JP2017148386A - Tongue image processing device, tongue image processing method and tongue image processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a tongue image processing device capable of individually detecting each tooth mark at high speed more accurately while reducing a processing load without being affected by an environmental factor during imaging of a tongue image.SOLUTION: A tongue image processing device comprises: an interface 1 for acquiring a tongue image obtained by imaging a tongue; an operation part 4 for designating a tooth mark detection area which is used for detecting a tooth mark which occurred on the tongue, in the tongue image; a tongue diagnosis area extraction part 21 for extracting a tongue diagnosis area using an area determination mask which is used for dividing an area into a tongue diagnosis area corresponding to the tongue in the designated tooth mark detection area and a background in the tooth mark detection area other than the tongue, which corresponds to a dimension of the tooth mark detection area, extracting a contour of the tongue from the tongue diagnosis area, and detecting a convex hull which is circumscribed to the contour; a tooth mark state detection part 22 for extracting a tooth mark area corresponding to one or each of plural tooth marks, based on the contour and convex hull; and an output control part 23 for outputting tooth mark area information indicating the tooth mark area.SELECTED DRAWING: Figure 1

Description

  The present invention belongs to the technical field of a tongue image processing apparatus, a tongue image processing method, and a tongue image processing program. More specifically, the present invention belongs to the technical field of a tongue image processing device, a tongue image processing method, and a program for the tongue image processing device for processing a tongue image obtained by imaging the tongue.

  In view of so-called health orientation in recent years, development of a method capable of objectively detecting or recognizing a person's health condition from the outside is underway. And as a prior art relevant to such a method, there exists the method described in the following patent document 1. FIG.

  In the method described in Patent Literature 1 below, the overall health level is determined based on image information of a part of the body such as the tongue. More specifically, in the method described in Patent Document 1, the tongue color, the tongue shape, the tooth mark on the peripheral edge of the tongue, the cleft on the tongue surface, and the moss on the tongue surface are obtained from the image of the tongue photographed with a digital still camera. The mathematical Mahalanobis distance between nine types of feature amount information such as the color of the color and a preset healthy body reference sample is obtained, and the value is used as a scale for obtaining the health level. Here, the above-mentioned tooth mark refers to the tooth formed on the side or tip of the tongue due to the fact that the side of the tongue is normally in contact with the tooth or the tongue is accidentally bitten by the tooth. It means a scar.

  At this time, the tooth trace as the feature amount plays an important role in the determination of the health level, like the other feature amounts. More specifically regarding the detection of such a tooth trace, in Patent Document 1, in the tongue image corresponding to the tongue, the distance from the center of gravity of the tongue to its contour is used to calculate the sum of the differences between two adjacent distances. Therefore, if the sum of the differences is large, it is determined that there is a tooth mark and the degree thereof is large, and if the sum of the differences is small, it is determined that there is almost no tooth mark and the degree thereof is small. Yes.

Japanese Patent No. 4487535

  However, according to the method of detecting a tooth mark in the method described in Patent Document 1, two indexes, that is, “the presence / absence of a tooth mark” and “the degree when there is a tooth mark (the degree of whether there are many or few tooth marks). ) ”Only. Therefore, in this case, there has been a problem that detailed evaluation such as the number, position, size, and color of the tooth mark region, which is important information in calculating the health degree using the tongue image, cannot be performed.

  Therefore, the present invention has been made in view of the above-mentioned problems, and an example of the problem is that it is more accurate, high-speed, and processing load can be reduced without being affected by environmental factors at the time of capturing a tongue image. A tongue image processing device, a tongue image processing method, and a program for the tongue image processing device capable of extracting a tongue region including a tooth root from a tongue image and separately detecting the respective tooth traces are provided. There is to do.

  In order to solve the above problems, the invention described in claim 1 is characterized in that tongue image information acquisition means such as an interface for acquiring tongue image information corresponding to a tongue image obtained by imaging the tongue, Designating means such as an operation unit used for designating a tooth mark detection area used for detection of the generated tooth mark in the tongue image, and a tongue area corresponding to the tongue in the designated tooth mark detection area; Area determination mask information corresponding to an area determination mask for separating a background area corresponding to the background in the tooth mark detection area other than the tongue and corresponding to the dimension of the tooth mark detection area Mask information acquisition means, such as an interface for acquiring information, and tongue region extraction means, such as a tongue examination region extraction unit, for extracting the tongue region from the tooth mark detection region using the region determination mask.

  In order to solve the above-mentioned problem, the invention according to claim 7 corresponds to the tongue image in a tongue image processing method executed in a tongue image processing apparatus for processing a tongue image obtained by imaging the tongue. A tongue image information acquisition step for acquiring the tongue image information, a designation step for designating in the tongue image a tooth mark detection region used for detection of a tooth mark formed on the tongue, and within the designated tooth mark detection region An area determination mask for separating a tongue area corresponding to the tongue and a background area corresponding to the background in the tooth mark detection area other than the tongue, and an area corresponding to the size of the tooth mark detection area A mask information acquisition step of acquiring region determination mask information corresponding to the determination mask, and a tongue region extraction step of extracting the tongue region from the tooth mark detection region using the region determination mask.

  In order to solve the above problems, the invention according to claim 8 is a tongue image processing apparatus for processing a tongue image obtained by imaging the tongue, and is formed on the tongue of the tongue image. Tongue image for acquiring tongue image information corresponding to the tongue image from a computer included in a tongue image processing apparatus having a designation unit such as an operation unit used for designating a tooth mark detection region used for detection of a tooth mark An area determination mask for separating information acquisition means, a tongue area corresponding to the tongue in the designated tooth mark detection area, and a background area corresponding to the background in the tooth mark detection area other than the tongue A mask information acquisition means for acquiring area determination mask information corresponding to an area determination mask corresponding to the size of the tooth mark detection area, and extracting the tongue area from the tooth mark detection area using the area determination mask; Tongue region extraction Stage, to function as.

  According to the invention described in any one of claims 1, 7, and 8, the tongue region is extracted using the region determination mask corresponding to the size of the tooth mark detection region. Therefore, since the tongue region is extracted using the region determination mask corresponding to the size of the tooth mark detection region, compared with the case where the tongue region detection process is performed using all of the tooth mark detection region, the tongue image is captured. It is possible to detect the tongue region more accurately, at a high speed and while reducing the processing load without being affected by environmental factors.

