin diseases currently being developed, there are provided imaging devices for simply enlarging the skin, devices for diagnosing the extent of skin pigmentation, and devices for providing aging information. In addition to the diagnostic technology, skin-related devices with the ability to manage skin and scalp conditions have also been developed and launched in the skin-related market.
Furthermore, a device having a function of classifying skin diseases by analyzing the image of the skin has been developed. However, the classification accuracy is lowered, the application range is very limited, and it is not applicable to everyday skin diseases.
Therefore, there is no diagnostic system or diagnostic method for skin disease that can be applied to both malignant skin disease and normal skin disease, and there is a limit to technology for diagnosing skin disease. In addition, the imaging apparatus for photographing existing skin diseases was limited to the use of dermoscopy.
SUMMARY OF THE INVENTION The present invention has been made to solve the above problems and it is an object of the present invention to provide a method and apparatus for acquiring a skin-related image through a general camera, not a dermoscopy, And to provide a system and method for diagnosing skin diseases.
As an embodiment of the present invention, a method for diagnosing a skin disease can be provided.
A method for diagnosing a skin disease according to an embodiment of the present invention includes: obtaining a skin disease image; Processing the skin disease image to remove noise from the obtained skin disease image and preserve the outline information of the skin disease lesion; Detecting a skin disease lesion from the treated skin disease image; Analyzing the characteristics of the skin lesion based on the detected skin lesion; And classifying and diagnosing the skin disease based on the characteristics of the analyzed skin disease lesion.
According to an embodiment of the present invention, a mask image can be acquired through a Convex Hull analysis on a skin disease image in the step of processing a skin disease image.
According to an embodiment of the present invention, in the step of detecting a skin disease lesion from an image, an optimal outline area of a skin disease lesion can be detected by applying a level set model based on the obtained mask image.
According to an embodiment of the present invention, in analyzing features of a skin disease lesion, the user response information on the skin disease related to the skin disease can be received and the texture feature of the skin disease lesion can be extracted and evaluated.
The texture features of skin lesions extracted and evaluated according to an embodiment of the present invention include FOS (First Order Statistics), GLCM (Gray Level Co-occurrence Matrix, LF, Intensity, Matrix), Local Binary Pattern (LBP), Discrete Wavelet Texture (DWT), or Fractal Dimension (FD).
According to an embodiment of the present invention, the step of classifying and diagnosing a skin disease includes AdaBoost (Adaptive Boosting), RF (Random Forest), SVM (Support) The present invention includes a step of classifying a skin disease by a plurality of classifiers that can be implemented by a vector machine and a probabilistic neural network (PNN), and diagnosing a skin disease by weighting a classification result by a plurality of classifiers .
Meanwhile, as an embodiment of the present invention, a computer-readable recording medium on which a program for causing the computer to execute the above-described method may be provided.
According to an embodiment of the present invention, a skin disease diagnosis system includes an image acquisition unit that acquires a skin disease image; An image processing unit for processing skin disease images to remove noise from acquired skin disease images and to preserve outline information of skin disease lesions; A lesion detection unit for detecting a lesion of the skin disease from the skin disease image processed by the image processing unit; A lesion analyzing unit for analyzing a characteristic of a skin disease lesion based on the lesion of the lesion detected by the lesion detecting unit; And a skin disease diagnosis unit for classifying and diagnosing a skin disease based on the characteristics of the skin disease lesion analyzed through the lesion analysis unit.
The image processing unit according to an embodiment of the present invention can acquire a mask image through Convex Hull analysis on a skin disease image.
The lesion detection unit according to an embodiment of the present invention can detect an optimized outline area of a skin disease lesion by applying a level set model based on a mask image obtained in an image processing unit.
The lesion analyzing unit according to an exemplary embodiment of the present invention may include a document information receiving unit for receiving user response information on a paper related to a skin disease and a texture feature extracting unit for extracting and evaluating a texture feature for a skin disease lesion.
