WO2018098987A1 - 中医舌诊图像处理系统及方法 - Google Patents

中医舌诊图像处理系统及方法 Download PDF

Info

Publication number
WO2018098987A1
WO2018098987A1 PCT/CN2017/082251 CN2017082251W WO2018098987A1 WO 2018098987 A1 WO2018098987 A1 WO 2018098987A1 CN 2017082251 W CN2017082251 W CN 2017082251W WO 2018098987 A1 WO2018098987 A1 WO 2018098987A1
Authority
WO
WIPO (PCT)
Prior art keywords
tongue
image
coating
color
diagnosis
Prior art date
Application number
PCT/CN2017/082251
Other languages
English (en)
French (fr)
Inventor
高伟明
张贯京
葛新科
徐勇
Original Assignee
深圳市易特科信息技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳市易特科信息技术有限公司 filed Critical 深圳市易特科信息技术有限公司
Publication of WO2018098987A1 publication Critical patent/WO2018098987A1/zh

Links

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/45For evaluating or diagnosing the musculoskeletal system or teeth
    • A61B5/4538Evaluating a particular part of the muscoloskeletal system or a particular medical condition
    • A61B5/4542Evaluating the mouth, e.g. the jaw
    • A61B5/4552Evaluating soft tissue within the mouth, e.g. gums or tongue
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/48Other medical applications
    • A61B5/4854Diagnosis based on concepts of traditional oriental medicine
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00Image enhancement or restoration
    • G06T5/70Denoising; Smoothing
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0012Biomedical image inspection