  In order to solve the above-mentioned problem, according to a second aspect of the present invention, in the tongue image processing apparatus according to the first aspect, the region determination mask has a semi-elliptical shape corresponding to the dimension and the tongue. A first mask corresponding to its own area, a second mask corresponding to the dimension, larger than the first mask and corresponding to the background area, an outer edge of the first mask corresponding to the dimension, and the first mask A third mask having an outer edge in a region between the outer edges of the two masks, and is superposed so that the respective center lines in the longitudinal direction coincide with each other. The tongue region is extracted with respect to a region of the tongue image corresponding to a region between the outer edge of the mask and the outer edge of the first mask.

  According to the invention described in claim 2, in addition to the operation of the invention described in claim 1, the first mask, the second mask, and the third mask are overlapped so that the center lines in the longitudinal direction coincide with each other. Thus, a region determination mask is configured, and the tongue region is extracted for the region of the tongue image corresponding to the region between the outer edge of the third mask and the outer edge of the first mask. Therefore, each tooth trace region can be detected separately while at a higher speed and reducing the processing load.

  In order to solve the above-mentioned problem, the invention according to claim 3 is the tongue image processing apparatus according to claim 1 or 2, wherein the tongue region extracting means uses the designated tooth mark detection region. Reduced tongue region extraction means such as a processing unit for extracting a reduced tongue region from the reduced tooth mark detection region using a tongue region extraction mask corresponding to the size of the reduced reduced tooth mark detection region, and the extracted reduced tongue region Magnifying means such as a processing unit for enlarging using a magnification corresponding to the reduced tongue region, and mask generating means such as a processing unit for generating a tongue region mask corresponding to the enlarged reduced tongue region, The tongue region extracting means is configured to extract the tongue region from the tooth mark detection region using the generated tongue region mask.

  According to the invention described in claim 3, in addition to the action of the invention described in claim 1 or 2, the tongue region corresponding to the size of the reduced tooth mark detection region obtained by reducing the specified tooth mark detection region. A reduced tongue region is extracted using the extraction mask, a tongue region mask corresponding to a region obtained by enlarging the reduced tongue region is generated, and the tongue region is extracted from the tooth mark detection region using the tongue region mask. Therefore, even when the original tongue image has an excessive amount of information as the tongue image information, each tooth mark region is extracted at a high speed while reducing the processing load by reducing the tooth mark detection region and extracting the tongue region. Can be detected.

  In order to solve the above-mentioned problem, the invention according to claim 4 is the tongue image processing apparatus according to any one of claims 1 to 3, wherein the tongue is extracted from the extracted tongue region. Contour extraction means such as a tongue examination region extraction unit for extracting a contour, convex hull detection means such as a tongue examination region extraction unit for detecting a convex hull circumscribing the contour based on the extracted contour, and the extraction Based on the contour and the detected convex hull, a tooth trace form detection unit that extracts a tooth trace area corresponding to each of one or a plurality of the tooth traces while identifying each other from the other tooth trace areas It further comprises tooth trace area extraction means and output means such as a display for outputting the tooth trace area information indicating the detected tooth trace area.

  According to the invention described in claim 4, in addition to the action of the invention described in any one of claims 1 to 3, a convex hull corresponding to the contour of the tongue extracted from the extracted tongue region is provided. Are detected while mutually identifying each tooth trace region based on the contour and the convex hull, and tooth mark region information indicating the detection result is output. Therefore, compared to the case where tongue region detection processing is performed using the entire tooth mark detection region, the processing load is reduced more accurately and quickly without being affected by environmental factors at the time of tongue image capturing. Each tooth trace region can be detected separately.

  In order to solve the above-mentioned problem, the invention according to claim 5 is the tongue image processing device according to claim 4, wherein the tooth trace region extracting means includes both ends of the tooth trace in the extracted contour and the A region of the tongue image corresponding to a semicircle centered at the midpoint of the line segment connecting each contact point with the detected convex hull and having the line segment as a diameter, and including the tooth mark Are extracted as the tooth trace regions while being distinguished from the other tooth trace regions.

  According to the invention described in claim 5, in addition to the action of the invention described in claim 4, the center of the line segment connecting the contact points between the both ends of the tooth trace and the convex hull in the contour is the center, and Tongue image areas corresponding to semicircles containing tooth marks with the diameter of the line segment are detected as tooth mark areas while distinguishing from other tooth mark areas, so the tooth mark area is detected with simple processing. can do.

  In order to solve the above-described problem, the invention according to claim 6 is the tongue image processing measure according to any one of claims 1 to 5, wherein the output means is configured to detect the detected tooth mark. A color detection unit such as a tooth trace form detection unit for detecting a color of the region, and a storage unit such as a recording unit for storing threshold information indicating a threshold for classifying the color into a preset color level. The output means outputs color information indicating the color level corresponding to the detected color based on the stored threshold information and color information indicating the detected color. It is comprised so that it may output for every said tooth mark area | region as area | region information.

  According to invention of Claim 6, in addition to the effect | action of the invention as described in any one of Claims 1-5, the color of a tooth trace area | region is detected, and the said color is related with a threshold value. Since the color information indicating which color level is supported is output as tooth mark area information for each tooth mark area, the color of the tooth mark area can be determined at high speed and for each tooth mark area by a simple method. .

  According to the present invention, the tongue region is extracted using the region determination mask corresponding to the size of the tooth mark detection region, the convex hull corresponding to the tongue region extracted from the tongue region is detected, and the contour and the convexity are detected. Based on the capsule, each tooth mark area is detected while being identified with each other, and tooth mark area information indicating the detection result is output.

  Therefore, since the tongue region is extracted using the region determination mask corresponding to the size of the tooth mark detection region, the tongue image is captured when compared with the case where the tongue region detection process is performed using all the tooth mark detection regions. Each tooth mark region can be detected separately without being affected by environmental factors or the like, while reducing the processing load more accurately and at high speed.