The skin disease diagnosis unit according to an embodiment of the present invention includes AdaBoost (Adaptive Boosting), RF (Random), and the like, using the user response information for the paperwork received at the paper information reception unit and the texture characteristics extracted and evaluated by the texture feature extraction unit, A skin disease classifier for classifying a skin disease by a plurality of classifiers that can be implemented with a support vector machine (SVM), a support vector machine (SVM), and a probabilistic neural network (PNN) And a weighting unit for diagnosing the disease.
The present invention analyzes the characteristics of a lesion appearing in an image of a skin disease and diagnoses a skin disease using a plurality of classifiers to improve the reliability of the diagnosis result and can be applied to not only malignant diseases but also daily diseases To diagnose various skin diseases.
In addition, the present invention can provide a medical service related to diagnosis of a skin disease to patients who can not receive medical care by visiting a hospital.
1 is a flowchart of a method for diagnosing a skin disease according to an embodiment of the present invention.
FIG. 2 illustrates a sequence of acquiring mask images in the image processing unit or the image processing unit according to an embodiment of the present invention.
FIG. 3 illustrates an optimized outline area of a skin lesion detected in a lesion detection step or a lesion detection part according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating an item of the medical history of daily skin diseases among the medical articles based on the received user response information according to an embodiment of the present invention.
FIG. 5 is a view showing an item of the inquiry about the malignant skin disease among the inquiry items on which the received user response information is based, according to an embodiment of the present invention.
FIG. 6 shows textural features of a skin disease lesion image according to an embodiment of the present invention.
FIG. 7 shows a classifier for classifying daily skin diseases according to an embodiment of the present invention and a method for diagnosing skin diseases classified by each classifier.
8 shows a classifier classifying malignant skin diseases according to an embodiment of the present invention and a method of diagnosing skin diseases classified by each classifier.
FIG. 9 shows the condition of a daily skin disease classifier implemented with AdaBoost (Adaptive Boosting) according to an embodiment of the present invention.
FIG. 10 illustrates the conditions of a routine skin disease classifier implemented in a Random Forest (RF) according to an embodiment of the present invention.
11 shows the condition of a normal skin disease classifier implemented with Support Vector Machine (SVM) according to an embodiment of the present invention.
FIG. 12 shows the condition of a normal skin disease classifier implemented with a probabilistic neural network (PNN) according to an embodiment of the present invention.
FIG. 13 shows conditions for each classifier algorithm implemented in the malignant skin disease classifier according to an embodiment of the present invention.
FIG. 14 is a block diagram showing a configuration of a skin disease diagnosis system according to an embodiment of the present invention.
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.
The terms used in this specification will be briefly described and the present invention will be described in detail.
While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments. Also, in certain cases, there may be a term selected arbitrarily by the applicant, in which case the meaning thereof will be described in detail in the description of the corresponding invention. Therefore, the term used in the present invention should be defined based on the meaning of the term, not on the name of a simple term, but on the entire contents of the present invention.
When an element is referred to as "including" an element throughout the specification, it is to be understood that the element may include other elements as well, without departing from the spirit or scope of the present invention. Also, the terms "part," " module, "and the like described in the specification mean units for processing at least one function or operation, which may be implemented in hardware or software or a combination of hardware and software . In addition, when a part is referred to as being "connected" to another part throughout the specification, it includes not only "directly connected" but also "connected with other part in between".
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.
1 is a flowchart of a method for diagnosing a skin disease according to an embodiment of the present invention.
The method of diagnosing a skin disease according to an embodiment of the present invention includes obtaining a skin disease image (S110); Processing the skin disease image to remove noise from the acquired skin disease image and preserve outline information of the skin disease lesion (S120); Detecting a skin disease lesion from the treated skin disease image (S130); Analyzing a feature of the lesion of the skin disease based on the detected lesion of the skin disease (S150); And classifying and diagnosing the skin disease based on the characteristics of the analyzed skin disease lesion (S160).
FIG. 2 illustrates a sequence of acquiring mask images in the image processing unit or the image processing unit according to an embodiment of the present invention.
As shown in FIG. 2 (a), the skin disease image can be acquired from a camera and can be acquired through a camera included in a general camera or a mobile device including dermoscopy (S110).