Definitions

  • the present invention relates to the field of image processing technologies, and in particular, to a Chinese medicine tongue diagnosis image processing system and method.
  • TCM tongue diagnosis images are used to analyze the tongue diagnosis image to provide a reference for disease diagnosis.
  • the main purpose of the present invention is to provide a TCM tongue diagnosis image processing system and method for the deficiencies and shortcomings of the prior art, which can automatically screen out the patient's tongue image from the tongue diagnosis database and extract from the tongue image.
  • the tongue part analyzes and recognizes various attributes of the tongue, and outputs a multi-dimensional tongue image feature vector that quantitatively describes various attributes, which provides a reference for Chinese doctors to carry out TCM tongue diagnosis.
  • the present invention provides a TCM tongue diagnosis image processing system, which is applied to a computer, and the computer is connected to a tongue diagnosis database through a database link, and the TCM tongue diagnosis image processing system includes:
  • a tongue image preprocessing module configured to acquire a tongue image of the patient from the tongue diagnosis database according to a patient's tongue diagnosis number, and perform denoising preprocessing on the acquired tongue image
  • a tongue extraction module configured to divide the pre-processed tongue image to obtain a tongue image
  • a moss identification module configured to divide the tongue image into a health condition corresponding to different organs according to the theory of traditional Chinese medicine The five parts of the tongue, and the color and texture features of the tongue coating in the five parts of the tongue;
  • a tongue coating analysis module configured to analyze a characteristic value of a thin thickness of the tongue coating, a moistening degree of the tongue coating, and a degree of rot of the tongue according to the color feature and the texture feature of the tongue coating;
  • a tongue detection module configured to detect the length and width of the tongue and quantify into a feature value describing a tongue condition
  • a tongue shape analysis module for detecting a tongue size and quantifying into a feature value describing a degree of fatness of the tongue
  • a tongue image output module which is used for combining the thickness values of the tongue coating, the moisturization of the tongue, the rot of the tongue, the condition of the tongue, and the degree of fatness of the tongue into a multi-dimensional tongue image feature vector, and
  • the output unit of the computer outputs the tongue image feature vector.
  • the tongue coating analysis module is further configured to: calculate a ratio of the color value of the tongue coating to the color value of the tongue, and obtain a characteristic value describing the thin thickness of the tongue coating by using the wavelet transform coefficient; and transform the two-color reflection model of the tongue coating to the color
  • the high-light pixel is obtained from the height space, and the region is extended from the high-light pixel to obtain the bright spot region. According to the size of the bright spot region, a feature value describing the moisturization degree of the tongue coating is obtained; and the roughness of the tongue coating is calculated according to the texture feature of the tongue coating. According to the roughness of the tongue coating, a characteristic value describing the degree of rot of the tongue is obtained.
  • the tongue extraction module is further configured to use a threshold segmentation according to a color parameter of the tongue image to obtain an initial contour line of the tongue body, and use the RG B color space for the boundary blur feature of the tongue color relative to the color of the human skin color.
  • the G component in the middle enhances the weak boundary of the tongue to obtain an enhanced tongue image, and extracts the tongue contour from the enhanced tongue image to obtain the tongue image.
  • the tongue detection module is further configured to detect the length and width of the tongue by using a multi-scale edge detection method, and describe the size of the tongue according to the length and width of the tongue and quantize the tongue as described. The characteristic value of the grain condition.
  • the tongue shape analysis module is further configured to fit the obtained tongue edge point into a quadratic curve by using a least square method, and determine the tongue size according to the size of the quadratic term and quantize the A characteristic value describing the degree of fatness of the tongue.
  • the present invention further provides a TCM tongue diagnosis image processing method, which is applied to a computer, and the computer is connected to a tongue diagnosis database through a database link, and the method includes the following steps:
  • the tongue image is divided into five parts of the tongue corresponding to the health status of different organs of the human body, and the color features and texture features of the tongue coating in the five parts of the tongue are respectively identified;
  • the step of analyzing the characteristic values of the thin thickness of the tongue coating, the degree of moisturization of the tongue coating, and the degree of the texture of the tongue coating according to the color characteristics and the texture characteristics of the tongue coating comprises the steps of: calculating the ratio of the color value of the tongue coating to the color value of the tongue body, Using the wavelet transform coefficient to obtain a characteristic value describing the thin thickness of the tongue coating; transforming the two-color reflection model of the tongue coating into the chromaticity space to obtain high-light pixels, and extending the region from the high-light pixel to obtain the bright spot area, according to the size of the bright spot area To obtain a characteristic value describing the degree of moisturization of the tongue coating; calculate the roughness of the tongue coating according to the texture characteristics of the tongue coating, and obtain a characteristic value describing the degree of corrosion of the tongue coating according to the roughness of the tongue coating.
  • the step of dividing the tongue image to obtain a tongue image comprises the steps of: according to the tongue
  • the color parameter of the head image is segmented to obtain the initial contour of the tongue;
  • the boundary blur feature for the color of the tongue relative to the color of the human skin uses the G component in the RGB color space to enhance the weak boundary of the tongue to obtain an enhanced tongue image;
  • the tongue image is obtained by extracting the tongue contour from the enhanced tongue image.
  • the step of detecting the length and width of the tongue and quantizing into a feature value describing the tongue condition comprises: detecting the length and width of the tongue by using a multi-scale edge detection method; according to the length of the tongue And the width describes the size of the tongue and quantifies the characteristic value of the tongue condition described.
  • the step of detecting the size of the tongue and quantifying into a feature value describing the degree of fatness of the tongue includes the steps of: fitting the obtained edge point of the tongue to a quadratic curve by least squares method; The size of the secondary term is used to determine the size of the tongue and quantify it as the characteristic value describing the degree of fatness of the tongue.
  • the tongue image of the patient can be automatically screened from the tongue diagnosis database, from the tongue image Automatically extracts the tongue part, automatically analyzes and recognizes various attributes of the tongue, and outputs a multi-dimensional tongue image feature vector that quantitatively describes various attributes, which provides a reference for Chinese doctors to carry out TCM tongue diagnosis, which is beneficial to promote The development of TCM tongue diagnosis.
  • FIG. 1 is a schematic diagram of an application environment of a preferred embodiment of a Chinese medical tongue image processing system according to the present invention
  • FIG. 2 is a schematic view showing a division of five parts of a tongue in a tongue image.
  • FIG. 3 is a flow chart of a preferred embodiment of the Chinese medical tongue image processing method of the present invention.
  • FIG. 1 is a schematic diagram of an application environment of a preferred embodiment of a Chinese medical tongue image processing system according to the present invention.
  • the TCM Tongue Diagnosis Image Processing System 10 is installed and operated in the computer 1, and the computer 1 establishes a communication connection with the Tongue Diagnostics Database 2 via the database link 3.
  • the computer 1 may be a computing device having data processing and communication functions such as a personal computer or a server.
  • the tongue diagnosis database 2 is a key value database or a document database, and stores a large number of tongue images of different patients for use by a Chinese medicine practitioner for tongue diagnosis.
  • the database link 3 can be a database link such as JDBC or ODBC.
  • the computer 1 further includes, but is not limited to, an input unit 11, a storage unit 12, a processing unit 13, and an output unit 14.
  • the input unit 11, the storage unit 12 and the output unit 14 are all connected to the processing unit 13 through a data bus and a control line, and can perform information interaction with the TCM tongue image processing system 10 through the processing unit 13.
  • the input unit 11 may be an input device such as a keyboard or a handwriting touch screen, and is used for the Chinese medicine practitioner to input the TCM tongue diagnosis information of the patient.
  • the patient's TCM tongue may be input on the input unit 11.
  • Information such as the clinic number or user name.
  • the storage unit 12 can be a read only memory unit ROM, an electrically erasable memory unit EEPROM or a flash memory unit FLASH.
  • the processing unit 13 may be a central processing unit (CPU), a microprocessor, a microcontroller (MCU), a data processing chip, or an information processing unit having data processing functions.
  • the output unit 14 may be a display screen for displaying a tongue image and a tongue image feature vector, or a printer for printing a tongue image and a tongue image feature vector.
  • the TCM tongue image processing system 10 includes, but is not limited to, a tongue image preprocessing module 101, a tongue extraction module 102, a moss identification module 103, a tongue coating analysis module 104, and a tongue.
  • module refers to a series of computer program instruction segments that can be executed by the processing unit 13 of the computer 1 and that can perform the image processing function of the Chinese medicine tongue diagnosis of the present invention, which are stored in the storage unit of the computer 1. 12 in.
  • the tongue image preprocessing module 101 is configured to acquire a tongue image of the patient from the tongue diagnosis database 2 according to a tongue diagnosis number (ID) input by the patient, and perform denoising preprocessing on the acquired tongue image. Subsequent tongue image analysis and processing provides the basis.
  • the tongue image is in the process of shooting, possibly due to the optical system, Motion, etc. cause blurring of the image, as well as noise from factors such as circuit and optics, which degrade the image quality; in addition, the patient is skewed due to failure to grasp the correct posture of the tongue.
  • pre-processing the tongue image the effects of these unfavorable factors on subsequent tongue image analysis are reduced or removed.
  • the tongue extraction module 102 is configured to perform a segmentation process on the tongue image to obtain a tongue image. Specifically, the tongue extraction module 102 uses a threshold segmentation to obtain an initial contour line of the tongue according to the color parameter of the tongue image, and uses the G component in the RGB color space for the boundary blur feature of the tongue color relative to the human skin color. Enhancing the weak boundary of the tongue to obtain an enhanced tongue image, and using the snake contour extraction method to extract the tongue contour from the enhanced tongue image to obtain a tongue image.