It is a block diagram showing a schematic structure of a tooth mark detection system concerning a 1st embodiment. It is a figure which illustrates the area | region determination mask which concerns on 1st Embodiment, (a) is a figure which illustrates designation | designated of the tongue examination area | region which concerns on 1st Embodiment, (b) is 1st which concerns on 1st Embodiment. It is a figure which shows a mask, (c) is a figure which shows the 2nd mask which concerns on 1st Embodiment, (d) is a figure which shows the 3rd-1 mask which concerns on 1st Embodiment, (e) FIG. 5 is a diagram showing a 3-2 mask according to the first embodiment, (f) is a diagram showing a configuration of a region determination mask according to the first embodiment, and (g) is a region according to the first embodiment. It is a figure which shows a determination mask. It is a flowchart which shows the tooth-dent detection process which concerns on 1st Embodiment. It is figure (I) explaining the tooth mark detection process which concerns on 1st Embodiment, (a) is a figure which illustrates designation | designated of the tongue examination area concerning 1st Embodiment, (b) is the said designation | designated It is a figure which illustrates a tongue examination area | region, (c) is a figure which illustrates the extraction result of the tongue examination area | region which concerns on 1st Embodiment, (d) illustrates the extraction result of the tongue outline which concerns on 1st Embodiment. (E) is a figure which illustrates the extraction result of the approximate curve which concerns on 1st Embodiment. It is a figure (II) explaining a tooth mark detection processing concerning a 1st embodiment, (a) is a figure which illustrates a detection result of a convex hull concerning a 1st embodiment, and (b) is a 1st embodiment. It is a figure which illustrates the detection result of the tooth trace concerning. It is a figure which shows the detail of the tongue examination area | region extraction process which concerns on 1st Embodiment, (a) is a flowchart which shows the said detail, (b) illustrates designation | designated of the tongue examination area | region in the said tongue examination area extraction process (C) is a diagram illustrating the designated tongue examination region, (d) is a diagram showing an area determination mask used for the tongue examination region extraction processing, and (e) is a diagram showing the tongue It is a figure which illustrates the extraction result of a medical examination area. It is a figure (I) which shows details of form detection processing concerning a 1st embodiment, (a) is a flow chart which shows the details, and (b) is a figure which illustrates tooth mark detection in the form detection processing concerned (C) is a figure which illustrates the setting of the tooth mark area | region used for the said form detection process, (d) is a figure which illustrates the color discrimination | determination image in the said form detection process, (e) is the said form detection It is a figure which illustrates the tooth mark area | region of the whole tongue in a process. It is a figure (II) which shows the detail of the form detection process which concerns on 1st Embodiment, (a) is a figure which illustrates only the tooth mark area | region in the said form detection process, (b) is the tooth | gear in the said form detection process It is an example of the enlarged view of a trace area | region. It is a figure which shows the detail of the tongue examination area | region extraction process which concerns on 2nd Embodiment, (a) is a flowchart which shows the said detail, (b) illustrates designation | designated of the tongue examination area | region in the said tongue examination area extraction process (C) is a diagram illustrating the designated tongue examination region, (d) is a diagram showing an area determination mask used for the tongue examination region extraction processing, and (e) is a diagram showing the tongue It is a figure which illustrates the temporary extraction result of a medical examination area | region, (f) is a figure which illustrates the tongue examination area mask etc. which concern on 2nd Embodiment, (g) is a tongue examination area which concerns on 2nd Embodiment. It is a figure which illustrates the final extraction result of.

  Next, modes for carrying out the present invention will be described based on the drawings. Each embodiment described below detects a tooth mark on the tongue from a tongue image corresponding to the tongue when determining the health level of the subject based on the state of the tongue of the subject, and the detection result Is an embodiment when the present invention is applied to a tooth mark detection system that presents the subject to the subject or the doctor in charge thereof.

(I) First Embodiment First, a first embodiment according to the present invention will be described with reference to FIGS. FIG. 1 is a block diagram illustrating a schematic configuration of the tooth mark detection system according to the first embodiment, FIG. 2 is a diagram illustrating an area determination mask according to the first embodiment, and FIG. 3 is a diagram illustrating the first embodiment. It is a flowchart which shows the tooth mark detection process which concerns on a form, and FIG.4 and FIG.5 is a figure explaining the tooth mark detection process which concerns on 1st Embodiment, respectively. FIG. 6 is a diagram showing details of the tongue examination region extraction process according to the first embodiment, and FIGS. 7 and 8 are diagrams showing details of the form detection process according to the first embodiment.

  As shown in FIG. 1, the tooth mark detection system SS according to the first embodiment includes, for example, a camera C composed of a digital still camera or the like similar to a conventional one, and a tooth mark according to the first embodiment connected to the camera C. And a detection device S. Based on the tongue image obtained by imaging the tongue of the subject with the camera C, the tooth mark on the tongue is detected by the tooth mark detection device S and the detection result is presented to the subject or the doctor in charge thereof. To do. In the following description, a tongue image obtained by imaging the subject's tongue is simply referred to as a “tongue image”.

  As shown in FIG. 1, the tooth mark detection apparatus S according to the first embodiment includes an interface 1, a processing unit 2 including a CPU, a ROM (Read Only Memory), a RAM (Random Access Memory), and the like, and a liquid crystal display. And the like, a display unit 3 including a keyboard and a mouse, and a recording unit 5 including a hard disk, for example. At this time, the tooth mark detection device S is specifically configured by, for example, a personal computer. The interface 1 corresponds to an example of a “tongue image information acquisition unit” and an example of a “mask information acquisition unit” according to the present invention, respectively, and the processing unit 2 includes a “reduced tongue region extraction unit” according to the present invention. It corresponds to an example, an example of “enlargement means”, and an example of “mask generation means”, respectively. Furthermore, the recording unit 5 corresponds to an example of the “storage unit” according to the present invention, the display 3 corresponds to an example of the “output unit” according to the present invention, and the operation unit 4 corresponds to the “designating unit” according to the present invention. It corresponds to an example.

  On the other hand, the processing unit 2 includes an image registration unit 20, a tongue examination region extraction unit 21, a tooth trace form detection unit 22, and an output control unit 23. At this time, the image registration unit 20, the tongue examination region extraction unit 21, the tooth trace form detection unit 22, and the output control unit 23 may be configured by, for example, a hardware logic circuit that configures the processing unit 2, or The program corresponding to the flowchart showing the tooth mark detection process according to the first embodiment to be described later may be executed by software by the CPU of the processing unit 2 reading and executing the program. At this time, the program may be recorded in advance in the ROM or the recording unit 5 constituting the processing unit 2, or may be acquired through a network such as the Internet as necessary.