In step S120 of processing the skin disease image, Convex Hull analysis is performed on the skin disease image by removing noise such as hair or air bubbles, and passing through a preprocessing step (see FIG. 2) for preserving information on the outline of the lesion A mask image can be obtained.
2 (a) is converted into a Gray Scale image (FIG. 2 (b)), the image data is adjusted to a gray scale intensity of 20 to 230 in order to maximize the outline information of the lesion and remove noise, Band pass filter. ≪ / RTI > The filtered image data may be linearly mapped in the range of 0 to 255 to obtain an image as shown in FIG. 2 (c).
The image shown in FIG. 2 (c) can be obtained as an image as shown in FIG. 2 (d) through an anisotropic diffusion filter to remove noise in the image and to preserve the outline information.
2 (d), the gray scale of the image is determined by the histogram analysis, and the average gray scale intensity is determined based on the histogram. When the binarization is performed by dividing the average gray scale intensity by the average intensity, the image of FIG. 2 (e) can be obtained.
Since the image of FIG. 2 (e) may contain fine noise in the lesion area, fine noise included in the lesion area can be removed through an erosion operation (FIG. 2 (f)).
Since the outline of the lesion area may be uneven, as shown in Fig. 2 (f), in which the fine noise included in the lesion area is removed, it is possible to use the Convex Hull, a calculation technique for creating an optimal polygon based on the outline points of the lesion area The lesion area can be optimized. The mask image (Fig. 2 (g)) can finally be acquired through Convex Hull.
According to an embodiment of the present invention, in step S130 of detecting a skin disease lesion from an image, a level set model based on the obtained mask image may be applied to detect an outline area optimized for a skin disease lesion (S130 ).
Since the lesion area shown in the mask image (FIG. 2 (g)) does not exactly coincide with the actual lesion area, it is possible to detect the optimal outline area of the skin lesion by applying the Level Set Model based on the mask image (S130) .
In the step S130 of detecting the optimized outline region of the skin disease lesion, the outline optimization can be performed by designating the outline region of the mask image (Fig. 2 (g)) as the input value of the level set model. In order to minimize the error due to the fine movement, it is possible to detect the improved outline region for the skin disease lesion by applying the level set model again after removing the weight of the inside and outside of the outline region.
After the improved outline area is detected, Convex Hull calculation is performed again to obtain the outline area of the final skin disease lesion.
3 (a) is a mask image, FIG. 3 (b) is a gray scale image, and FIG. 3 (c) is a Dermoscope Image.
On the other hand, the detection of the outline area of the final skin lesion can be performed manually by the user without applying the level set model (S140). The characteristics of the skin lesion can be analyzed based on the image of the lesion area within the outline of the finally detected skin lesion (S150). In the step S150 of analyzing the features of the skin disease lesion according to an embodiment of the present invention, the user response information about the skin disease related article can be received and the texture feature of the skin disease lesion can be extracted and evaluated.
The user response information according to an embodiment of the present invention may be configured as a response to an inquiry item related to the skin disease shown in FIGS.
Fig. 4 shows the items for routine skin diseases, and Fig. 5 shows items for the malignant skin diseases.
The common skin diseases according to an embodiment of the present invention include acne vulgaris, atopic dermatitis, granuloma annulare, keloid, melanocytic nevus, Urticaria, . ≪ / RTI > The malignant skin disease according to one embodiment of the present invention is malignant melanoma. However, the daily disease and malignant disease according to one embodiment of the present invention are not limited thereto.
The skin lesion characteristics can be extracted by analyzing the image texture of the finally detected lesion area (S150).
As shown in FIG. 6, texture characteristics of skin disease lesions extracted and evaluated according to an embodiment of the present invention are commonly used in ordinary skin diseases and malignant skin diseases such as FOS (First Order Statistics), GLCM (Gray Level Co- (LBP), Discrete Wavelet Texture (DWT), or Fractal Dimension (FD) can be applied to the input image data, The texture features of the skin disease lesion image according to the embodiment are not limited thereto.
The detected texture feature may be optimized through Principal Component Analysis or Genetic Algorithm.