  • the snake contour extraction method is an image image contour extraction method for image processing in the prior art, which is not specifically described in the embodiment of the present invention.
  • the moss identification module 103 is configured to divide the tongue image into five parts of the tongue corresponding to the health status of different organs of the human according to the theory of traditional Chinese medicine, and respectively identify the color features and texture features of the tongue coating in the five parts of the tongue. . Part A shown in Figure 2 corresponds to the kidney, part B corresponds to the stomach and spleen, part C corresponds to the heart and lungs, and parts D and E correspond to the liver and bladder.
  • the moss identification module 103 is also used to identify the moss and tongue color of the five parts of the tongue respectively, and is described by a multi-dimensional vector.
  • the quantitative description is as follows: the proportion of the tongue coating, the tongue coating Color, tongue color ratio, tongue color, tongue color ratio, tongue ratio, tongue color, tongue color ratio, tongue color, tongue color ratio.
  • the tongue color has five types: light red, pale white, red, dark red, and cyan; and the color of the tongue is divided into white, yellow, yellow, and gray.
  • the tongue coating analysis module 104 is configured to analyze the characteristic values of the thin thickness of the tongue coating, the degree of moisturization of the tongue coating, and the degree of greasy tongue coating according to the color features and texture features of the tongue coating. Specifically, the tongue coating analysis module 104 is configured to calculate the ratio of the color value of the tongue coating to the color value of the tongue (for example, converting the color characteristic value of the tongue to the Luv space, and calculating the ratio of the u value of the tongue to the u value of the tongue), and then utilizing Wavelet transform (for example, 2D Gabor wavelet transform) coefficients are used to describe the thickness characteristics of the tongue coating region, thus obtaining a feature value describing the thin thickness of the tongue coating; the tongue coating analysis module 104 is used to apply a two-color reflection model of the tongue coating (for example, a two-color reflection model proposed by Shafer) Transforming into the chromaticity space to obtain high-light pixels, extending from the high-light pixels to obtain the bright spot area, and then obtaining
  • the tongue detecting module 105 is configured to detect the length and width of the tongue and quantize it as a feature value describing the tongue condition. Specifically, the tongue detection module 105 detects the length and width of the tongue by using a multi-scale edge detection method, and describes the size of the tongue according to the length and width of the tongue and quantizes it as a feature value describing the condition of the tongue, for example. The values are 1, 2, and 3, which indicate the absence, presence, and severity of the tongue.
  • the multi-scale edge detection method is a prior art for detecting tongue continuity in the prior art, and is not specifically described in the embodiment of the present invention.
  • the tongue analysis module 106 is configured to detect the size of the tongue and quantify it as a feature value describing the degree of fatness of the tongue; specifically, the tongue shape analysis module 106 uses the least square method to extract the edge of the tongue. Combine the quadratic curve, and judge the size of the tongue according to the size of the quadratic item and quantify it as a characteristic value describing the degree of fatness of the tongue, for example, the values are 1, 2, 3, respectively, indicating the sputum and middle of the tongue. Three levels of fat.
  • the tongue image output module 107 is configured to combine the characteristic values of the thin thickness of the tongue coating, the moisturizing degree of the tongue, the degree of tongue rot, the condition of the tongue, and the degree of fatness of the tongue into a multi-dimensional tongue image feature vector, and
  • the tongue image feature vector is output through the output unit 14 to provide a reference for the Chinese doctor to perform the Chinese medicine tongue diagnosis.
  • the tongue image output module 107 combines the characteristic values of the thin thickness of the tongue coating, the moisturizing degree of the tongue, the rot of the tongue, the condition of the tongue, and the degree of the fatness of the tongue into a multi-dimensional tongue image feature vector, and the tongue image is
  • the feature is displayed on the display screen of the output unit 14, or the printer controlling the output unit 14 prints the tongue image feature, so that the Chinese doctor can understand the patient's tongue image, thereby facilitating the diagnosis of the patient by the doctor's tongue diagnosis. Condition.
  • FIG. 3 is a flow chart of a preferred embodiment of the Chinese medical tongue image processing method of the present invention.
  • the TCM tongue diagnosis image processing method includes the following steps:
  • Step S31 obtaining a tongue image of the patient from the tongue diagnosis database according to the tongue diagnosis number of the patient, and performing denoising preprocessing on the acquired tongue image; specifically, the tongue image preprocessing module 101 is based on the tongue diagnosis input by the patient.
  • the number is obtained from the tongue diagnosis database 2, and the obtained tongue image is subjected to denoising preprocessing, which provides a basis for subsequent tongue image analysis and processing.
  • the image of the tongue may be degraded during the shooting process due to blurring of the image caused by optical system, motion, etc., and noise due to factors such as circuit and optics.
  • the patient is unable to grasp the correct tongue posture and makes the tongue The body is skewed and so on.
  • Step S32 performing a segmentation process on the tongue image to obtain a tongue image
  • the tongue extraction module 102 performs segmentation processing on the tongue image to obtain a tongue image, and includes the following steps:: using a threshold segmentation according to a color parameter of the tongue image The initial contour of the tongue; for the boundary blur feature of the tongue color relative to the human skin color, the G component in the RGB color space is used to enhance the weak boundary of the tongue to obtain an enhanced tongue image; the snake contour extraction method is used to enhance the shape The tongue image extracts the contour of the tongue to obtain a tongue image.
  • Step S33 according to the theory of traditional Chinese medicine, the tongue image is divided into five parts of the tongue corresponding to the health status of different organs of the human body, and the color characteristics and texture features of the tongue coating in the five parts of the tongue are respectively identified; specifically, the moss
  • the recognition module 103 divides the tongue into five parts of the tongue corresponding to the health status of different organs of the human body according to the theory of traditional Chinese medicine.
  • the part A shown in Fig. 2 corresponds to the kidney
  • the part B corresponds to the stomach and the spleen
  • the part C corresponds to the heart and the lung
  • D The liver and bladder correspond to the E part.
  • the moss identification module 103 respectively identifies the color of the tongue coating (including the mossy and the tongue color) of the five parts of the tongue, and describes it by a multi-dimensional vector. For example, taking the area A as an example, the quantitative description is as follows: the proportion of the tongue coating, the tongue coating Color, tongue color ratio, tongue color, tongue color ratio, tongue ratio, tongue color, tongue color ratio, tongue color, tongue color ratio. Among them, the tongue color has five types: light red, pale white, red, dark red, and cyan; and the tongue color is divided into white, yellow, yellow, and gray.
  • Step S34 analyzing the characteristic values of the thin thickness of the tongue coating, the degree of moisturization of the tongue coating, and the degree of the texture of the tongue coating according to the color characteristics and texture characteristics of the tongue coating; specifically, the tongue coating analysis module 104 calculates the proportion of the color value of the tongue coating to the color value of the tongue body.
  • the tongue coating analysis module 104 transforms the two-color reflection model of the tongue coating (for example, the two-color reflection model proposed by Shafer) into the chromaticity space to obtain a high-light pixel, and extends the region from the high-light pixel to obtain a bright spot.
  • the two-color reflection model of the tongue coating for example, the two-color reflection model proposed by Shafer
  • Step S35 detecting the length and width of the tongue and quantizing it as a feature value describing the condition of the tongue.
  • the tongue detection module 105 uses a multi-scale edge detection method to detect the length and width of the tongue pattern, and describes the size of the tongue according to the length and width of the tongue and quantizes it as a feature value describing the tongue condition.
  • the values are 1, 2, and 3, which indicate the absence, presence, and severity of the tongue.
  • Step S36 detecting the size of the tongue surface and quantifying into a feature value describing the degree of fatness of the tongue; specifically, the tongue shape analysis module 106 uses the least square method to fit the obtained tongue edge point to a quadratic curve.
  • the size of the quadratic term the size of the tongue is judged and quantified as a characteristic value describing the degree of fatness of the tongue.
  • the values are 1, 2, and 3, which indicate the three degrees of sputum, middle, and fat of the tongue.
  • Step S37 combining characteristic values of thin thickness of tongue coating, moisturizing degree of tongue coating, degree of tongue rot, degree of tongue and tongue, and degree of tongue fatness into a multi-dimensional tongue image feature vector, and said by an output unit Tongue image feature vector output.
  • the tongue image output module 107 combines the characteristic values of the thin thickness of the tongue coating, the moisturizing degree of the tongue, the rot of the tongue, the condition of the tongue, and the degree of the fatness of the tongue into a multi-dimensional tongue image feature vector, and the tongue image is
  • the feature is displayed on the display screen of the output unit 14, or the printer controlling the output unit 14 prints the tongue image feature, so that the Chinese doctor can understand the patient's tongue image, thereby facilitating the diagnosis of the patient by the doctor's tongue diagnosis. Condition.
  • the traditional Chinese medicine tongue image processing system and method of the present invention can automatically screen out the tongue image of the patient from the tongue diagnosis database, extract the tongue part from the tongue image, and automatically analyze various attributes of the tongue body. Identification, output quantitative description of a multi-dimensional tongue image feature vector of various attributes, providing a reference for Chinese doctors to carry out TCM tongue diagnosis, which is conducive to promoting the development of TCM tongue diagnosis.
  • the tongue image of the patient can be automatically screened from the tongue diagnosis database, from the tongue image Automatically extracts the tongue part, automatically analyzes and recognizes various attributes of the tongue, and outputs a multi-dimensional tongue image feature vector that quantitatively describes various attributes, which provides a reference for Chinese doctors to carry out TCM tongue diagnosis, which is beneficial to promote The development of TCM tongue diagnosis.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Veterinary Medicine (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Public Health (AREA)
  • Animal Behavior & Ethology (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Orthopedic Medicine & Surgery (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Rheumatology (AREA)
  • Alternative & Traditional Medicine (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • Image Processing (AREA)

Abstract

公开了一种中医舌诊图像处理系统(10)及方法。该系统(10)应用于计算机(1)中,该计算机(1)通过数据库链接(3)连接至舌诊数据库(2)。该系统(10)包括舌像预处理模块(101)、舌体提取模块(102)、苔质识别模块(103)、舌苔分析模块(104)、舌纹检测模块(105)、舌形分析模块(106)以及舌像输出模块(107)。该中医舌诊图像处理系统(10)及方法能够从舌诊数据库(2)中自动筛选出患者的舌头图像,从舌头图像中提取出舌体部分,从舌体中分析并识别出舌苔薄厚度、舌苔润燥度、舌苔腐腻状况、舌纹状况以及舌体胖瘦程度等舌体属性,将各个舌体属性的特征值组合成一个多维的舌像特征向量,并通过输出单元将舌像特征向量输出,为中医舌诊提供参考依据。

Description

说明书 发明名称:中医舌诊图像处理系统及方法 技术领域
[0001] 本发明涉及图像处理技术领域, 尤其涉及一种中医舌诊图像处理系统及方法。
背景技术
[0002] 传统中医包括 "望、 闻、 问、 切"四诊, 而舌诊又是望诊的关键内容, 属中医临 床必察之项。 在中医学理论中, 人体被认为是一个有机统一的整体, 其每个部 分的变化都与整体有着密不可分的关系。 也正是如此, 舌像被认为是人体生理 病变的最直观反映, 如气血津液、 人之精气等内部脏腑的重要信息均可通过舌 像获得。 因此中医舌诊的优势显而易见, 无论人体内五脏六腑多么复杂的病理 症状, 均可直观、 快捷地通过观测舌像得知原委, 且舌诊也可以指导相关的处 方用药以及病情预防。 然而, 中医舌诊也有其传统的弊端, 它过于依赖中医师 的主观观察, 诊断结果通常也因人而异, 且基本不具有可重复性, 而这也极大 的阻碍了中医舌诊的进一步发展。 针对中医舌诊的弊端, 对于舌诊客观化的研 究就显得更加重要。 如果能很好的克服这一难题, 不仅能够很好的迎合中医所 提倡的标准化和定量化要求, 而且可以更好的推动舌诊实际应用价值的提升。 因此, 为了更好的推广中医舌诊, 利用计算机辅助处理并识别中医舌诊图像显 得尤为重要, 采用图像处理和数据挖掘技术, 对舌诊图像进行分析为病情诊断 提供参考依据, 是发展中医舌诊的一条创新之路, 以便推动中医舌诊的进一步 发展。
技术问题
[0003] 本发明的主要目的在于针对现有技术的不足与缺陷, 提供一种中医舌诊图像处 理系统及方法, 能够从舌诊数据库中自动筛选出患者的舌头图像, 从舌头图像 中提取出舌体部分, 对舌体的各种属性进行分析和识别, 输出定量描述各种属 性的一个多维的舌像特征向量, 为中医生进行中医舌诊提供参考依据。
问题的解决方案
技术解决方案 [0004] 为实现上述目的, 本发明提供了一种中医舌诊图像处理系统, 应用于计算机中 , 该计算机通过数据库链接连接至舌诊数据库, 所述中医舌诊图像处理系统包 括:
[0005] 舌像预处理模块, 用于根据患者的舌诊编号从所述舌诊数据库中获取患者的舌 头图像, 并对获取的舌头图像进行去噪声预处理;
[0006] 舌体提取模块, 用于将预处理后的舌头图像进行分割处理得到舌体图像; [0007] 苔质识别模块, 用于根据中医理论将舌体图像分成对应于人不同器官健康状况 的舌体五大部分, 以及分别对所述舌体五大部分中舌苔的颜色特征和纹理特征 进行识别;
[0008] 舌苔分析模块, 用于根据所述舌苔的颜色特征和纹理特征分析舌苔薄厚度、 舌 苔润燥度以及舌苔腐腻度的特征值;
[0009] 舌纹检测模块, 用于检测舌纹的长度和宽度并量化为一个描述舌纹状况的特征 值;
[0010] 舌形分析模块, 用于通过检测舌面大小并量化为一个描述舌体胖痩程度的特征 值;
[0011] 舌像输出模块, 用于将舌苔薄厚度、 舌苔润燥度、 舌苔腐腻状况、 舌纹状况以 及舌体胖痩程度的特征值组合成一个多维的舌像特征向量, 以及通过所述计算 机的输出单元将所述舌像特征向量输出。
[0012] 优选的, 所述舌苔分析模块还用于: 计算舌苔颜色值占舌体颜色值的比例, 利 用小波变换系数得到一个描述舌苔薄厚度的特征值; 将舌苔的双色反射模型变 换到色度空间得到高光象素, 从高光象素出发进行区域延伸来获得亮斑区域, 根据亮斑区域的大小来得到一个描述舌苔润燥度的特征值; 根据舌苔的测纹理 特征计算舌苔的粗糙度, 根据该舌苔的粗糙度得到一个描述舌苔腐腻度的特征 值。
[0013] 优选的, 所述舌体提取模块还用于根据舌头图像的颜色参数采用阈值分割得到 舌体的初始轮廓线, 针对舌体颜色相对于人体皮肤颜色的边界模糊特点利用 RG B色彩空间中的 G分量来增强舌体的弱边界得到增强型舌头图像, 以及从增强型 舌头图像提取舌体轮廓得到所述舌体图像。 [0014] 优选的, 所述舌纹检测模块还用于采用多尺度边缘检测方法检测出舌纹的长度 和宽度, 根据舌纹的长度和宽度描述出舌纹的大小并量化为所述描述舌纹状况 的特征值。
[0015] 优选的, 所述舌形分析模块还用于利用最小二乘法将得到的舌体边缘点拟合为 二次曲线, 以及根据二次项的大小来判断舌面大小并量化为所述描述舌体胖痩 程度的特征值。
[0016] 为实现上述目的, 本发明还提供一种中医舌诊图像处理方法, 应用于计算机中 , 该计算机通过数据库链接连接至舌诊数据库, 该方法包括如下步骤:
[0017] 根据患者的舌诊编号从所述舌诊数据库中获取患者的舌头图像, 并对获取的舌 头图像进行去噪声预处理;
[0018] 将预处理后的舌头图像进行分割处理得到舌体图像;
[0019] 根据中医理论将舌体图像分成对应于人不同器官健康状况的舌体五大部分, 以 及分别对所述舌体五大部分中舌苔的颜色特征和纹理特征进行识别;
[0020] 根据所述舌苔的颜色特征和纹理特征分析舌苔薄厚度、 舌苔润燥度以及舌苔腐 腻度的特征值;
[0021] 检测舌纹的长度和宽度并量化为一个描述舌纹状况的特征值;
[0022] 检测舌面大小并量化为一个描述舌体胖痩程度的特征值;
[0023] 将所述舌苔薄厚度、 舌苔润燥度、 舌苔腐腻状况、 舌纹状况以及舌体胖痩程度 的特征值组合成一个多维的舌像特征向量, 以及通过所述计算机的输出单元将 所述舌像特征向量输出。
[0024] 优选的, 所述根据舌苔的颜色特征和纹理特征分析舌苔薄厚度、 舌苔润燥度以 及舌苔腐腻度的特征值的步骤包括步骤: 计算舌苔颜色值占舌体颜色值的比例 , 利用小波变换系数得到一个描述舌苔薄厚度的特征值; 将舌苔的双色反射模 型变换到色度空间得到高光象素, 从高光象素出发进行区域延伸来获得亮斑区 域, 根据亮斑区域的大小来得到一个描述舌苔润燥度的特征值; 根据舌苔的测 纹理特征计算舌苔的粗糙度, 根据该舌苔的粗糙度得到一个描述舌苔腐腻度的 特征值。
[0025] 优选的, 所述将舌头图像进行分割处理得到舌体图像的步骤包括步骤: 根据舌 头图像的颜色参数采用阈值分割得到舌体的初始轮廓线; 针对舌体颜色相对于 人体皮肤颜色的边界模糊特点利用 RGB色彩空间中的 G分量来增强舌体的弱边界 得到增强型舌头图像; 从增强型舌头图像提取舌体轮廓得到所述舌体图像。
[0026] 优选的, 所述检测舌纹的长度和宽度并量化为一个描述舌纹状况的特征值的步 骤包括: 采用多尺度边缘检测方法检测出舌纹的长度和宽度; 根据舌纹的长度 和宽度描述出舌纹的大小并量化为所述描述舌纹状况的特征值。
[0027] 优选的, 所述检测舌面大小并量化为一个描述舌体胖痩程度的特征值的步骤包 括步骤: 利用最小二乘法将得到的舌体边缘点拟合为二次曲线; 根据二次项的 大小来判断舌面大小并量化为所述描述舌体胖痩程度的特征值。
发明的有益效果
有益效果
[0028] 相较于现有技术, 本发明所述中医舌诊图像处理系统及方法采用上述技术方案 , 达到了如下技术效果: 能够从舌诊数据库中自动筛选出患者的舌头图像, 从 舌图像中自动提取出舌体部分, 对舌体的各种属性进行自动分析和识别, 输出 定量描述各种属性的一个多维的舌像特征向量, 为中医生进行中医舌诊提供参 考依据, 有利于推动中医舌诊的发展。
对附图的简要说明
附图说明
[0029] 图 1是本发明中医舌诊图像处理系统优选实施例的应用环境示意图;
[0030] 图 2是舌体图像中的舌体五大部分划分示意图。
[0031] 图 3是本发明中医舌诊图像处理方法优选实施例的流程图。
[0032] 本发明目的实现、 功能特点及优点将结合实施例, 参照附图做进一步说明。
实施该发明的最佳实施例
本发明的最佳实施方式
[0033] 为更进一步阐述本发明为达成上述目的所采取的技术手段及功效, 以下结合附 图及较佳实施例, 对本发明的具体实施方式、 结构、 特征及其功效进行详细说 明。 应当理解, 此处所描述的具体实施例仅仅用以解释本发明, 并不用于限定 本发明。
[0034] 参照图 1所示, 图 1是本发明中医舌诊图像处理系统优选实施例的应用环境示意 图。 在本实施例中, 所述中医舌诊图像处理系统 10安装并运行于计算机 1中, 所 述计算机 1通过数据库链接 3与舌诊数据库 2建立通信连接。 所述计算机 1可以为 一种个人计算机、 服务器等具有数据处理和通信功能的计算装置。 所述舌诊数 据库 2为一种键值数据库或者文档数据库, 存储有大量不同患者的舌诊图像, 用 于供中医师进行中医舌诊吋使用。 所述数据库链接 3可以为 JDBC或 ODBC等幵放 式数据库链接。
[0035] 所述计算机 1还包括, 但不仅限于, 输入单元 11、 存储单元 12、 处理单元 13以 及输出单元 14。 所述输入单元 11、 存储单元 12和输出单元 14均通过数据总线和 控制线连接至处理单元 13, 并能通过处理单元 13与所述中医舌诊图像处理系统 1 0进行信息交互。
[0036] 在本实施例中, 所述输入单元 11可以为键盘或手写触摸屏等输入设备, 用于供 中医师输入患者的中医舌诊信息, 例如, 可以在输入单元 11上输入患者的中医 舌诊编号或者用户名等信息。 所述存储单元 12可以为一种只读存储单元 ROM, 电可擦写存储单元 EEPROM或快闪存储单元 FLASH等存储器。 所述处理单元 13 可以为一种中央处理器 (CPU) 、 微处理器、 微控制器 (MCU) 、 数据处理芯 片、 或者具有数据处理功能的信息处理单元。 所述输出单元 14可以为一种用于 显示舌诊图像以及舌像特征向量的显示屏, 也可以为一种用于打印舌诊图像以 及舌像特征向量的打印机。
[0037] 在本实施例中, 所述中医舌诊图像处理系统 10包括, 但不局限于, 舌像预处理 模块 101、 舌体提取模块 102、 苔质识别模块 103、 舌苔分析模块 104、 舌纹检测 模块 105、 舌形分析模块 106以及舌像输出模块 107。 本发明所称的模块是指一种 能够被所述计算机 1的处理单元 13执行并且能够完成本发明中医舌诊图像处理功 能的一系列计算机程序指令段, 其存储在所述计算机 1的存储单元 12中。
[0038] 所述舌像预处理模块 101用于根据患者输入的舌诊编号 (ID) 从所述舌诊数据 库 2中获取患者的舌头图像, 并对获取的舌头图像进行去噪声预处理, 为后续的 舌头图像分析与处理提供基础。 舌头图像在拍摄过程中, 可能由于光学系统、 运动等造成图像的模糊, 以及源自电路和光学等因素的噪声而使得图像质量发 生退化; 另外患者由于未能掌握正确的伸舌姿势而使得舌体发生歪斜等等。 通 过对舌图像进行预处理, 减小或去除这些不利因素对后续舌头图像分析造成的 影响。
[0039] 所述舌体提取模块 102用于将舌头图像进行分割处理得到舌体图像。 具体地, 所述舌体提取模块 102根据舌头图像的颜色参数采用阈值分割得到舌体的初始轮 廓线, 针对舌体颜色相对于人体皮肤颜色的边界模糊特点, 利用 RGB色彩空间 中的 G分量来增强舌体的弱边界得到增强型舌头图像, 以及采用 snake轮廓提取 方法从增强型舌头图像提取舌体轮廓得到舌体图像。 所述 snake轮廓提取方法为 现有技术中图像处理的像图像轮廓提取方法, 本发明实施例不作具体赘述。
[0040] 所述苔质识别模块 103用于根据中医理论将舌体图像分成对应于人不同器官健 康状况的舌体五大部分, 以及分别对舌体五大部分中舌苔的颜色特征和纹理特 征进行识别。 如图 2所示的 A部分对应于肾, B部分对应胃和脾, C部分对应心脏 和肺, D和 E部分对应肝脏和膀胱。 所述苔质识别模块 103还用于分别对舌体五大 部分的苔质和舌质颜色进行识别, 用一个多维的向量来描述, 例如以 A区为例, 定量描述为: 舌苔比例, 舌苔主颜色, 舌苔主颜色比例, 舌苔次颜色, 舌苔次 颜色比例, 舌质比例, 舌质主颜色, 舌质主颜色比例, 舌质次颜色, 舌质次颜 色比例。 其中, 舌质颜色有淡红、 淡白、 红、 暗红、 青紫 5种类型; 而舌苔颜色 分为白、 淡黄、 黄、 灰 4种类型。
[0041] 所述舌苔分析模块 104用于根据所述舌苔的颜色特征和纹理特征分析舌苔薄厚 度、 舌苔润燥度以及舌苔腐腻度的特征值。 具体地, 舌苔分析模块 104用于计算 舌苔颜色值占舌体颜色值的比例 (例如, 将舌体的颜色特征值转换到 Luv空间, 计算舌苔 u值占舌体 u值的比例) , 再利用小波变换 (例如 2D Gabor小波变换) 系数来描述舌苔区域的厚度特征, 这样得到一个描述舌苔薄厚度的特征值; 舌 苔分析模块 104用于将舌苔的双色反射模型 (例如 Shafer提出的双色反射模型) 变换到色度空间得到高光象素, 从高光象素出发进行区域延伸来获得亮斑区域 , 再根据亮斑区域的大小来得到一个描述舌苔润燥度的特征值; 舌苔分析模块 1 04还用于根据舌苔的测纹理特征计算舌苔的粗糙度 (例如 Rosenfeld/Tamum粗糙 度模型) , 以及根据该舌苔的粗糙度得到一个描述舌苔腐腻度的特征值。