  The tongue examination region extraction unit 21 corresponds to an example of a “tongue region extraction unit”, an example of a “contour extraction unit”, and an example of a “convex hull detection unit” according to the present invention. Reference numeral 22 corresponds to an example of a “dental mark region extracting unit” and an example of a “color detecting unit” according to the present invention.

  In the configuration described above, the interface 1 of the tooth mark detection device S acquires the tongue image information corresponding to the tongue image from the camera C when the subject's tongue is imaged by the camera C, as will be described later. The image is output to the image registration unit 20. Here, in the tongue image, in addition to the tongue itself having a tooth mark detected by the tooth mark detection process according to the first embodiment, when the subject takes out the tongue, The teeth, the upper lip, and the lower lip are included. Next, the image registration unit 20 records (registers) the output tongue image information in the recording unit 5 in association with identification information for identifying a subject or a doctor in charge from another person, for example. In the following description, the subject or the doctor in charge thereof is simply referred to as “subject”.

  On the other hand, the recording unit 5 includes an image recording unit 50, a mask database 51, and a color determination threshold database 52. At this time, the image recording unit 50 records the tongue image information in association with the identification information, the mask database 51 records one or a plurality of data corresponding to the region determination mask according to the first embodiment, and the color determination threshold database. 52 records one or more color determination threshold values according to the first embodiment. In FIG. 1, the mask database is represented as “mask DB”, and the color determination threshold database is represented as “color determination threshold DB”. The area determination mask according to the first embodiment and the color determination threshold according to the first embodiment will be described later in detail.

  Next, the tongue examination area extraction unit 21 extracts the tongue examination area according to the first embodiment from the tongue image corresponding to the tongue image information using the area determination mask, and corresponds to the extracted tongue examination area. Tongue area information to be generated is output to the tooth trace form detection unit 22. Here, the tongue region is a region including a portion of the tongue having a tooth mark, excluding the upper teeth and the upper lip, among the tongue images corresponding to the tongue image information, according to the first embodiment. This refers to the region that is the detection target of the tooth trace. Then, the tooth mark form detection unit 22 detects the form of the tooth marks including the number and color of one or a plurality of tooth marks on the tongue corresponding to the tongue examination area information, and outputs the detection result. To the unit 23. This form detection will also be described in detail later. Thereafter, the output control unit 23 generates a display screen for presenting (displaying) the detection result to a subject or the like based on the shape detection result of each tooth mark output from the tooth mark shape detection unit 22, and the display Information corresponding to the screen is displayed on the display 3, and the detection result is displayed.

  Next, the tooth mark detection process according to the first embodiment will be specifically described with reference to FIGS.

  First, the region determination mask used for extracting the tongue examination region according to the first embodiment will be described in detail with reference to FIG. As described above, the region determination mask is used when the tongue examination region corresponding to the tongue is extracted from the tongue image including the image corresponding to the upper teeth and the upper lip. That is, in the extraction of the tongue examination area according to the first embodiment, as illustrated in FIG. 2A, the tongue examination area corresponding to the tongue T should be cut out from the tongue image TG in which the upper teeth and the upper lip are simultaneously reflected. A rectangular extraction frame FM having a width W / length H corresponding to the size of the tongue T in the tongue image TG is manually designated by the doctor in charge.

  At this time, even if the extraction frame FM is used to cut out the portion corresponding to the tongue T from the tongue image TG, the background of the tongue T is present inside the extraction frame FM as shown in FIG. Images of both sides of the chin, neck, and collar of the jacket are included. On the other hand, in order to exclude the background of the tongue T, it is difficult to designate the extraction frame FM as a shape that exactly follows the complicated outer shape of the tongue T. Therefore, in the extraction of the tongue examination area according to the first embodiment, an area determination mask having a shape corresponding to the dimension of the specified rectangular extraction frame FM is prepared in advance, and data corresponding to the mask is recorded in the mask database 51. deep. This region determination mask is for distinguishing and determining a tongue examination region corresponding to the tongue T itself and a region corresponding to the other background, etc. on the tongue image TG illustrated in FIG. This is a mask for image processing. The region determination mask has a shape corresponding to the width W and length H of the designated extraction frame FM, and masks the region of the tongue image TG other than the region along the outer shape of the tongue T (ie, , To be excluded from the target of the tooth trace detection process according to the first embodiment). Here, since the shape of the area determination mask is determined by the size (width W × length H) of the extraction frame FM, for example, if the size of the extraction frame FM is accidentally the same between another subject and another subject, Even if the subjects are different, extraction of the tongue examination region is performed using the same region determination mask.

  More specifically, the area determination mask MK according to the first embodiment includes a semi-elliptical mask (width Wa × length Ha) illustrated by a solid line in FIG. 2B, and two points in FIG. A rectangular (horizontal Wb × vertical Hb) mask M2 exemplified by a chain line, a semi-elliptical mask (width Wc × length Hc) exemplified by a broken line in FIG. 2D, and FIG. A rectangular (horizontal Wd × longitudinal Hd) mask M3-2 exemplified by a one-dot chain line is superimposed on each other so that the center lines thereof coincide with each other as shown in FIG. 2 (f). . At this time, as illustrated in FIG. 2, the relationship between the width Wa, the width Wb, the width Wc, and the width Wd is width Wa <width Wd <width Wb <width Wc, and the length Ha, length Hb, length Hc, and length Wc. The relationship of Hd is length Ha <vertical Hd <vertical Hb and length Hc. The relationship between these dimensions and the width W and length H of the extraction frame FM is preset experimentally or empirically, and can be preset as follows as an example. In addition, you may vary each constant shown below for every dimension of the extraction frame FM.

Width Wa = Width W × 0.40, Length Ha = Length H × 0.85
Width Wb = width W × 0.96, length Hb = length H × 0.96
Width Wc = width W × 1.04, length Hc = length H × 1.1
Horizontal Wd = Width W × 0.65, Vertical Hd = Length H × 0.65
As illustrated in FIG. 2G, the area determination mask MK is defined by a mask area MK1 defined by the mask M1, a mask M1, a part of the mask M3-1, and a part of the mask M3-2. And a mask region MK3 defined by a part of the mask M3-1, a part of the mask M3-2, and the mask M2. In FIG. 2G, the mask area MK2 is indicated by cross hatching, and the mask area MK3 is indicated by upper right hatching. Data indicating the region determination mask MK having the above configuration is recorded in the recording unit 5 in a nonvolatile manner as the mask database 51. At this time, the mask area MK1 corresponds to the area of the tongue T itself in the tongue image TG cut out by the extraction frame FM, and the mask area MK3 corresponds to the background area in the tongue image TG cut out by the extraction frame FM, This corresponds to an area in which it cannot be determined whether the mask area MK2 is the tongue T itself or the background. In the extraction of the tongue examination region according to the first embodiment, the mask region MK1 is determined as the region of the tongue T itself, the mask region MK3 is determined as the background region, and the portion of the tongue image TG corresponding to the mask region MK2 Extraction of the tongue examination area, which will be described later, is performed only on the processing target.