After the step S161 of classifying the skin disease by a plurality of classifiers using the response information of the user for the questionnaire item shown in FIG. 4 or 5 and the texture feature of the extracted skin disease lesion as the input information, A skin disease can be diagnosed (S162) by weighting the result of classification by the classifiers of FIG.
The classifier may be implemented as AdaBoost (Adaptive Boosting), RF (Random Forest), SVM (Support Vector Machine), and PNN (Probabilistic Neural Network). The classifier for normal skin diseases is AdaBoost1 (Support Vector Machine1), SV2 (Support Vector Machine2), and PNN (Probabilistic Neural Network) are composed of 7 classifiers in total. And the classifier for malignant skin diseases may be composed of three classifiers as AdaBoost (Adaptive Boosting), RF (Random Forest) and SVM (Support Vector Machine) like the classifier shown in FIG.
According to one embodiment of the present invention, in the step of classifying a skin disease through a classifier (S161), when a classifier for an ordinary skin disease composed of seven classifiers is used in case of a normal skin disease, A classifier for malignant skin disease consisting of three classifiers may be used.
According to an embodiment of the present invention, the step of diagnosing a skin disease (S162) by weighting the results of classification by a plurality of classifiers, It can be determined as a final disease.
For example, AdaBoost (Adaptive Boosting) is classified as benign for malignant skin disease, Malignant is classified as RF (Random Forest), SVM (Support Vector Machine ), Benign lesions can be diagnosed as benign by receiving two benign lesions.
Among the seven classifiers, AdBoost1 (Adaptive Boosting1), AdaBoost2 (Adaptive Boosting2), Atopic dermatitis, RF1 (Random Forest1), Atopic dermatitis, RF2 (Random Forest2) (Support vector machine 1), keloid, support vector machine 2 (SVM2), acne and PNN (probabilistic neural network) are classified as keloid.
AdaBoost (Adaptive Boosting) among the algorithms implemented in the classifier is a learning algorithm for designing a 'robust' classifier from a linear combination of 'Weak Learners' classifiers.
The AdaBoost (Adaptive Boosting) algorithm is used to train T weak classifiers h t , t ∈ { 1, ..., T} . Each individual weak classifier has a simple form and exhibits relatively inaccurate accuracy. Weak classifiers are mostly decision trees with only one branch (Split) or at most three branches. Each classifier has a Weighted Vote at the final decision making stage,
(1) D t (i) = 1 / m, i = 1, ..., m. ( m is Number of features input)
(2) t = 1, ... , with respect to T, (T is the classifier is used)
a. D t (i) Find the classifier h t that minimizes the weighted error.
b.
If? j <0.5 is not satisfied, the process is terminated. That is, at this stage to calculate the classification errors (ε j) of weak classifiers, the classification error (ε j) is 0.5 or more, the categorizer may be excluded, and select a category at most 0.5 deceptive classification error (ε j).
c. The classifier h t
d. Update the data weights.
(2) b. If no classifier lower than the error ratio of 50% is found in the step, the algorithm is terminated.
When the AdaBoost (Adaptive Boosting) algorithm is terminated, the final strong classifier receives a new vector, and the learned weak classifier can perform classification using the weighted sum of h t . That is, the final classification model can be generated by further increasing the weights to the samples in which the classification error occurs and repeating the above processes. In other words, optimal weights can be found by recalculating the classification error of weak classifiers.
As shown in FIG. 9, AdaBoost (Adaptive Boosting) 1 has a learning cycle of 100, AdaBoost (Adaptive Boosting) 2 has a learning cycle of 200, and the weak classifier class AdaBoost (Adaptive Boosting) 1 and AdaBoost (Adaptive Boosting) 2 are both set as Tree. In case of malignant skin disease classifier, as shown in FIG. 13, learning
Among the algorithms implemented in the classifier, RF (Random Forest) uses arbitrarily confusing decision trees, collects class votes from Leaf Nodes of many trees, and selects the class with the highest number of votes at a time It is an algorithm that can learn the above classes.