[0042] 所述舌纹检测模块 105用于检测舌纹的长度和宽度并量化为一个描述舌纹状况 的特征值。 具体地, 舌纹检测模块 105采用多尺度边缘检测方法检测出舌纹的长 度和宽度, 并根据舌纹的长度和宽度描述出舌纹的大小并量化为一个描述舌纹 状况的特征值, 例如取值为 1, 2, 3, 分别表示舌纹的无、 有、 严重。 在本实施 例中, 所述多尺度边缘检测方法为现用技术中检测舌纹连续性的现有技术, 本 发明实施例不作具体赘述。
[0043] 所述舌形分析模块 106用于检测舌面大小并量化为一个描述舌体胖痩程度的特 征值; 具体地, 舌形分析模块 106利用最小二乘法将得到的舌体边缘点拟合为二 次曲线, 以及根据二次项的大小来判断舌面大小并量化为描述一个舌体胖痩程 度的特征值, 例如取值为 1, 2, 3, 分别表示舌体的痩、 中、 胖的三种程度。
[0044] 所述舌像输出模块 107用于将舌苔薄厚度、 舌苔润燥度、 舌苔腐腻度、 舌纹状 况以及舌体胖痩程度的特征值组合成一个多维的舌像特征向量, 并通过输出单 元 14将所述舌像特征向量输出, 为中医生进行中医舌诊提供参考依据。 具体地 , 舌像输出模块 107将舌苔薄厚度、 舌苔润燥度、 舌苔腐腻状况、 舌纹状况以及 舌体胖痩程度的特征值组合成一个多维的舌像特征向量, 将所述舌像特征显示 在输出单元 14的显示屏上, 或者控制所述输出单元 14的打印机打印所述舌像特 征, 为中医生了解患者的舌像情况, 从而有利于通过中医舌诊来辅助医生诊断 患者的病情。
[0045] 本发明还提供了一种中医舌诊图像处理方法, 应用于计算机 1中。 如图 3所示, 图 3是本发明中医舌诊图像处理方法优选实施例的流程图。 在本实施例中, 参考 图 1和图 2所示, 所述中医舌诊图像处理方法包括如下步骤:
[0046] 步骤 S31, 根据患者的舌诊编号从舌诊数据库中获取患者的舌头图像, 并对获 取的舌头图像进行去噪声预处理; 具体地, 舌像预处理模块 101根据患者输入的 舌诊编号从舌诊数据库 2中获取患者的舌头图像, 并对获取的舌头图像进行去噪 声预处理, 为后续的舌头图像分析与处理提供基础。 舌头图像在拍摄过程中, 可能由于光学系统、 运动等造成图像的模糊, 以及源自电路和光学等因素的噪 声而使得图像质量发生退化; 另外患者由于未能掌握正确的伸舌姿势而使得舌 体发生歪斜等等。 通过对舌图像进行预处理, 减小或去除这些不利因素对后续 舌头图像分析造成的影响。
[0047] 步骤 S32, 将舌头图像进行分割处理得到舌体图像; 具体地, 舌体提取模块 102 将舌头图像进行分割处理得到舌体图像, 包括如下步骤: 根据舌头图像的颜色 参数采用阈值分割得到舌体的初始轮廓线; 针对舌体颜色相对于人体皮肤颜色 的边界模糊特点, 利用 RGB色彩空间中的 G分量来增强舌体的弱边界得到增强型 舌头图像; 采用 snake轮廓提取方法从增强型舌头图像提取舌体轮廓得到舌体图 像。
[0048] 步骤 S33, 根据中医理论将舌体图像分成对应于人不同器官健康状况的舌体五 大部分, 并分别对舌体五大部分中舌苔的颜色特征和纹理特征进行识别; 具体 地, 苔质识别模块 103根据中医理论将舌体分成对应于人不同器官健康状况的舌 体五大部分, 如图 2所示的 A部分对应于肾, B部分对应胃和脾, C部分对应心脏 和肺, D和 E部分对应肝脏和膀胱。 苔质识别模块 103分别对舌体五大部分的舌苔 颜色 (包括苔质和舌质颜色) 进行识别, 用一个多维的向量来描述, 例如以 A区 为例, 定量描述为: 舌苔比例, 舌苔主颜色, 舌苔主颜色比例, 舌苔次颜色, 舌苔次颜色比例, 舌质比例, 舌质主颜色, 舌质主颜色比例, 舌质次颜色, 舌 质次颜色比例。 其中, 舌质颜色有淡红、 淡白、 红、 暗红、 青紫 5种类型; 而舌 苔颜色分为白、 淡黄、 黄、 灰 4种类型。
[0049] 步骤 S34, 根据舌苔的颜色特征和纹理特征分析舌苔薄厚度、 舌苔润燥度以及 舌苔腐腻度的特征值; 具体地, 舌苔分析模块 104计算舌苔颜色值占舌体颜色值 的比例 (例如, 将舌体的颜色特征值转换到 Luv空间, 计算舌苔 u值占舌体 u值的 比例) , 再利用小波变换 (例如 2D Gabor小波变换) 系数来描述舌苔区域的厚 度特征, 这样得到一个描述舌苔薄厚度的特征值; 舌苔分析模块 104将舌苔的双 色反射模型 (例如 Shafer提出的双色反射模型) 变换到色度空间得到高光象素, 从高光象素出发进行区域延伸来获得亮斑区域, 再根据亮斑区域的大小来得到 一个描述舌苔润燥度的特征值; 舌苔分析模块 104根据舌苔的测纹理特征计算舌 苔的粗糙度 (例如 Rosenfeld/Tamum粗糙度模型) , 根据该舌苔的粗糙度得到一 个描述舌苔腐腻度的特征值。 [0050] 步骤 S35, 检测舌纹的长度和宽度并量化为一个描述舌纹状况的特征值。 具体 地, 舌纹检测模块 105采用多尺度边缘检测方法进行舌纹检测舌纹的长度和宽度 , 以及根据舌纹的长度和宽度描述出舌纹的大小并量化为一个描述舌纹状况的 特征值, 例如取值为 1, 2, 3, 分别表示舌纹的无、 有、 严重。
[0051] 步骤 S36, 检测舌面大小并量化为一个描述舌体胖痩程度的特征值; 具体地, 舌形分析模块 106利用最小二乘法将得到的舌体边缘点拟合为二次曲线, 根据二 次项的大小来判断舌面大小并量化为一个描述舌体胖痩程度的特征值, 例如取 值为 1, 2, 3, 分别表示舌体的痩、 中、 胖的三种程度。
[0052] 步骤 S37, 将舌苔薄厚度、 舌苔润燥度、 舌苔腐腻度、 舌纹状况以及舌体胖痩 程度的特征值组合成一个多维的舌像特征向量, 并通过输出单元将所述舌像特 征向量输出。 具体地, 舌像输出模块 107将舌苔薄厚度、 舌苔润燥度、 舌苔腐腻 状况、 舌纹状况以及舌体胖痩程度的特征值组合成一个多维的舌像特征向量, 将所述舌像特征显示在输出单元 14的显示屏上, 或者控制所述输出单元 14的打 印机打印所述舌像特征, 为中医生了解患者的舌像情况, 从而有利于通过中医 舌诊来辅助医生诊断患者的病情。
[0053] 本发明所述中医舌诊图像处理系统及方法能够从舌诊数据库中自动筛选出患者 的舌头图像, 从舌头图像中提取出舌体部分, 对舌体的各种属性进行自动分析 和识别, 输出定量描述各种属性的一个多维的舌像特征向量, 为中医生进行中 医舌诊提供参考依据, 有利于推动中医舌诊的发展。
[0054] 以上仅为本发明的优选实施例, 并非因此限制本发明的专利范围, 凡是利用本 发明说明书及附图内容所作的等效结构或等效功能变换, 或直接或间接运用在 其他相关的技术领域, 均同理包括在本发明的专利保护范围内。
工业实用性
[0055] 相较于现有技术, 本发明所述中医舌诊图像处理系统及方法采用上述技术方案 , 达到了如下技术效果: 能够从舌诊数据库中自动筛选出患者的舌头图像, 从 舌图像中自动提取出舌体部分, 对舌体的各种属性进行自动分析和识别, 输出 定量描述各种属性的一个多维的舌像特征向量, 为中医生进行中医舌诊提供参 考依据, 有利于推动中医舌诊的发展。

Claims

权利要求书
[权利要求 1] 一种中医舌诊图像处理系统, 应用于计算机中, 该计算机通过数据库 链接连接至舌诊数据库, 其特征在于, 所述中医舌诊图像处理系统包 括: 舌像预处理模块, 用于根据患者的舌诊编号从所述舌诊数据库中 获取患者的舌头图像, 并对获取的舌头图像进行去噪声预处理; 舌体 提取模块, 用于将预处理后的舌头图像进行分割处理得到舌体图像; 苔质识别模块, 用于根据中医理论将舌体图像分成对应于人不同器官 健康状况的舌体五大部分, 以及分别对所述舌体五大部分中舌苔的颜 色特征和纹理特征进行识别; 舌苔分析模块, 用于根据所述舌苔的颜 色特征和纹理特征分析舌苔薄厚度、 舌苔润燥度以及舌苔腐腻度的特 征值; 舌纹检测模块, 用于检测舌纹的长度和宽度并量化为一个描述 舌纹状况的特征值; 舌形分析模块, 用于通过检测舌面大小并量化为 一个描述舌体胖痩程度的特征值; 舌像输出模块, 用于将舌苔薄厚度 、 舌苔润燥度、 舌苔腐腻度、 舌纹状况以及舌体胖痩程度的特征值组 合成一个多维的舌像特征向量, 以及通过所述计算机的输出单元将所 述舌像特征向量输出。
[权利要求 2] 如权利要求 1所述的中医舌诊图像处理系统, 其特征在于, 所述舌苔 分析模块还用于: 计算舌苔颜色值占舌体颜色值的比例, 利用小波变 换系数得到一个描述舌苔薄厚度的特征值; 将舌苔的双色反射模型变 换到色度空间得到高光象素, 从高光象素出发进行区域延伸来获得亮 斑区域, 根据亮斑区域的大小来得到一个描述舌苔润燥度的特征值; 根据舌苔的测纹理特征计算舌苔的粗糙度, 根据该舌苔的粗糙度得到 一个描述舌苔腐腻度的特征值。
[权利要求 3] 如权利要求 1所述的中医舌诊图像处理系统, 其特征在于, 所述舌体 提取模块还用于根据舌头图像的颜色参数采用阈值分割得到舌体的初 始轮廓线, 针对舌体颜色相对于人体皮肤颜色的边界模糊特点利用 R GB色彩空间中的 G分量来增强舌体的弱边界得到增强型舌头图像, 以及从增强型舌头图像提取舌体轮廓得到所述舌体图像。
[权利要求 4] 如权利要求 1所述的中医舌诊图像处理系统, 其特征在于, 所述舌纹 检测模块还用于采用多尺度边缘检测方法检测出舌纹的长度和宽度, 根据舌纹的长度和宽度描述出舌纹的大小并量化为所述描述舌纹状况 的特征值。
[权利要求 5] 如权利要求 1所述的中医舌诊图像处理系统, 其特征在于, 所述舌形 分析模块还用于利用最小二乘法将得到的舌体边缘点拟合为二次曲线 , 以及根据二次项的大小来判断舌面大小并量化为所述描述舌体胖痩 程度的特征值。
[权利要求 6] —种中医舌诊图像处理方法, 应用于计算机中, 该计算机通过数据库 链接连接至舌诊数据库, 其特征在于, 该方法包括如下步骤: 根据患 者的舌诊编号从所述舌诊数据库中获取患者的舌头图像, 并对获取的 舌头图像进行去噪声预处理; 将预处理后的舌头图像进行分割处理得 到舌体图像; 根据中医理论将舌体图像分成对应于人不同器官健康状 况的舌体五大部分, 以及分别对所述舌体五大部分中舌苔的颜色特征 和纹理特征进行识别; 根据所述舌苔的颜色特征和纹理特征分析舌苔 薄厚度、 舌苔润燥度以及舌苔腐腻度的特征值; 检测舌纹的长度和宽 度并量化为一个描述舌纹状况的特征值; 检测舌面大小并量化为一个 描述舌体胖痩程度的特征值; 将所述舌苔薄厚度、 舌苔润燥度、 舌苔 腐腻度、 舌纹状况以及舌体胖痩程度的特征值组合成一个多维的舌像 特征向量, 以及通过所述计算机的输出单元将所述舌像特征向量输出
[权利要求 7] 如权利要求 6所述的中医舌诊图像处理方法, 其特征在于, 所述根据 舌苔的颜色特征和纹理特征分析舌苔薄厚度、 舌苔润燥度以及舌苔腐 腻度的特征值的步骤包括步骤: 计算舌苔颜色值占舌体颜色值的比例 , 利用小波变换系数得到一个描述舌苔薄厚度的特征值; 将舌苔的双 色反射模型变换到色度空间得到高光象素, 从高光象素出发进行区域 延伸来获得亮斑区域, 根据亮斑区域的大小来得到一个描述舌苔润燥 度的特征值; 根据舌苔的测纹理特征计算舌苔的粗糙度, 根据该舌苔 的粗糙度得到一个描述舌苔腐腻度的特征值。
[权利要求 8] 如权利要求 6所述的中医舌诊图像处理方法, 其特征在于, 所述将舌 头图像进行分割处理得到舌体图像的步骤包括如下步骤: 根据舌头图 像的颜色参数采用阈值分割得到舌体的初始轮廓线; 针对舌体颜色相 对于人体皮肤颜色的边界模糊特点利用 RGB色彩空间中的 G分量来增 强舌体的弱边界得到增强型舌头图像; 从增强型舌头图像提取舌体轮 廓得到所述舌体图像。
[权利要求 9] 如权利要求 6所述的中医舌诊图像处理方法, 其特征在于, 所述检测 舌纹的长度和宽度并量化为一个描述舌纹状况的特征值的步骤包括步 骤: 采用多尺度边缘检测方法检测出舌纹的长度和宽度; 根据舌纹的 长度和宽度描述出舌纹的大小并量化为所述描述舌纹状况的特征值。
[权利要求 10] 如权利要求 6所述的中医舌诊图像处理方法, 其特征在于, 所述检测 舌面大小并量化为一个描述舌体胖痩程度的特征值的步骤包括步骤: 利用最小二乘法将得到的舌体边缘点拟合为二次曲线; 根据二次项的 大小来判断舌面大小并量化为所述描述舌体胖痩程度的特征值。
PCT/CN2017/082251 2016-11-29 2017-04-27 中医舌诊图像处理系统及方法 WO2018098987A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201611075072.8A CN106725310A (zh) 2016-11-29 2016-11-29 中医舌诊图像处理系统及方法
CN201611075072.8 2016-11-29

Publications (1)

Publication Number Publication Date
WO2018098987A1 true WO2018098987A1 (zh) 2018-06-07

Family

ID=58898525

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/082251 WO2018098987A1 (zh) 2016-11-29 2017-04-27 中医舌诊图像处理系统及方法

Country Status (2)

Country Link
CN (1) CN106725310A (zh)
WO (1) WO2018098987A1 (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107977958A (zh) * 2017-11-21 2018-05-01 郑州云海信息技术有限公司 一种图像诊断方法和装置
CN109259730A (zh) * 2018-10-09 2019-01-25 广东数相智能科技有限公司 一种基于舌诊的预警分析方法和存储介质
CN109350011A (zh) * 2018-11-19 2019-02-19 陶磊 一种基于移动智能终端的舌诊装置
CN109785311B (zh) * 2019-01-14 2021-06-04 深圳和而泰数据资源与云技术有限公司 一种疾病诊断装置、电子设备及存储介质
CN109745012A (zh) * 2019-02-26 2019-05-14 北京小成素问信息技术有限公司 便携式舌诊镜、舌体图像获取方法和舌诊系统
CN110751625A (zh) * 2019-09-12 2020-02-04 北京中科芯健医疗科技有限公司 舌象采集装置及舌象采集方法
CN110916628A (zh) * 2019-12-31 2020-03-27 天津慧医谷科技有限公司 一种中医望闻问切四诊信息采集诊断装置
CN111696073A (zh) * 2020-04-29 2020-09-22 陕西尚善优选食品科技有限公司 一种基于深度学习的舌面诊检测方法
CN112806964A (zh) * 2021-01-28 2021-05-18 清华大学 舌象分析方法和系统
CN113160973A (zh) * 2021-04-15 2021-07-23 武汉未康未病医学有限公司 一种看舌头App气血津液数学模型
CN113239790A (zh) * 2021-05-11 2021-08-10 上海道生医疗科技有限公司 舌裂纹的特征识别及长度度量方法及系统
CN115147372B (zh) * 2022-07-04 2024-05-03 海南榕树家信息科技有限公司 一种基于医学图像分割的中医舌像智能辨治方法及系统
CN116013504A (zh) * 2022-12-29 2023-04-25 上海市中医医院 一种基于图像检索分析的中医舌诊装置
CN117853345B (zh) * 2024-03-07 2024-05-24 吉林大学 一种用于中医舌诊舌像成像的图像优化方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1803087A (zh) * 2006-01-19 2006-07-19 上海交通大学 舌像颜色自动识别方法
CN1830380A (zh) * 2006-04-20 2006-09-13 上海交通大学 中医舌像分析与诊断系统
CN1931087A (zh) * 2006-10-11 2007-03-21 哈尔滨工业大学 舌象纹理自动分析方法
CN1973757A (zh) * 2006-10-11 2007-06-06 哈尔滨工业大学 基于舌象特征的病证计算机分析系统
CN101919685A (zh) * 2010-04-30 2010-12-22 广州中医药大学 一种舌诊智能控制及诊断系统及其诊断方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1803087A (zh) * 2006-01-19 2006-07-19 上海交通大学 舌像颜色自动识别方法
CN1830380A (zh) * 2006-04-20 2006-09-13 上海交通大学 中医舌像分析与诊断系统
CN1931087A (zh) * 2006-10-11 2007-03-21 哈尔滨工业大学 舌象纹理自动分析方法
CN1973757A (zh) * 2006-10-11 2007-06-06 哈尔滨工业大学 基于舌象特征的病证计算机分析系统
CN101919685A (zh) * 2010-04-30 2010-12-22 广州中医药大学 一种舌诊智能控制及诊断系统及其诊断方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHOU YUE ET AL.: "Feature analysis method of tongue image by chinese medical diagnosis based on image processing", INFRARED AND LASER ENGINEERING, vol. 31, no. 6, 22 December 2002 (2002-12-22), XP055227821 *

Also Published As

Publication number Publication date
CN106725310A (zh) 2017-05-31

Similar Documents

Publication Publication Date Title
WO2018098987A1 (zh) 中医舌诊图像处理系统及方法
WO2018098986A1 (zh) 中医舌像自动检测系统及方法
Javidi et al. Vessel segmentation and microaneurysm detection using discriminative dictionary learning and sparse representation
Zortea et al. A simple weighted thresholding method for the segmentation of pigmented skin lesions in macroscopic images
US8958625B1 (en) Spiculated malignant mass detection and classification in a radiographic image
US20210174505A1 (en) Method and system for imaging and analysis of anatomical features
JP6005663B2 (ja) 血管画像における動静脈比の自動測定
Dai et al. Optic disc segmentation based on variational model with multiple energies
Ramezani et al. Automatic detection of malignant melanoma using macroscopic images
US20160228008A1 (en) Image diagnosis device for photographing breast by using matching of tactile image and near-infrared image and method for aquiring breast tissue image
Marques et al. On the role of texture and color in the classification of dermoscopy images
TWI719587B (zh) 眼底圖像定量分析的前置處理方法及存放裝置
CN114820603B (zh) 基于ai舌诊图像处理的智能健康管理方法及相关装置
Chen et al. Hybrid facial image feature extraction and recognition for non-invasive chronic fatigue syndrome diagnosis
KR20200108686A (ko) 딥러닝 알고리즘을 이용한 근감소증 분석 프로그램 및 애플리케이션
Liu et al. Automatic detection and quantification of brain midline shift using anatomical marker model
CN113781488A (zh) 舌象图像的分割方法、装置及介质
US11980491B2 (en) Automatic recognition method for measurement point in cephalo image
Li et al. Automated image analysis of uterine cervical images
Chen et al. Application of artificial intelligence in tongue diagnosis of traditional Chinese medicine: a review
CN111223117A (zh) 舌面图像分割装置、方法及计算机存储介质
Pasero et al. Application of an automatic ulcer segmentation algorithm
CN117274278A (zh) 基于模拟感受野的视网膜图像病灶部位分割方法及系统
TWI663960B (zh) 足畸形檢測模型、足畸形檢測系統及足畸形檢測方法
Heravi et al. Low price foot pressure distribution screening technique: Optical podoscope with accurate foot print segmentation using hidden Markov random field model

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17876588

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17876588

Country of ref document: EP

Kind code of ref document: A1