  Next, the color determination threshold according to the first embodiment will be described. The color determination threshold value is a threshold value used when the color (redness) of each detected one or a plurality of tooth marks is classified into a plurality of color levels by the tooth mark shape detection unit 22. The color determination threshold is set in advance for each color level based on, for example, medical knowledge. More specifically, for example, when each color of each tooth mark is divided into five color levels, four different color determination threshold values are set as the first color determination threshold value to the fourth color determination threshold value. . As a result of detecting the color (redness) of each tooth mark, if the color is redness less than the first color determination threshold, the color is classified as the first color level, and the color is equal to or higher than the first color determination threshold. If it is reddish below the second color determination threshold, it is classified as the second color level, and if it is reddish above the second color determination threshold and below the third color determination threshold, it is classified as the third color level. If the color is reddish not less than the third color determination threshold and less than the fourth color determination threshold, the color is classified as the fourth color level, and if the color is reddish not less than the fourth color determination threshold, the fifth color. Classified as level. Then, the output control unit 23 displays the classified color level as a result on the display 3 together with the position of the tooth trace corresponding to the color level, for example, and presents it to the subject.

  Next, the tooth mark detection process according to the first embodiment will be described collectively.

  That is, as shown in FIG. 3, the tooth mark detection process according to the first embodiment is performed by, for example, a doctor in charge as a user using the operation unit 4 of the tooth mark detection apparatus S to instruct execution of the tooth mark detection process. This is started when an operation is performed. When the tooth mark detection process is started, an image including the front surface of the subject's tongue is first captured using the camera C (step S1). Here, the image imaged by said step S1 is the tongue image TG containing the test subject's tongue T illustrated, for example to Fig.4 (a).

  When an image including the tongue image TG is captured in step S1, the image registration unit 20 of the processing unit 2 acquires image information corresponding to the image via the interface 1, and temporarily stores the image information. 5 is recorded in a nonvolatile manner (step S2). At this time, the image registration unit 20 records (registers) the image information in the recording unit 5 in association with the identification information corresponding thereto.

  Next, the tongue examination region extraction unit 21 of the processing unit 2 reads out image information recorded in the recording unit 5 based on, for example, an operation in the operation unit 4, and displays a tongue image TG corresponding to the read image information on the display 3. The tongue examination region according to the first embodiment is extracted from the tongue image TG using the extraction frame FM (step S3). More specifically, in the extraction of the tongue examination area in step S3, first, as illustrated in FIG. 4A, the tongue image TG is displayed on the display 3, and the extraction frame FM is determined based on the above operation on the display. Then, as illustrated in FIG. 4B, the tongue image in the determined extraction frame FM is cut out as a cut-out tongue image TM. At this time, the extraction frame FM illustrated in FIG. 4A is the position and length of each side (that is, the position and size of the extraction frame FM in the entire tongue image TG) by the operation of the operation unit 4 with a mouse, for example. Is determined on the display 3. Thereafter, the tongue examination region extraction unit 21 extracts a tongue examination region corresponding to the tongue T illustrated in FIG. 4C from the cut tongue image TM using the region determination mask MK according to the first embodiment. . The tongue examination area in this case is extracted in a state where the tongue examination area corresponding to the tongue T itself and the background BG are identified as illustrated in FIG. Step S3 will be described in detail later.

  Next, the tongue examination region extraction unit 21 extracts the tongue contour RL corresponding to the outer edge of the extracted tongue examination region as illustrated in FIG. Step S4). Furthermore, in order to remove the angular protrusions and fine undulations in the extracted tongue contour RL, the tongue examination region extraction unit 21 displays an approximate curve EL for the tongue contour RL with a broken line in FIG. As illustrated, the calculation is performed based on the tongue contour RL (step S5). Then, the tooth trace region extraction unit 21 further detects the convex hull CH about the approximate curve EL calculated in step S5 as illustrated in FIG. 5A (step S6).

  Here, the convex hull detected in step S6 will be described. First, in the approximate curve EL calculated in step S5, the center of the tongue T is exemplified as shown in FIGS. 4 (e) and 5 (a). There are irregularities corresponding to the tongue contour RL when viewed from the side. And the convex hull CH calculated by step S6 is a figure comprised by connecting the front-end | tips of the convex part which adjoins in the unevenness | corrugation of approximate curve EL with a straight line so that it may be illustrated by Fig.5 (a). Therefore, as a result, the convex hull CH is a figure obtained by sequentially connecting the convex portions of the tongue T as viewed from the center of the tongue T.

  The tongue contour RL extraction process in step S4, the approximate curve EL calculation process in step S5, and the convex hull detection process in step S6 are the same extraction process and calculation process as in the prior art. .

  Next, as illustrated in FIG. 5B, the tongue examination region extraction unit 21 separates the concave portion of the approximate curve EL from the center of the tongue T and the convex hull CH in the convex hull CH and the approximate curve EL. The distance L of the farthest part is detected for each of the parts (that is, the parts where the convex hull CH and the approximate curve EL are separated as a result of the presence of each tooth trace). Here, in FIG. 5B, solid circles indicate the above-mentioned most distant portions (that is, the deepest portions as the recesses) in the respective recesses (that is, the tooth marks TR) in the approximate curve EL. In the following description, the deepest portion in each recess (tooth trace TR) is referred to as a “valley”. Further, in FIG. 5B, the broken-line circles indicate the respective convex portions (that is, two contact points of the approximate curve EL and the convex hull CH for the concave portions) at both ends of the concave portions. In the following description, each convex portion is referred to as a “vertex” of the tooth trace TR. Then, the tongue examination area extraction unit 21 uses, for example, a method such as numbering as the tooth trace TR, with the concave portion having the distance L longer than the tooth trace detection threshold set experimentally or empirically in advance. Detection is performed while identifying each trace TR. Thereafter, the tongue examination region extraction unit 21 outputs the detection result to the tooth trace form detection unit 22 of the processing unit 2 so as to be mutually distinguishable (step S7). In the case illustrated in FIG. 5B, a total of 12 tooth marks TR are detected for the tongue T.