The random tree is basically based on a Decision Tree. This decision tree can continue to expand until it is properly classified. Each tree thus becomes a highly distributed classifier by learning the training data almost completely. To eliminate this disadvantage of high dispersion, a plurality of trees are combined to obtain an average.
When the trees are similar to each other, the random tree can select a subset of the different features among all features to be learned so that each tree is different. To perform reliable classification, the random tree can verify the branch using the OOB (Out of Bag) metric. That is, training is performed on any data subset retrieved from a given node, and the remaining data is used to measure the branching performance, and the OOB data can generally be set to about 1/3 of the total data.
As shown in FIG. 10, in the case of the RF (Random Forest) application, the RF1 has a tree depth of 5, the tree number is 10, the RF2 has a tree depth of 10 and the tree number is 50, The RF2 is set to all 15, and in the case of the malignant skin disease classifier, the
SVM (Support Vector Machine) among the algorithms implemented in the classifier is one of the learning models related to the training algorithm for analyzing the data and patterns. Generally, input variables can be derived from one of two kinds. Conventional classification methods including neural networks are designed to minimize the error rate while SVM (Support Vector Machine) is designed not only to minimize the error rate but also maximize the general classification ability by maximizing the margin existing between the two classes .
SVM (Support Vector Machine) classifier can show different classification results according to mapping method using Kernel function, which is divided into Linear, Polynomial, Radial Basis Function (RBF) and Multi-layer Perceptron (MLP). Kernel-based mapping can move the nonlinear data, which is hard to divide in the input space where data is actually arranged, to the high dimensional space called Feature Space, and then perform the linear discrimination of SVM (Support Vector Machine) in this new space.
Since the performance of SVM (Support Vector Machine) depends greatly on the kind of kernel, kernel type can be set differently according to classifier. 11, SVM1 is set as a Radial Basis Function (RBF) and SVM2 is set as a Multi-layer Perceptron. In case of malignant skin disease, as shown in FIG. 11, Radial Basis Function RBF). However, the present invention is not limited thereto.
Among the algorithms implemented in the classifier, the probabilistic neural network (PNN) is modeled as an algorithm learned in two or more training patterns. Based on the data belonging to the existing class and the distance matrix for the new input object, Probability can be calculated and distinguished.
The input layer is defined as a kind of distribution layer that inputs the same input type for all training types. The pattern layer obtains the inner product of the weight vector and the input type X for each training pattern, Can be input to perform a nonlinear operation. Summation Layer summarizes the nonlinear output results obtained from the Pattern Layer. The Output Layer has one weight with two input neurons and can output a binary number consisting of 0 and 1.
12, the Spread Value is set to 5.15, and the number of input layers is determined according to the number of features of each disease, as shown in FIG. 12, in the embodiment of the present invention. And the hidden layer can be set to five. The Spread Value can be derived from the classification performance evaluation for 0.1 interval values from 0 to 10.
Referring to FIG. 14, a skin disease diagnosis system according to an embodiment of the present invention includes an
The
The skin
The contents of the above-described method can be applied in connection with the system according to the embodiment of the present invention. Therefore, the description of the same contents as those of the above-described method with respect to the system is omitted.
One embodiment of the present invention may also be embodied in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. In addition, the computer-readable medium may include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes any information delivery media, including computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism.
It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.
The scope of the present invention is defined by the appended claims rather than the foregoing detailed description, and all changes or modifications derived from the meaning and range of the claims and their equivalents should be construed as being included in the scope of the present invention.
100:
200:
300:
400: lesion analysis unit
410: Document image information receiving section
420: texture feature extracting unit
500: Skin Disease Diagnosis Department
510: Skin Disease Classification Division
520: Whether or not the weight is provided
Claims (13)
(a) obtaining a skin disease image;
(b) processing the skin disease image to remove noise from the acquired skin disease image and preserve outline information of the skin disease lesion;
(c) detecting the skin disease lesion from the treated skin disease image;
(d) analyzing the characteristics of the skin disease lesion based on the detected skin disease lesion; And
(e) classifying and diagnosing a skin disease based on the analyzed characteristics of the skin disease lesion,
The step (d) is performed by receiving user response information on the medical history related to the skin disease, extracting the texture feature information on the skin disease lesion, and evaluating based on the received questionnaire response information and the extracted texture feature information ,
Wherein the step (e) includes the steps of classifying the skin disease by a plurality of classifiers using the user response information for the paperwork received in step (d) and the extracted texture feature information as input information, Diagnosing the skin disease by weighting the frequency of the expected skin disease judgment obtained as a classification result by the classifiers,
Wherein the plurality of classifiers are a plurality of different classifiers selected from AdaBoost (Adaptive Boosting), RF (Random Forest), SVM (Support Vector Machine) and PNN (Probabilistic Neural Network).
Wherein the mask image is acquired through Convex Hull analysis on the skin disease image in the step (b).
Wherein the step (c) comprises applying a level set model based on the obtained mask image to detect an optimized outline area of the skin disease lesion.
The texture characteristics of the skin disease lesion include FOS (First Order Statistics), GLCM (Gray Level Co-occurrence Matrix, LF (Law's Feature), Intensity, GLRLM (Gray Level Run Length Matrix), LBP (Discrete Wavelet Texture) or FD (Fractal Dimension).
An image acquiring unit acquiring a skin disease image;
An image processor for processing the skin disease image to remove noise from the obtained skin disease image and preserve outline information of a skin disease lesion;
A lesion detection unit for detecting the lesion of the skin disease from the skin disease image processed by the image processing unit;
A lesion analyzer for analyzing a characteristic of the skin lesion based on the lesion detected by the lesion detector; And
And a skin disease diagnosis unit for classifying and diagnosing a skin disease based on the characteristics of the skin disease lesion analyzed through the lesion analysis unit,
The lesion analyzing unit includes a document information receiving unit and a texture feature extracting unit. The lesion analyzing unit analyzes the response information received from the user regarding the document related to the skin disease through the document information receiving unit, The feature of the skin disease lesion is evaluated based on the texture feature information extracted by the extracting unit,
Wherein the skin disease diagnosis unit includes a skin disease classifying unit and a weighting unit, the skin disease classifying unit classifies the skin disease by a plurality of classifiers using the response information and the texture feature information as input information, And the skin disease diagnosis unit diagnoses a skin disease based on the given weight, and the skin disease diagnosis unit diagnoses the skin disease based on the weighted value,
Wherein the plurality of classifiers are a plurality of different classifiers selected from AdaBoost (Adaptive Boosting), RF (Random Forest), SVM (Support Vector Machine) and PNN (Probabilistic Neural Network).
Wherein the image processor acquires a mask image through Convex Hull analysis on the skin disease image.
Wherein the lesion detection unit detects an optimized outline area of the skin disease lesion by applying a level set model based on the mask image obtained in the image processing unit.