  Next, based on the detection result of each tooth mark TR up to step S7, the tooth mark shape detection unit 22 includes each tooth mark TR and corresponds to the tongue T according to the first embodiment. Extract from the tongue examination area. And the tooth trace form detection part 22 detects the form (especially color of each tooth mark area | region) of each tooth mark area | region based on the extracted image of each tooth mark area | region, respectively, and the detection result is the process part 2 Is output to the output controller 23 (step S8). Step S8 will be described in detail later.

  Thereafter, the output control unit 23 generates a display screen for displaying the detection result of the form of each tooth mark region for each tooth mark area based on the detection result of the form in step S8, and outputs the display screen to the display 3. A test subject is presented as a result of the tooth trace detection process according to the embodiment (step S9). The display screen at this time is preferably a display screen that displays, for example, each tooth mark region and its form detection result in a one-to-one correspondence.

  Thereafter, the processing unit 2 detects the tooth according to the first embodiment by detecting, for example, whether the power switch as the tooth mark detection device S is turned off or a predetermined end operation is performed in the operation unit 4. It is determined whether or not to end the mark detection process (step S10). When the tooth mark detection process is continuously executed in the determination in step S10 (step S10; NO), the processing unit 2 returns to step S1 and repeats the above-described tooth mark detection process. On the other hand, when the tooth trace detection process is ended in the determination in step S10 (step S10; YES), the processing unit 2 ends the tooth mark detection process as it is.

  Next, the tongue diagnosis region extraction process in step S3 will be described more specifically with reference to FIG.

  As shown in FIG. 6A, in step S3, the tongue examination region extraction unit 21 first records image information corresponding to the tongue image TG illustrated in FIG. 6B based on, for example, an operation in the operation unit 4. When read from the unit 5 (step S30), the tongue image TG corresponding to the read image information is displayed on the display 3. When a region to be a tongue examination region is designated by an operation of a doctor in charge using the extraction frame FM displayed superimposed on the tongue image TG (step S31; see FIG. 6B), the tongue. The examination region extraction unit 21 cuts out the cutout tongue image TM corresponding to the designated region from the original tongue image TG (see FIG. 6C).

  Next, the tongue examination area extraction unit 21 has an area determination mask MK having a shape corresponding to the width W and length H of the cut out tongue image TM (in other words, the width W and length H of the extraction frame FM). Is obtained from the mask database 51 (step S32; see FIG. 6D). The area determination mask MK at this time is configured by three mask areas MK1 to MK3 having shapes corresponding to the width W and length H of the cut-out tongue image TM, as shown in FIG. ing. Then, the tongue examination region extraction unit 21 determines the mask region MK1 as the region of the tongue T itself and the mask region MK3 as the background region in the cut out tongue image TM, and the tongue image TG corresponding to the mask region MK2. The tongue examination region is extracted by taking only the part of as a processing target (step S33). Here, as a specific method of extracting the tongue examination region, the same method as the conventional one can be used. More specifically, the tongue examination region extraction unit 21 extracts a tongue examination region using, for example, a so-called Graph Cut method, with only a portion of the tongue image TG corresponding to the mask region MK2 as a processing target. Thereafter, the tongue examination area extraction unit 21 performs a preset noise removal process on the extraction result of step S33, and further cuts out a rectangular area including only the extracted tongue examination area from the original cutout tongue image TM. (Step S34). As a result, the tongue examination region is extracted in a state where the tongue examination region corresponding to the tongue T itself and the background BG are identified as illustrated in FIG. Thereafter, the tongue examination region extraction unit 21 proceeds to step S4 in FIG.

  Finally, the form detection process of the tooth mark area in step S8 will be described more specifically with reference to FIGS.

  As shown in FIG. 7A, in step S8, the tooth trace form detection unit 22 firstly detects the tooth traces TR up to step S7, as shown in FIG. Obtained from the extraction unit 21 (step S80). In the case illustrated in FIG. 7B, a total of ten tooth marks TR are detected on one tongue T so as to be distinguishable from each other by numbers. Next, the tooth mark form detection unit 22 sets a tooth mark area according to the first embodiment for each of the tooth marks TR, and is recorded in the set tooth mark area and the color determination threshold value database 52. Based on the color determination image corresponding to the color determination threshold, a color determination range for each tooth mark region is set (step S81). More specifically, as shown in FIG. 7C, the tooth trace form detection unit 22 first determines the middle of the line segment connecting the vertices constituting the tooth trace TR for each of the tooth traces TR as illustrated in FIG. 7C. An arc whose center is the point CR and whose radius is r of a half of the line segment (that is, an arc whose diameter is the line segment) is set on the tongue T side of the line segment. And the tooth trace form detection part 22 sets the area | region of the tongue T in the semicircle enclosed with the said set circular arc and the said line segment containing the midpoint CR with the tooth trace area | region AR about the said tooth trace TR. To do. Thereafter, the tooth mark shape detection unit 22 superimposes the color discrimination image illustrated in FIG. 7D and the tooth mark area AR set for each tooth mark TR as illustrated in FIG. 7E. Thus, the color determination range for each tooth mark area AR is set (step S81).

  Next, the tooth mark form detection unit 22 extracts only the set tooth mark areas AR from the tongue examination area corresponding to the tongue T as illustrated in FIG. 8A (step S82). Then, the tooth mark form detection unit 22 performs color level determination and classification using the color determination threshold for each of the extracted tooth mark regions AR, as illustrated in an enlarged view in FIG. (Step S83). Thereafter, the tooth mark shape detection unit 22 outputs the result of the determination and classification of the color level to the output control unit 23 so as to be identifiable for each tooth mark area AR as the detection result of the form of the tooth mark area AR. Thereafter, the tooth trace shape detection unit 22 proceeds to step S9 in FIG.

  Then, the output control unit 23 presents the detection result of the form of each tooth mark area AR in step S8 to the subject or the like via the display 3 so as to be identifiable for each tooth mark area AR (step S9).