The texture characteristics of the skin disease lesion include FOS (First Order Statistics), GLCM (Gray Level Co-occurrence Matrix, LF (Law's Feature), Intensity, GLRLM (Gray Level Run Length Matrix), LBP (Discrete Wavelet Texture) or FD (Fractal Dimension).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020160031509A KR101812406B1 (en) | 2016-03-16 | 2016-03-16 | The method and system for diagnosing skin disease |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020160031509A KR101812406B1 (en) | 2016-03-16 | 2016-03-16 | The method and system for diagnosing skin disease |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| KR20170107778A KR20170107778A (en) | 2017-09-26 |
| KR101812406B1 true KR101812406B1 (en) | 2017-12-27 |
Family
ID=60036678
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| KR1020160031509A KR101812406B1 (en) | 2016-03-16 | 2016-03-16 | The method and system for diagnosing skin disease |
Country Status (1)
| Country | Link |
|---|---|
| KR (1) | KR101812406B1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20200015281A (en) | 2018-08-03 | 2020-02-12 | (주) 프로큐라티오 | Method and Device for diagnosing abnormal health conditions |
| KR102165699B1 (en) | 2019-05-24 | 2020-10-14 | 동서대학교 산학협력단 | Skin disease care System for user specific real time service |
| KR20210120408A (en) | 2020-03-26 | 2021-10-07 | 재단법인 아산사회복지재단 | Method and Device for checking the health condition by skin reaction |
| WO2023218485A1 (en) * | 2022-05-11 | 2023-11-16 | Artificial Learning Systems India Pvt Ltd | System and method for classifying skin conditions |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102208087B1 (en) * | 2018-02-28 | 2021-01-27 | 경북대학교 산학협력단 | Apparatus and method for Auto classification of skin disease |
| KR102311654B1 (en) * | 2019-02-01 | 2021-10-08 | 장현재 | Smart skin disease discrimination platform system constituting API engine for discrimination of skin disease using artificial intelligence deep run based on skin image |
| KR102265525B1 (en) * | 2019-05-31 | 2021-06-16 | 재단법인 경북아이티융합 산업기술원 | Method and system for diagnosing skin based on an artificial intelligence |
| EP3986251A4 (en) * | 2019-06-18 | 2023-07-26 | Digital Diagnostics Inc. | Using a set of machine learning diagnostic models to determine a diagnosis based on a skin tone of a patient |
| KR102556350B1 (en) * | 2019-07-25 | 2023-07-17 | 광운대학교 산학협력단 | Method and Apparatus for Calculating Ratio of Lesion Area |
| KR102097740B1 (en) * | 2019-07-25 | 2020-04-06 | 주식회사 딥노이드 | System for Classifying and standardizing of Medical images automatically using Artificial intelligence |
| KR102097742B1 (en) * | 2019-07-31 | 2020-04-06 | 주식회사 딥노이드 | System for Searching medical image using artificial intelligence and Driving method thereof |
| WO2021141187A1 (en) * | 2020-01-08 | 2021-07-15 | 주식회사 에프앤디파트너스 | Skin image-based skin disease identification apparatus |
| KR102389917B1 (en) * | 2020-01-31 | 2022-04-25 | 인하대학교 산학협력단 | System for derermining scabies based using artificial intelligence and thereof method |
| KR102192953B1 (en) * | 2020-03-31 | 2020-12-18 | 신현경 | A system for treating skin damage based on artificial intelligence and providing remote medical service |
| KR102358316B1 (en) * | 2020-04-20 | 2022-02-08 | 서울대학교병원 | Sticker for white balance reference point of medical clinical photos and method and apparatus for analyzing medical clinical photos datum using the same |
| CN113558570A (en) * | 2020-04-29 | 2021-10-29 | 皮智股份有限公司 | Artificial intelligent cloud skin and skin lesion identification method and system |
| WO2022031067A1 (en) * | 2020-08-05 | 2022-02-10 | 주식회사 에이아이투유 | Method and apparatus for diagnosing skin disease by using chatbot |
| KR102419567B1 (en) * | 2020-08-05 | 2022-07-12 | 주식회사 에이아이투유 | Method and apparatus for predicting diagnosis of skin diseases |
| KR102427395B1 (en) * | 2020-11-23 | 2022-07-29 | 한림대학교 산학협력단 | Control method, device and program for skin cancer judgment based on artificial intelligence |
| KR102481179B1 (en) * | 2020-12-30 | 2022-12-26 | (주)엔티엘헬스케어 | Cervical diagnostic camera apparatus with light reflection reduction implemented |
| KR102451102B1 (en) * | 2021-01-05 | 2022-10-06 | 주식회사 오션라이트에이아이 | Companion animal skin lesion classification system and method based on artificial intelligence using clinical and image information |
| KR20230024232A (en) * | 2021-08-11 | 2023-02-20 | 고려대학교 산학협력단 | Method and apparatus for diagnosing skin diseases using image processing |
| WO2023018257A1 (en) * | 2021-08-11 | 2023-02-16 | 고려대학교 산학협력단 | Method and apparatus for estimating cause of contact dermatitis by using user medical interview data and skin image data |