  As described above, according to the tooth mark detection process according to the first embodiment, the tongue examination region is extracted using the region determination mask MK corresponding to the dimensions (width W and length H) of the extraction frame FM ( (See step S3 in FIG. 3). Therefore, since the tongue examination area is extracted using the area determination mask MK corresponding to the size of the extraction frame FM, the tongue image TG is compared with the case where the tongue examination area is extracted using all the tongue images TG. The tongue examination region can be detected without being affected by environmental factors at the time of imaging, and more accurately and at high speed while reducing the processing load.

  In addition, the area determination mask MK is configured by superimposing the mask M1, the mask M2, the mask M3-1, and the mask M3-2 so that the respective center lines in the longitudinal direction coincide with each other, and the outer edge of the third mask. Extraction of the tongue examination region is performed for the region of the tongue image TG corresponding to the mask region MK2 between the outer edge of each of the mask 3-1 and the mask M3-2 (see FIG. 2). Each tooth mark area AR can be detected separately while reducing the above.

  Further, the corresponding convex hull CH is detected from the tongue contour RL of the tongue T extracted from the extracted tongue examination region, and the respective tooth mark regions AR are mutually identified based on the tongue contour RL and the convex hull CH. Each of them is detected, and the detection result is output (see step S4 to step S7 in FIG. 3). Therefore, compared with the case where the entire tongue image TG is used to extract the tongue diagnosis region, the processing load is more accurately and quickly reduced without being influenced by environmental factors at the time of capturing the tongue image TG. Each tooth mark area AR can be detected separately while planning.

  Furthermore, a region of the tongue T corresponding to a semicircle including the tooth mark TR having the center of the line segment connecting the vertices of both ends of the tooth mark TR in the tongue outline RL and having the diameter of the line segment as a center. Since each of the tooth mark areas AR is detected while being distinguished from the other tooth mark area AR (see FIG. 7C), the tooth mark area AR can be detected by a simple process.

  Further, the color of each tooth mark area AR is detected, and information indicating which color level the color corresponds to in relation to the color determination threshold is output for each tooth mark area AR (step S83 in FIG. 7). Reference), the color of the tooth mark area AR can be discriminated at high speed for each tooth mark area AR by a simple method.

(II) Second Embodiment Next, a second embodiment which is another embodiment according to the present invention will be described with reference to FIG. At this time, FIG. 9 is a diagram showing details of the tongue examination region extraction processing according to the second embodiment. In FIG. 9A, the same step number is assigned to the same process as the tongue examination region extraction process according to the second embodiment shown in FIG. 6A, and the detailed description is omitted.

  In the tooth mark detection processing according to the first embodiment described above, the cut tongue image TM cut out from the tongue image TG corresponding to the tongue T of the subject using the extraction frame FM is set as the detection target of the tooth mark TR. On the other hand, in the second embodiment described below, the cut-out tongue image TM is reduced to reduce the burden on the processing unit 2 and speed up the processing, and the reduced cut-out tongue image TM is displayed. On the other hand, a tongue examination region is extracted using the region determination mask MK, and each tooth trace TR is extracted from an enlarged image of the extraction result.

  Further, the hardware configuration of the tooth mark detection system according to the second embodiment is basically the same as the hardware configuration of the tooth mark detection system SS according to the first embodiment. Therefore, in the following second embodiment, the tooth numbers according to the second embodiment are used only for portions and processes different from the tooth mark detection system SS by using the same member numbers as those of the tooth mark detection system SS according to the first embodiment. A trace detection system will be described.

  As shown in FIG. 9A, in the tooth trace detection system according to the second embodiment, the tongue examination region extraction process (see step S3 in FIG. 3) with respect to the tooth trace detection system SS according to the first embodiment. Only the contents of are different.

  That is, as shown in FIG. 9A, in the tongue examination area extraction process according to the second embodiment, first, steps S30 and S31 similar to the tongue examination area extraction process according to the first embodiment are performed as shown in FIG. This is performed using the extraction frame FM on the tongue image TG illustrated in b). Next, the tongue examination area extraction unit 21 according to the second embodiment reduces the cut-out tongue image TM corresponding to the extraction frame FM by a preset reduction rate, and reduces the reduction illustrated in FIG. 9C. A cutout tongue image TMM is generated (step S31-1). Thereafter, the tongue examination region extraction unit 21 according to the second embodiment is smaller than the region determination mask MK having a shape corresponding to the size of the cut-out tongue image TMM (therefore, the region determination mask MK according to the first embodiment). 9 (d) is acquired from the mask database 51, and the same steps S32 to S33 as the tongue examination region extraction process according to the first embodiment are performed using the acquired data. As a result, a small tongue examination region corresponding to the tongue T including the background BG is extracted as illustrated in FIG.

  Next, the tongue examination area extraction unit 21 according to the second embodiment enlarges the small tongue examination area illustrated in FIG. 9E with an enlargement ratio corresponding to the reduction ratio used in step S31-1. (Step S33-1) Further, by inverting the enlarged tongue examination region and the background BG, the tongue examination region mask RM according to the second embodiment illustrated on the left in FIG. 9F is generated (Step S33). -2). In the tongue examination region mask RM, the region corresponding to the background BG in the original cutout tongue image TM illustrated on the right in FIG. 9F is masked, and only the region corresponding to the tongue T can be extracted. Note that the reduction ratio and the enlargement ratio are preferably set experimentally or empirically.

  Thereafter, the tongue examination region extraction unit 21 according to the second embodiment uses the tongue examination region mask RM and the original cutout tongue image TM illustrated on the right in FIG. The tongue examination region of the tongue T is extracted from the TM as illustrated in FIG. 9G (step S33-3). And the tongue examination area | region extraction part 21 which concerns on 2nd Embodiment performs step S34 similar to the tongue examination area | region extraction process which concerns on 1st Embodiment with respect to the said extracted tongue examination area | region, and after that, step S4 of FIG. Migrate to Thereafter, step S4 and subsequent steps similar to the tongue examination region extraction process according to the first embodiment are executed (see FIG. 3).

  As described above, according to the tooth mark detection process according to the second embodiment, in addition to the effect of the tooth mark detection process according to the first embodiment, the extracted tongue image TMM obtained by reducing the cut-out tongue image TM. A small tongue examination region is extracted using the region determination mask MK corresponding to the size, a tongue examination region mask RM corresponding to the enlarged region is generated, and the tongue examination region is determined using the tongue examination region mask RM. Extract. Therefore, even when the original tongue image TG has an excessive amount of information, each tooth mark area AR can be reduced at high speed while reducing the processing load by reducing the extracted tongue image TM and extracting the tongue examination area. Can be detected.