| KR20230024234A (en) * | 2021-08-11 | 2023-02-20 | 고려대학교 산학협력단 | Method and apparatus for remote skin disease diagnosing using augmented and virtual reality |
| KR20240025797A (en) * | 2022-08-19 | 2024-02-27 | 조창숙 | Skin diagnosis apparatus, skin diagnosis system including the same and skin diagnosis method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140036054A1 (en) * | 2012-03-28 | 2014-02-06 | George Zouridakis | Methods and Software for Screening and Diagnosing Skin Lesions and Plant Diseases |
-
2016
- 2016-03-16 KR KR1020160031509A patent/KR101812406B1/en active IP Right Grant
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140036054A1 (en) * | 2012-03-28 | 2014-02-06 | George Zouridakis | Methods and Software for Screening and Diagnosing Skin Lesions and Plant Diseases |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20200015281A (en) | 2018-08-03 | 2020-02-12 | (주) 프로큐라티오 | Method and Device for diagnosing abnormal health conditions |
| KR102165699B1 (en) | 2019-05-24 | 2020-10-14 | 동서대학교 산학협력단 | Skin disease care System for user specific real time service |
| KR20210120408A (en) | 2020-03-26 | 2021-10-07 | 재단법인 아산사회복지재단 | Method and Device for checking the health condition by skin reaction |
| WO2023218485A1 (en) * | 2022-05-11 | 2023-11-16 | Artificial Learning Systems India Pvt Ltd | System and method for classifying skin conditions |
Also Published As
| Publication number | Publication date |
|---|---|
| KR20170107778A (en) | 2017-09-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101812406B1 (en) | The method and system for diagnosing skin disease | |
| Dahab et al. | Automated brain tumor detection and identification using image processing and probabilistic neural network techniques | |
| Karn et al. | Robust retinal blood vessel segmentation using hybrid active contour model | |
| Zhang et al. | Retinal vessel segmentation using multi-scale textons derived from keypoints | |
| US6996549B2 (en) | Computer-aided image analysis | |
| EP1356421B1 (en) | Computer-aided image analysis | |
| Mookiah et al. | Decision support system for age-related macular degeneration using discrete wavelet transform | |
| Adla et al. | Deep learning-based computer aided diagnosis model for skin cancer detection and classification | |
| US20120184840A1 (en) | Automated Measurement of Brain Injury Indices Using Brain CT Images, Injury Data, and Machine Learning | |
| AU2002243783A1 (en) | Computer-aided image analysis | |
| Melekoodappattu et al. | Detection and classification of breast cancer from digital mammograms using hybrid extreme learning machine classifier | |
| Rehman et al. | Microscopic retinal blood vessels detection and segmentation using support vector machine and K‐nearest neighbors | |
| Molaei et al. | FDCNet: Presentation of the fuzzy CNN and fractal feature extraction for detection and classification of tumors | |
| Sharif et al. | Efficient hybrid approach to segment and classify exudates for DR prediction | |
| Rahmawaty et al. | Classification of breast ultrasound images based on texture analysis | |
| Ganeshbabu | Glaucoma image classification using discrete orthogonal stockwell transform | |
| Bagaria et al. | Bone fractures detection using support vector machine and error backpropagation neural network | |
| Behara et al. | Artificial intelligence in medical diagnostics: A review from a South African context | |
| Ganesan et al. | Internet of medical things with cloud-based e-health services for brain tumour detection model using deep convolution neural network | |
| Biswal et al. | Robust classification of neovascularization using random forest classifier via convoluted vascular network | |
| Tamilmani et al. | Early detection of brain cancer using association allotment hierarchical clustering | |
| Wang | Deep Learning-based and Machine Learning-based Application in Skin Cancer Image Classification | |
| Ghoushchi et al. | Clustering of brain tumors in brain MRI images based on extraction of textural and statistical features | |
| Jusoh et al. | An overview of retinal blood vessels segmentation | |
| Rajini et al. | Brain Tumor Segmentation from MRI Images using Deep Learning-based CNN with SVM Classifier |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A201 | Request for examination | ||
| E902 | Notification of reason for refusal | ||
| E701 | Decision to grant or registration of patent right | ||
| GRNT | Written decision to grant |