  Note that programs corresponding to the flowcharts shown in FIG. 3, FIG. 6 (a), FIG. 7 (a), and FIG. 9 (a) are recorded on a recording medium such as an optical disc or via a network such as the Internet. It is also possible to cause the microcomputer or the like to function as the processing unit 2 according to each embodiment by reading and executing these by using, for example, a microcomputer or the like.

  As described above, the present invention can be used, for example, in the field of the tooth mark detection system for detecting the tooth mark TR of the tongue T for the purpose of determining the health level. In particular, the tongue image TG of the tongue T is used. If it is acquired and applied to the field of a tooth trace detection system that detects the tooth trace TR, a particularly remarkable effect is obtained.

DESCRIPTION OF SYMBOLS 1 Interface 2 Processing part 3 Display 4 Operation part 5 Recording part 20 Image registration part 21 Tongue examination area extraction part 22 Tooth mark form detection part 23 Output control part 50 Image recording part 51 Mask database 52 Color determination threshold value database C Camera S Tooth mark Detecting device SS Tooth mark detection system TG Tongue image T Tongue CR Midpoint FM Extraction frame MK Region determination mask M1, M2, M3-1, M3-2 Mask MK1, MK2, MK3 Mask region TM, TMM Cut tongue image BG Background RL tongue contour EL approximate curve CH convex hull TR tooth mark AR tooth mark area

Claims (8)

Tongue image information acquisition means for acquiring tongue image information corresponding to a tongue image obtained by imaging the tongue;
Designation means used for designating in the tongue image a tooth mark detection region used for detection of a tooth mark formed on the tongue;
An area determination mask for separating a tongue area corresponding to the tongue in the designated tooth mark detection area and a background area corresponding to the background in the tooth mark detection area other than the tongue, wherein the tooth Mask information acquisition means for acquiring area determination mask information corresponding to an area determination mask corresponding to the size of the mark detection area;
Tongue region extracting means for extracting the tongue region from the tooth mark detection region using the region determination mask;
A tongue image processing apparatus comprising: The tongue image processing apparatus according to claim 1,
The region determination mask is
A first mask having a semi-elliptical shape corresponding to the dimensions and corresponding to the region of the tongue itself;
A second mask corresponding to the dimensions, larger than the first mask and corresponding to the background region;
A third mask corresponding to the dimensions and having an outer edge in a region between the outer edge of the first mask and the outer edge of the second mask;
Are superimposed so that the respective centerlines in the longitudinal direction coincide with each other,
The tongue region extracting means extracts the tongue region for a region of the tongue image corresponding to a region between an outer edge of the third mask and an outer edge of the first mask. Image processing device. In the tongue image processing apparatus according to claim 1 or 2,
The tongue region extracting means includes
Reduced tongue region extracting means for extracting a reduced tongue region from the reduced tooth mark detection region using a tongue region extraction mask corresponding to the size of the reduced tooth mark detection region obtained by reducing the designated tooth mark detection region;
Enlarging means for enlarging the extracted reduced tongue area using an enlargement ratio corresponding to the reduced tongue area;
Mask generating means for generating a tongue region mask corresponding to the enlarged reduced tongue region;
With
The tongue image processing apparatus, wherein the tongue region extraction unit extracts the tongue region from the tooth mark detection region using the generated tongue region mask. In the tongue image processing apparatus according to any one of claims 1 to 3,
Contour extracting means for extracting the contour of the tongue from the extracted tongue region;
A convex hull detecting means for detecting a convex hull circumscribing the contour based on the extracted contour;
Based on the extracted contour and the detected convex hull, a tooth trace region extracting means for extracting a tooth trace region corresponding to each of the one or a plurality of the tooth traces while identifying each other from the other tooth trace regions. When,
Output means for outputting tooth mark area information indicating the detected tooth mark area;
A tongue image processing apparatus further comprising: In the tongue image processing apparatus according to claim 4,
The tooth trace region extracting means is centered on a midpoint of a line segment connecting each contact point between both ends of the tooth trace and the detected convex hull in the extracted contour, and the line segment is a half diameter A tongue image processing apparatus, wherein each of the tongue image areas corresponding to a semicircle including the tooth mark is extracted as the tooth mark area while being distinguished from the other tooth mark areas. . In the tongue image processing measure according to any one of claims 1 to 5,
The output means includes
Color detection means for detecting the color of the detected tooth trace region;
Storage means for storing threshold information indicating a threshold for classifying the color into a preset color level;
With
The output means outputs color information indicating the color level corresponding to the detected color based on the stored threshold information and color information indicating the detected color, to the tooth mark region. A tongue image processing apparatus, wherein information is output for each tooth mark region as information. In a tongue image processing method executed in a tongue image processing apparatus for processing a tongue image obtained by imaging a tongue,
A tongue image information acquisition step of acquiring tongue image information corresponding to the tongue image;
A designation step for designating a tooth mark detection region used for detection of a tooth mark on the tongue in the tongue image;
An area determination mask for separating a tongue area corresponding to the tongue in the designated tooth mark detection area and a background area corresponding to the background in the tooth mark detection area other than the tongue, wherein the tooth A mask information acquisition step for acquiring area determination mask information corresponding to an area determination mask corresponding to the dimension of the mark detection area;
A tongue region extracting step of extracting the tongue region from the tooth mark detection region using the region determination mask;
Including a tongue image processing method. A tongue image processing apparatus for processing a tongue image obtained by imaging a tongue, which is used to designate a tooth mark detection region used for detecting a tooth mark formed on the tongue of the tongue image. A computer included in the tongue image processing apparatus including the specifying unit;
Tongue image information acquisition means for acquiring tongue image information corresponding to the tongue image;
An area determination mask for separating a tongue area corresponding to the tongue in the designated tooth mark detection area and a background area corresponding to the background in the tooth mark detection area other than the tongue, wherein the tooth Mask information acquisition means for acquiring area determination mask information corresponding to the area determination mask corresponding to the dimension of the mark detection area, and
Tongue region extracting means for extracting the tongue region from the tooth mark detection region using the region determination mask;
A tongue image processing program characterized by being made to function as: