JP4809056B2 - Face classification device for cheek makeup, face classification program, and recording medium on which the program is recorded - Google Patents

Description

The present invention relates to a face classification device for cheek makeup , a face classification program, and a recording medium on which the program is recorded. In particular, the present invention provides a face adjustment method using cheek makeup that is effective for each cheek morphological feature. The present invention relates to a face classification apparatus , a face classification program, and a recording medium on which the program is recorded.

  Conventionally, at the top of a user's questionnaire in cheek makeup, troubles such as “I don't know where to apply cheek” or “I can't blur cheek” have been ranked. Therefore, conventionally, there is a blusher makeup simulation system in which blusher is applied to a face image according to each selected type and displayed only by the user selecting a makeup image type (see, for example, Patent Document 1). .

  There is also a method for classifying the subject's facial features according to the features of the facial form and showing makeup guidelines based on the classified type (see, for example, Patent Document 2).

In addition, in general cheek makeup advice and counseling, the expression of cheeking uses the expressions "teach cheek along the cheekbone", "smile and cheek the highest position", etc. Yes. The purpose of cheek makeup is mainly to produce images such as “adorable” and “adult-like”, and to make three-dimensional corrections such as “make cheeks clear” and “show kokeh cheeks”. .
JP 2000-279228 A JP 2001-224430 A

  However, it is preferable to apply a make-up suitable for each of the cheek makeups according to morphological features such as the contour of the face and the length of the cheek. However, in the prior art, there is nothing that accurately shows the position where the face is applied, the range to be blurred, and the shape by making the face correspond to the morphological characteristics. In addition, there has been no proposal for a face adjustment method using a make-up of a cut surface that shows the blood color of the face using cheeks and balances the three-dimensional effect.

  In addition, as shown in the conventional method described above, interpretations such as “perform cheek along the cheekbone” and “perform cheek at the highest position by laughing” differ depending on the individual, and depend on the level of laughter. Therefore, it may not be the intended finish.

  In addition, as for image production and three-dimensional correction methods, none of the cheek makeup methods taking into account different characteristics of individual cheeks were accurately shown. Furthermore, in the cheek makeup, the range and shape to which it is applied influences the evaluation of the finished product, but the range and shape are not clearly shown.

The present invention has been made in view of the above-described problems, and is a face classification device , a face classification program, and the program for providing a face adjustment method by cheek makeup that is effective for each cheek morphological feature. An object of the present invention is to provide a recording medium on which is recorded.

  In order to solve the above problems, the present invention employs means for solving the problems having the following characteristics.

The invention described in claim 1 is a face classification device for cheek makeup for classifying a subject's cheek length, skeleton, and fleshiness, a photographing means for photographing a face image of the subject, a photographed facial image Control means for performing control for positioning to at least one of four types classified in advance based on at least two indicators of cheek length and skeleton and fleshiness of cheek face And display means for displaying the classified results in correspondence with the four types on a map positioned in each quadrant of the plane coordinate space.

According to the first aspect of the present invention, cheeks can be easily and effectively classified based on at least two indicators of cheek length and skeleton and fleshiness of the cheek face. Accordingly, it is possible to provide a face adjustment method using an effective cheek makeup for each morphological feature.

The invention described in claim 2 is characterized in that the control means uses a preset average face and each of the four types of representative faces in order to evaluate the face of the subject.

According to the second aspect of the invention, by using the average face image and each of the four types of representative face images, the accuracy of cheek analysis can be improved and classification can be performed easily and efficiently.

According to a third aspect of the present invention, the control means has a vertical width drawn from right under the pupil to the face outline, a horizontal line from the side of the nose to the face outline, a horizontal width, and the vertical width of the cheek. The ratio of the width ratio is evaluated based on the average face, and the skeleton and fleshiness are evaluated by the degree that the cheek bone is conspicuously slender and the flesh is inconspicuous and fluffy. To do.

According to the third aspect of the present invention, the cheek length and the skeleton and fleshiness can be evaluated on an appropriate scale.

According to a fourth aspect of the present invention, the display means approximates the cheek length and the skeleton and fleshiness of the average face based on the classified face forms. An adjustment method for adjusting cheek length and skeleton / feeling of fleshiness with cheek makeup is displayed.

According to the invention described in claim 4, by displaying an adjustment method for adjusting the cheek length, skeleton and fleshiness with cheek makeup, it is possible to easily grasp the appropriate cheek makeup method for each type. Can do.

According to a fifth aspect of the present invention, the cheek makeup comprises: a starting line for inserting a cheek, a line connecting the anticle from the tip of the nose, a horizontal line drawn from the center of the eye, and a line drawn from the point intersecting the facial contour to the mouth corner. It is characterized by an intersection.

According to the fifth aspect of the present invention, since it is possible to designate a starting point for inserting a common cheek for each type, a highly accurate cheek makeup can be easily realized.

According to a sixth aspect of the present invention, the range of the cheek blur is parallel to the first line drawn from the point where the horizontal line passing through the center of the eye intersects the contour of the face and the tip of the nose. A second line drawn down to the corner of the mouth, a third line drawn vertically from the corner of the mouth, and the third line moved in parallel until the horizontal line passing through the center of the eye intersects the contour of the face. It is set on the basis of a parallelogram surrounded by four lines.

According to the sixth aspect of the present invention, since a common cheek blurring range can be specified for each type, a highly accurate cheek makeup can be easily realized.

The invention described in claim 7 is a face classification program for causing a computer to function as each unit included in the face classification device according to any one of claims 1 to 6 .

According to the seventh aspect of the present invention, cheeks can be easily and effectively classified. In addition, it is possible to provide a face adjustment method using an effective cheek makeup for each morphological feature. Further, by installing the program, face classification in the present invention can be easily realized by a general-purpose personal computer or the like.

The invention described in claim 8 is a computer-readable recording medium on which the face classification program according to claim 7 is recorded.

According to the eighth aspect of the present invention, the face classification program can be easily installed on a plurality of other computers using a recording medium. Further, by installing the program, it is possible to easily realize the face classification method and the face adjustment method by cheek makeup in the present invention on a general-purpose personal computer or the like.

  According to the present invention, cheeks can be easily and effectively classified by morphological features. In addition, it is possible to provide a face adjustment method using an effective cheek makeup for each morphological feature.

<Outline of the present invention>
The present invention analyzes the morphological features of cheeks in length, width, skeletal feeling, and fleshiness, maps (MAP) an identification method for classifying cheeks by morphological features, and follows the classified results. We will show you how to adjust your face with effective cheek makeup.

  Furthermore, the present invention is a suitable cheek makeup method for appropriately showing the position, range, and shape of applying a cheek makeup, and taking the characteristics of the individual cheeks into a preferable color and adjusting the three-dimensional effect of the cheeks. I will provide a.

  More specifically, it is considered that the present invention is important for favoring the appearance of the cheek length and the skeleton / feeling of flesh in the technique of changing the form of applying cheek makeup according to cheek characteristics. Analyze and classify individual cheek length, skeleton and texture with tools that map, and present cheek makeup according to the type of classification, that is, the position on the map. A suitable cheek makeup according to the conditions is simply and efficiently performed.

  In addition, the index of analysis and classification in the map is expressed by the space axis (choke length) and the form axis (choke skeleton and fleshiness). The space axis is long or short, or the skeleton and fleshiness is cheekbones. Means whether it is conspicuously slender and straight, or the cheek bone is conspicuously fluffy. In addition, by placing an example of a face having a standard balance cheek shape, such as an average face, in the quadrant classified by two axes and the center where the two axes intersect, the accuracy of analysis and classification is improved. Classification can be performed simply and efficiently.

<Embodiment example>
Next, a preferred embodiment of a face classification method, a face classification device, a classification determination map, a face classification program, and a recording medium on which the program is recorded according to the present invention will be described with reference to the drawings. .

<Device configuration>
First, input the evaluation information about the face of the subject, analyze and evaluate which class on the map belongs based on the cheek classification map prepared in advance, and further display the cheek makeup method A device configuration example of the face classification device will be described. Note that the face classification device to be described later can also be used when performing analysis processing or the like related to the creation of a classification determination map to be described later.

  Here, the face classification device uses, for example, a general-purpose personal computer, a server, etc., and installs an execution program (face classification program) that can cause the computer to execute each process according to the present invention. Face classification processing can be realized.

  FIG. 1 is a diagram illustrating an example of a device configuration of a face classification device according to the present invention. 1 includes an input device 11, an output device 12, a drive device 13, an auxiliary storage device 14, a memory 15, a CPU (Central Processing Unit) 16 that performs various controls, and a network connection. The apparatus 17 and the imaging apparatus 18 are configured to be connected to each other via a system bus B.

  The input device 11 has a pointing device such as a keyboard and a mouse operated by a user, and inputs various operation signals such as execution of a program from the user. The output device 12 has a display for displaying various windows and data necessary for operating the computer main body for performing the processing in the present invention, and displays the program execution progress and results by the control program of the CPU 16. can do. Further, the output device 12 may have a function of a printer or the like. In that case, various information that can be acquired such as evaluation item data, a classification determination map, a face classification result, and a cheek makeup method are described. It can be printed on a print medium such as paper and provided to a subject or the like.

  Here, in the present invention, the execution program installed in the computer main body is provided by the recording medium 19 such as a CD-ROM, for example. The recording medium 19 on which the program is recorded can be set in the drive device 13, and the execution program included in the recording medium 19 is installed in the auxiliary storage device 14 from the recording medium 19 via the drive device 13. As the recording medium 19, other than a CD-ROM, for example, a recording medium for recording information optically, electrically or magnetically, such as a flexible disk or a magneto-optical disk, a ROM (Read Only Memory), a flash memory. Various types of recording media such as a semiconductor memory that electrically records information can be used.

  The auxiliary storage device 14 is a storage means such as a hard disk, such as an execution program according to the present invention, a control program provided in a computer, a face image for each morphological feature displayed on a map, an average face image, etc. Information such as various types of image data, subject cheek classification results, cheek makeup methods according to cheek morphological features, and the like can be accumulated, and input / output can be performed as necessary.

  The CPU 16 performs processing of the entire computer, such as various operations and input / output of data with each hardware component, based on a control program such as an OS (Operating System) and an execution program read and stored by the memory 15. To control each process described later in the face classification process or the like. In addition, the CPU 16 sends various information such as the classification result of the subject's cheeks necessary during the execution of the program, information necessary for displaying the classification determination map in the present invention on the output means 12, and the like from the auxiliary storage device 14. It can be acquired, and the result executed by the program can be stored in the auxiliary storage device 14.

  The network connection device 17 obtains an execution program from another terminal or the like connected to a communication network represented by the Internet or the like or a LAN (Local Area Network) by connecting to a communication network or the like. The execution result obtained by execution or the execution program itself in the present invention can be provided to another terminal or the like.

  The imaging device 18 includes a CCD (Charge Coupled Devices) camera or the like, and captures a front face image of a subject, an oblique face image from a predetermined angle, and the like. The captured image is accumulated in the auxiliary storage device 14 and the like, read out as necessary, and subjected to processing such as face classification.

  With the apparatus configuration as described above, the face classification processing according to the present invention can be executed. Further, by installing the program, the face classification processing in the present invention can be easily realized by a general-purpose personal computer or the like.

<Face Classification Method in the Present Invention>
Next, the classification determination map used as a guideline for the face classification method and the cheek makeup method according to the present invention will be described specifically, including the contents of each study up to the present invention.

<1. Analysis of cheek characteristics and examination of classification method>
<1-1. Sample collection for evaluation term extraction>
First, in order to examine what expressions are used in capturing cheek features, we collected facial photographs of people with different facial features. Specifically, samples were collected by photographing the face images of 21 women in their 20s to 40s working at a cosmetics company, facing the front and at an angle of 45 degrees.

<1-2. Examination of expression terms>
Next, the terminology suitable for cheek feature evaluation was examined. The above 21 face photographs were observed one by one by four research staffs, and the items recalled as expressions of each cheek were spoken and recorded. Here, FIG. 2 is a diagram illustrating an example of a keyword recalled about cheek characteristics. In the example shown in FIG. 2, the number of words recalled about cheeks was 49 words.

<1-3. Arrangement of elements>
Next, the recorded expression terms were organized by element based on the semantic content. Here, the elements are “upper cheek”, “cheek bone”, “lower cheek”, “meat”, “raised shape”, “center of gravity”, “recess”, “length”, and “width”. did. Next, the contents of each word are examined, and those with high semantic similarity are collected into one, and those that are not suitable for general expression are excluded, and 12 adjective pairs are selected as evaluation terms used for preliminary examination. did.

  Here, FIG. 3 is a diagram showing an example of an evaluation sheet on which evaluation terms used in the preliminary examination are described. Note that the memory 20 provided in FIG. 3 shows the evaluation stages for the left and right elements in five stages, and the center of the memory is the left and right elements (for example, “there is a sense of flesh”, “ "I have a skeletal feeling").

  The layout of the evaluation sheet and the arrangement of each element in the present invention are not limited to this. For example, an area where numbers can be input is provided instead of the memory, and 5 levels of numbers may be written in the area. In this case, for example, in FIG. 3 (1), “I feel flesh” is set to “1”, “I feel skeleton” is set to “5”, “I can say neither” is set to “3”, etc. in advance what numerical value corresponds to what is assumed to be set. In the present invention, the evaluation stage is not limited to five stages.

  Further, for the evaluation using the evaluation sheet shown in FIG. 3, for example, the evaluation sheet shown in FIG. 3 is displayed on the display of the face classification device 10 described above, and the evaluator or the like performs a predetermined evaluation on the face of the subject using the input device 11 or the like. Evaluation may be performed by inputting information. Note that the evaluation information input method can be set by, for example, using a pointing device such as a mouse to position and select a cursor at a predetermined position of a predetermined evaluation item. The data input method is not limited to this, and for example, evaluation information may be input using a keyboard or the like.

<1-4. Selection of evaluation terms>
Next, in order to examine the appropriateness of the evaluation terms, the face was actually evaluated using the 12 adjective pairs shown in FIG. 3 described above. Specifically, the subjects were evaluated by the four researchers in charge using the evaluation sheet shown in FIG. 3 for the samples collected from the 21 women described above. At this time, the skeletal sensation, fleshiness, and facial contours of the entire face were also added to the evaluation items in order to examine the relevance to cheek evaluation.

  From the examination results described above, it was judged that items with differences in evaluation were important as evaluation items. In addition, words with high correlation are represented by a single word, and expressions that are difficult to understand are corrected to select cheek evaluation terms. Furthermore, the cheek feeling and skeletal feeling are highly correlated with the evaluation of the flesh feeling and skeletal feeling of the entire face, so that the whole face is judged as a target and shared with the cheek judgment.

<1-5. Extraction of cheek analysis index>
<1-5-1. Face evaluation experiment>
Next, in order to extract an index for analyzing cheek characteristics, an evaluation sheet of selected words was used, and 34 faces of general monitors up to their 40s, mainly 20s, were actually evaluated. Here, FIG. 4 is a figure which shows an example of the evaluation item used for cheek analysis index extraction. As shown in FIG. 4, the method used for the evaluation is 8 items such as “chubby”, “clean”, “wide”, “narrow” for cheeks, and “the body has a feeling of flesh-has a skeletal feeling” for the entire face. The total number of items was 9 items. Also in FIG. 4, as in FIG. 3, the bipolar scale (memory) has five levels.

  The evaluation was performed by 4 to 6 persons including a beauty engineer and a product development staff in addition to the research staff. The reason for increasing the number of evaluators is to avoid variation in evaluation among individuals. However, the definitions of the evaluation items were shared, and confirmation was performed in advance so as to reduce the variation in evaluation.

  The above evaluation can also be performed using the face classification device 10. That is, the evaluation sheet shown in FIG. 4 can be displayed on the display of the face classification device 10, and evaluation can be performed by the evaluator and the like using the input device 11 to input predetermined evaluation information by the method described above. .

<1-5-2. Index extraction by factor analysis>
Next, the average evaluation score of each subject was calculated for each item from the evaluation score of each evaluator described above. Moreover, factor analysis was performed using this value. Here, factor analysis is a statistical technique for reducing information of variables correlated to each other to a small number of potential factors. The variable here corresponds to the average evaluation score of each item in this case. Therefore, by performing factor analysis, correlated items can be collected and a small number of evaluation factors can be extracted.

  Also, commercially available statistical analysis software (for example, SPSS (Statistical Package for Social Science) ver. 11.0J) is used for factor analysis, and the software is installed in a computer device such as the face classification device 10 for example. Factor analysis can be performed. The principal component method was designated as the factor analysis method, and the varimax method was designated as the rotation method.

  As a result of analysis based on the above contents, two factors were extracted. FIG. 5 is a diagram illustrating the factor loading obtained from the factor analysis result. Here, the numerical values shown in FIG. 5 indicate that the larger the absolute value, the stronger the relationship with each factor. In addition, in order to clarify the factor structure, the two items of “cheap is broad-narrow” and “cheap is slack-not slack” are deleted from this analysis.

  As shown in FIG. 5, items having a large value in the first factor are “there is a skeletal feeling—there is a fleshiness”, “the cheeks are clear—the chubby”, “the cheek bones are not lying down” ", It is interpreted as a factor of" skeleton and fleshiness ".

  In addition, the items with large values in the second factor are “the cheek is short-long”, “the cheek bone is not protruding forward—it is protruding”, “the cheek is protruding—flat”, Since it is composed of the items “below the nose and above”, it is interpreted as a factor of “the length and irregularities of the cheek”. The cumulative contribution, which is an index of the degree of explanation, was 70.1%.

  Here, FIG. 6 is a diagram illustrating an example in which each target is plotted based on the score of the extracted first and second factors. Analyzing the cloth arrangement shown in Fig. 6, each object (34 people in general monitors up to the 40s centered on the 20s mentioned above) is positioned on the axis of the skeleton and fleshiness and the axis of the cheek length and unevenness. It can be seen that they are separated.

  Therefore, it can be estimated that the cheek feature can be captured by the two axes of “the cheek is long—short / unevenness large—small” and the “skeletal feeling—feeling of flesh”. However, because of this, the point of interest to determine the characteristics is ambiguous, so further analysis was performed.

<1-5-3. Organizing the focus of indicators>
The face with the highest feature of each axis, that is, the face with the highest absolute value of each score, was extracted from the plot of FIG. The examination results will be described with representative examples.

<Skin and flesh>
First, regarding the axis of “skeleton / feeling of flesh”, a sample of S15 having a high negative score and S8 having a high positive score will be described as an example. In order to more clearly analyze the features with skeletal flesh, for example, two samples were subjected to image analysis using the face classification device 10 or the like and examined. In addition, commercially available image processing software (for example, Photoshop etc.) can be used for the image analysis installed in the face classification apparatus 10. In addition, as image processing, monochrome posterization processing that shows gradations of light and dark in steps was performed, and the distribution of shadows on the face was examined. Here, FIG. 7 is a diagram illustrating an example of an image analysis of a face having a high skeleton and fleshiness. FIG. 7 shows an example in which gradations are displayed in 16 steps for the images of S15 and S8.

  In FIG. 7, in the face image 21-1 of S 15 with a high negative score, that is, a high evaluation of the skeletal feeling, the cheekbone line 22-1 is square, and the flesh line from the eyes to the mouth is linear. It can be seen that the cheek line 23-1 is sharp.

  On the other hand, in the face image 21-2 of S8 where the positive score is high, that is, the evaluation of fleshiness is high, the cheekbone line 22-2 is gently rounded and the flesh line 23-2 from the eyes to the mouth. Is also curvy and rounded. Further, since the flesh is good, the cheek line 23-2 has a clear bulge, and it can be seen that the bulge of the highest portion, that is, the bright area 24 is present in the area under the left and right eyes. This study made it possible to clarify the difference in characteristics between the skeleton and fleshiness. That is, these elements can be used as a basis for determining the skeleton and the feeling of flesh.

<Length of cheeks and irregularities>
Next, the cheek length was measured from the front photo of the face to be evaluated, and appropriate criteria were examined. For example, the length of the cheek was measured with several candidates such as the distance from the lower eyelid to the mouth corner position, the length of the vertical line descending from the pupil to the face line, and the like.

  Finally, based on the sense of the person who judged that the cheek was long, the criteria were determined in consideration of the high degree of matching with that person. The standard is that the horizontal width is the horizontal line from the side of the nose to the contour, the vertical width is the vertical line from just below the pupil to the contour of the face, and the ratio of the horizontal width to the vertical width is used.

  Here, in the “standard balance face” created by averaging the face photographs of women in their 20s by image composition, that is, the average face reflecting the average value of women in their 20s, the ratio was about 1: 2. . Considering this value and the distribution of the values obtained from the above-mentioned measurement of 34 persons, the standard length is in the range of 1.95 to 2.05, and the length is less than 1.95. When it exceeded 2.05, it was decided that the cheek was long.

<1-6. Creating a cheek classification map>
Based on the above results, a map (MAP) was created to analyze and classify cheek characteristics. At the time of preparation, the obtained indexes were further arranged in consideration of convenience and versatility such as beauty counseling.

  FIG. 8 is a diagram illustrating an example of a cheek classification determination map. Here, with regard to the above-described second factor “cheo length / unevenness”, it was decided that only the cheek length was used as an index in consideration of simple analysis and determination criteria uniquely determined. Therefore, the cheek analysis and classification index was combined by combining the two factors of the “factory skeleton / feeling of flesh” of the first factor and the “length of cheek” which was the second factor.

  FIG. 8 shows an example in which the vertical axis is “cheek length”, the horizontal axis is “skeleton / feeling of flesh”, the cheek length axis is the space axis, and the axis of skeleton / feeling of flesh is the form axis. However, the arrangement (vertical and horizontal) of the two axes in the present invention is not particularly limited, and for example, the elements of the two axes may be reversed.

  Here, in FIG. 8, in order to facilitate the cheek analysis, expressions (characters) indicating the characteristics of each axis may be given. Further, a face image (photograph) that significantly represents the characteristics of each quadrant may be created by computer graphics by the face classification device 10 and placed at a corresponding position. At this time, not only the front face image but also a face image from an oblique direction (for example, 45 degrees oblique) may be created and arranged. As a result, as described above, it is possible to solve the point that the evaluation accuracy is deteriorated only with the front face, and when analyzing the cheek skeleton and fleshiness, it is possible to grasp the cheekbone and cheek flesh characteristics with high accuracy. .

  That is, as shown in FIG. 8, two axes, a first axis indicating the face type including the balance of the facial space, a second axis indicating the face type consisting of a skeleton-carved feeling, and the two axes. And a coordinate space in which a face image (for example, an average face image) having a standard and ideal stereoscopic shadow set in advance is arranged in the center where two axes intersect.

  Here, the cheek analysis, the classification index, and the point of determination will be described. FIG. 9 is a diagram illustrating an example of how to determine the cheek analysis. The cheek evaluation factor is composed of two indicators, “skeleton / feeling of flesh” and “height length / concaveness”. As for the point of determination, the “skeleton / feeling of flesh” is judged by the cheek bone, the flesh inside, and the face line from the bottom of the ear to the chin.

  More specifically, the point for determining the “skeletal sensation” is that when the subject turns diagonally as shown in the oblique face image 30 of FIG. For example, the fact that the face line 32 below the ears on the near side is straight, the line 33 on the side near the mouth on the near side is straight, and so on.

  In addition, as for the point of ascertaining “feeling of flesh”, as shown in the oblique face image 30 of FIG. 9 (a), when the subject turned diagonally, “the cheekbone line 31 on the back side appears to be gently rounded” “The fact that the lower face line 32 is curved from the ear on the near side”, “The horizontal line 33 on the front side of the mouth is curved”, and the like are points to be determined.

  In addition, as shown in the front face image 34 of FIG. 9B, the “choke length” has a horizontal line 35 from the nose to the face outline, and a vertical line from right under the black eye to the face outline. When 36 is the vertical width, the determination is made based on, for example, a width to width ratio of 1: 2 ± 0.05. Specifically, if the horizontal width 35 and vertical line have a width to width ratio of 1: 1.8, it is determined that the cheek length is shorter, and the width to height ratio is 1: 2. If it is 2, it is determined that the cheek is longer.

  As described above, it is possible to accurately analyze and classify cheek characteristics by making a judgment along the cheek classification judgment map shown in FIG. 8 and the judgment points in the above-described indices.

  Note that the face image of the subject used for the analysis described above is captured in advance by the imaging device 18 of the face classification device 10 and stored in the auxiliary storage device 14, and analysis is performed using the stored image. It can be carried out. Further, in the analysis, the evaluation results in each item of the above-described evaluation information are totaled, the results are compared with preset values, and the classification result of the subject is shown in the cheek classification determination map shown in FIG. Positioning is performed at a predetermined position by a circle or a star, and displayed on a display or printed paper.

<2. Examination of cheek makeup method>
<2-1. Study of target cheek makeup>
Next, a face adjustment method using a cheek makeup according to the morphological features classified by the cheek classification determination map described above will be described. In general, the method of applying cheek makeup differs depending on the image to be produced. Therefore, in the present invention, examination was made focusing on “rendering of blood color”.

  The reason for paying attention to the blood color is based on the following three points. First of all, because blood color is an element that everyone originally has when it is in a healthy state, we thought it would lead to naturally bringing out the original healthy beauty of each individual. It is.

  In addition, as an example of cheeks with good blood color, for example, an infant's cheeks are self-explanatory, but in an example where `` the blood color is good and preferable '', the blood color is in the area connecting the eyes, nose, mouth, and ears. appear. Therefore, the second reason for focusing on the blood color is that it is possible to derive the position where the cheek makeup is to be derived according to the rules common to all people using the eyes, nose, mouth and ears as guidelines. Furthermore, the third reason for focusing on blood color is that the production of blood color is intended to be used by many people for cheeks and is a cheerful element that feels beautiful, so it meets the needs of many people. It is.

  In addition, the purpose of using teak was surveyed for 68 women in their 20s and 30s who usually make cheek makeup. As a result, the most frequently selected item was “To show the blood color well”. In addition, as a result of a questionnaire survey of 40 women in their twenties who usually do cheek makeup on the cheeks that they think are beautiful, the most frequently answered responses are “There is a beam”, “Moderately plump” and “Good blood” there were.

  In general, when applying makeup, a makeup base, foundation or the like is used to adjust the color unevenness of the bare skin. At this time, the natural color on the bare skin is almost erased. Therefore, it can be said that adding a blood color with a cheek makeup is a natural production method of regaining the original blood color element.

<2-2. Examination of the location and range for applying cheek makeup>
The cheek makeup applied by beauty technicians is more beautiful than what ordinary people do on their own. This can be attributed to the fact that beauty technicians learn the characteristics of individual faces based on empirical rules and expertise, and learn how to make them look beautiful and have beautiful finishes. Therefore, we analyzed how to apply teak makeup by beauty technicians and tried to extract the laws.

  First of all, using facial photographs of 10 models with different cheek characteristics, we conducted a questionnaire to 27 beauty technicians, "If you want to express natural blood color with cheek makeup, how much will you blur around?" did. In addition, the center and range of the position where the cheek makeup was applied were directly written on the paper on which each face was printed.

  FIG. 10 is a diagram illustrating an example of a total result of the centers and starting points for performing cheek makeup that produces a blood color. FIG. 10 shows an example in which the above-mentioned beauty technician's writing is totalized for a certain model (model A) and a line leading to the center of the cheek makeup is drawn.

  First, it is the position of the center of the cheek makeup that expresses the blood color, but it is concentrated at almost the same location as shown in FIG. 10, and it can be seen that it is near the center of the area connecting the eyes, nose, mouth, and ears.

  Furthermore, as a result of studying that the part is guided with reference to an easily identifiable element of the face, as shown in FIG. 10, an icicle (a part that is depressed when the mouth in front of the ear is opened) is connected by a line 41 to When a line 43 is drawn at the corner of the mouth from the point where the horizontal line 42 drawn from the center of the face intersects the outline of the face, it was found that the intersection of both coincides with the answer of the beauty technician.

  Next, the range in which the cheek makeup is blurred will be described. FIG. 11 is a diagram illustrating an example of a totaling result of a range in which a cheek makeup that produces a blood color is blurred. In FIG. 11, model A is used as in FIG. As shown in FIG. 11, it was found that the range where the makeup is blurred is within the area connecting the eyes, nose, mouth, and ears. Furthermore, this range was found to fall within the parallelogram drawn using the line leading to the center to which the cheek makeup was applied.

  Specifically, as shown in FIG. 11, a line (first line) 44 drawn to the nose from the point where the horizontal line 42 passing through the center of the eye drawn when guiding the center of the cheek makeup first intersects the contour of the face. A line (second line) 45 is provided in parallel to the mouth corner. Further, a vertical line (third line) 46 is drawn upward from the mouth corner, and the vertical line 46 is moved in parallel until the horizontal line 42 passing through the center of the eye intersects the contour of the face (fourth line). Line) 47 is provided. These lines 44 to 47 can constitute a parallelogram as shown in FIG. It can also be seen that the range of blurring the cheek makeup written by the beauty technician is within this parallelogram.

<2-3. Definition of location and range for applying cheek makeup>
Based on the above, it has become possible to derive the position of the starting point of cheeking for all faces and the standard of the range to blur using the eyes, mouth and nose, which are the three major elements that determine the face space. . In the following description, the position of the starting point for applying cheek makeup is referred to as “best point”.

  In other words, the starting point of the bloody color production cheek makeup is the intersection of the line connecting the licking from the tip of the nose, the horizontal line drawn from the center of the eye, and the line drawn from the point intersecting the face outline to the mouth corner. The range of blurring is a parallelogram drawn with the best point as a guideline. Specifically, a line is drawn to the tip of the nose from the intersection of the horizontal line passing through the center of the eye and the outline of the face, and then the line is lowered in parallel to the mouth corner. Furthermore, it is a parallelogram that is formed when a vertical line is drawn upward from the corner of the mouth and the vertical line is moved in parallel to a point where it intersects the horizontal line passing through the center of the eye.

  In this way, it was found that the best point of the starting point for applying the cheek and the range of blurring can be derived by a method common to all types.

<2-4. Examination of cheek makeup shape by cheek characteristics>
From the rule of thumb of beauty technicians, it was predicted that manipulating the shape of the cheek makeup taking into account the characteristics of the cheek would look good on the person and give a beautiful finish. Actually, even if the conditions of the starting point for applying the teak and the range to be blurred are satisfied, the finish may not be suitable.

  Therefore, the direction of adjustment was taken into account depending on the position of the cheek classification map, and the hypothesis was made that the shape of the cheek makeup was also changed according to the directionality. That is, adjustment is performed so that a person with a short cheek looks long, a person with a long skeleton feels short, and a person with a noticeable skeleton feels plump and a person with a fleshiness feels clear.

  Here, a questionnaire was conducted to verify the above hypothesis. In the questionnaire, two types of cheek makeup were applied to four models with different cheek characteristics that correspond to the four quadrants of the cheek classification map. One of the cheek makeups is “OK cheek”, which uses the rules of the standard of the best point and the range of blurring, and further changes the blur shape according to the cheek characteristics. The other is “NG teak” which is not considered in the form of blur. About these two types, the information of “OK” and “NG” was hidden, and general women (40 women in their 20s) evaluated them.

  Specifically, each photograph was viewed one by one, and the finish of each cheek makeup was evaluated on a bipolar 5-level scale. The evaluation items were “natural-unnatural”, “matched—not matched”, and the like. The average value of the evaluation of 40 people for the model of 4 people was calculated and examined. FIG. 12 is a diagram illustrating an example of a result of evaluation of a cheek makeup finish.

  When statistical tests were performed on the evaluation, it was found that there was a significant difference between the evaluations of “OK cheek” and “NG cheek” (results of analysis of variance of two factors of cheek makeup 2 level and model 4 level) The main effect of teak makeup is significant at the 0.5% level in all evaluation items). As a result, it was found that “OK cheek” had a natural finish and a cheek three-dimensional appearance, and the flesh of the cheek looked slightly fluffier than “NG cheek”. This result also agrees with the above-mentioned beautiful cheek element, “Happy and moderately plump”. In addition, it was found that the balance of the whole face appears to be in order by applying this cheek makeup. In other words, with the “OK cheek” mentioned above, the cheeks can be seen three-dimensionally, you can feel the firmness of your skin softly, and restore the natural and healthy bloody beauty that the person originally has. Can do. In addition, the blood color production can adjust the balance while naturally drawing out the individual beauty.

<2-5. Difference in cheek makeup shape due to cheek characteristics>
The shape of the cheek makeup applied to each model in the above questionnaire is based on the idea of adjusting the cheek length and the appearance of the skeleton and flesh. FIG. 13 is a diagram showing an example of a map in which the shapes of cheek makeups classified by morphological features assumed by a beauty technician are arranged at predetermined positions. The “OK cheek” used in the questionnaire represents this form shown in FIG.

  In addition, when the shape of the cheek makeup of the models belonging to the four categories for which “OK cheek” was performed was analyzed by extracting the reddish color by image analysis, it was found to match the makeup area assumed by the beauty technician. I understood.

  Here, FIG. 14 is a diagram showing an example of a map in which cheek makeup shapes and ranges according to cheek morphological features are arranged at predetermined positions. In FIG. 14, the cheek shape 51-1 corresponding to the standard face is an ellipse that is inscribed in the parallelogram as a guide. Further, in FIG. 14, the standard inscribed ellipse 52 is shown with a dotted line, and the shape according to the feature of the cheek is filled in. From this, it can be seen that it is possible to indicate the direction and shape of the adjustment corresponding to the position of the cheek classification determination map.

Here, FIG. 15 is a diagram illustrating an example of the inscribed ellipse. The ellipse inscribed in the parallelogram as shown in FIG. 11 described above has the long axis A as a straight line connecting the midpoints 53-1 and 53-2 of the vertical sides of the parallelogram as a guide for blurring, and the short axis. Let B be a straight line perpendicular to the midpoint 54 of the long axis A and in contact with the upper and lower sides of the parallelogram. Further, when the half lengths of the axes A and B are a and b, the inscribed ellipse can be represented by (x ² / a ² ) + (y ² / b ² ) = 1. In the case of an average face, the relationship is a: b = 1: 0.55 ± 0.02.

  In FIG. 14, since the shape is simplified, the density to which the cheek makeup is applied is not expressed. However, the blurring method of the cheek makeup is basically the darkest starting point. Apply it so that it fades toward the boundary of the shape and the blur on the boundary blends naturally into the skin.

  Here, FIG. 16 is a diagram for explaining an example of a cheek makeup blurring method. FIG. 16 (a) shows an example of the result of measuring the brightness (density) by dividing the area into 12 stages from the center (starting point) of the cheek makeup to the outside, and FIG. 16 (b) shows the result from the center of the cheek. An example of the result of measuring the RGB value by dividing the area into 12 steps to the outside is shown.

  First, as shown in FIG. 16 (a), the cheek makeup blurring method is applied so that the center is darkest and lighter toward the outside. Here, the brightness of the skin in the vicinity of the outermost area was also measured, and when the result was plotted in FIG. 16A, the application state of the cheek makeup was brighter (lighter) from the center to the outer side, It was found that the area was almost equal to the lightness of the skin.

In addition, in order to investigate whether it is possible to predict the ideal value of brightness according to the stage of the area using a mathematical formula, the regression analysis method is performed with the number of stage of the area as the independent variable and the measured value of the brightness as the dependent variable. It was. As a result, determining the counting ^{R 2} representing a true prediction became 0.936. Moreover, the obtained regression equation was y = 1.203x + 68.682. It was found that by using this formula, the brightness value referred to at each stage can be predicted.

  Next, the change in blur color was examined based on the measured values. The RGB values of each area divided into 12 stages and the skin of the portion adjacent to the outermost area were measured. When the values were plotted on the R, G, and B graphs shown in FIG. 16B, it was found that the RGB values of the skin approached from the center of the area toward the outside.

  As for the R value, half of the area is increased to the right and the redness is increased step by step, but in the area outside it, the value is leveled out, and the change width of the value becomes small. In addition, the G value and the B value showed a change proportional to each stage, and were found to change linearly so as to connect the center value and the skin value.

  In addition, when a regression analysis method was used to examine whether or not an ideal value corresponding to the stage of the area can be predicted by a mathematical expression, a good regression formula was obtained for each value. At this time, the number of steps in the area was set as an independent variable, and each of the values of R, G, and B was set as a dependent variable.

Regarding the R value, the regression equation in the quadratic curve was well applied, and the obtained regression equation was y = −0.284x ² + 5.940x + 222.591, and the determination coefficient R ² was as high as 0.946. From this result, it was found that the small change in the value in the outer area is reasonable.

Further, the linear regression equation based on the linear equation is well applied to the G value and the B value, and the obtained regression equation has a G value of y = 4.269x + 1141833 and a B value of y = 4.269x + 125.333. , and the coefficient of determination ^{R 2} is G value is 0.986, B values were both as high as 0.964. From these results, it was shown that the G value and B value change in proportion to the stage.

  Here, FIG. 17 is a diagram illustrating an example of measured values of the brightness, hue, and saturation of the skin adjacent to each area divided into the above-described 12 stages and the outermost area. FIG. 17A shows an example of the relationship between brightness and hue, and FIG. 17B shows an example of the relationship between brightness and saturation.

  According to FIG. 17 (a), the center area to which the cheek makeup has been applied is reddish, but as it goes outward, the redness decreases and changes to yellow and approaches the skin color. In addition, according to FIG. 17B, the saturation is high and bright from the center where the cheek makeup is applied to the half (six steps) area, and there is a difference from the saturation of the skin, but the outside is the skin. It can be seen that it is applied to the skin so that it naturally approaches the saturation of the skin.

<Identification of shape according to cheek characteristics>
Here, the shape of the cheek makeup on the standard cheek shown in FIG. 14 and the shape of the cheek makeup on the cheek according to each feature can be described as follows. First, a make shape 51-1 in the case of a standard cheek is an inscribed ellipse 52 drawn with respect to a parallelogram that serves as a rough shape standard. Further, the make shapes 51-2 and 51-5 in the case where the cheek is short are shorter on the lower side than the standard inscribed ellipse 52. Further, the make shapes 51-3 and 51-4 in the case where the cheek is long are slightly longer than the standard inscribed ellipse 52.

  Further, the make shapes 51-4 and 51-5 in the case where the cheek bone is conspicuous change the upper side of the standard inscribed ellipse 52. Specifically, the upper part near the center of the face has a larger arc than the standard inscribed ellipse 52. In addition, the upper part near the contour of the face has a smaller arc than the standard inscribed ellipse 52. Further, in the make shapes 51-2 and 51-3 when the cheek bone is not conspicuous, the upper part near the center of the face is lowered from the standard inscribed ellipse 52 to be in-curved.

<2-6. Adjustment direction according to cheek characteristics>
Next, the direction of adjustment corresponding to the position on the map that has been clarified through the above-described hypothesis verification will be described with reference to the drawings. FIG. 18 is a diagram showing an example of a map describing the direction of adjustment by cheek makeup. As shown in FIG. 18, the person who has a short cheek with respect to the space axis (vertical axis) is adjusted so that the person looks long. In addition, the person who is fluffy and the flute is not conspicuous with respect to the form axis (horizontal axis) is adjusted to make the flesh look slender, conversely, the flesh is conspicuous, and the slender straight person is not conspicuous Make adjustments to make the cheeks plump. As shown in FIG. 18, since there is a cheek classification determination map, the direction of cheek adjustment can be easily indicated.

<How to use make brush>
Further, in order to easily realize each cheek makeup shape for each type as shown in FIG. 14, it is also effective to change the way of placing and moving the makeup brush (brush) used for cheek makeup. It is. Here, FIG. 19 is a diagram showing an example of a map in which the orientation of the brush corresponding to the cheek length is arranged at a predetermined position.

  As shown in FIG. 19, when the cheek is long, the cut surface 62 of the elliptical make brush (brush) 61 is placed perpendicular to the starting point, that is, the major axis is vertical. Thereby, it can adjust so that a cheek may not be seen short. When the cheek is short, the tip cutting surface of the make brush (brush) 61 is placed horizontally with respect to the starting point, that is, the major axis is placed horizontally. As a result, the cheek is adjusted so as not to be seen for a long time. These make-up methods can more easily adjust the appearance of the cheek length.

  In addition, the cheek skeleton / feeling of flesh can be adjusted by using the make-up brush as described above. FIG. 20 is a diagram illustrating an example of a map in which how to move the brush corresponding to the cheek skeleton and flesh is arranged at a predetermined position. As shown in FIG. 20, when the fleshiness is conspicuous, the brush is moved linearly so as to draw a V-shape from the ear side to the center of the face and from the center to the ear side within the range of blur. Thereby, it can adjust so that a cheek can be seen clearly. If the skeletal sensation is conspicuous, the brush is moved in a curved line so as to draw a U-shape from the ear side to the center of the face and from the center of the face to the ear side within the range of blur. Thereby, it can adjust so that a cheek can be seen plumply.

  Here, in the case of a standard cheek, since it is located at the center of two axes, the middle of the methods at both ends may be used. In other words, the brush is placed obliquely with respect to the starting point, and the brush is moved in the middle of the V and U shapes.

  In addition, the classification classification map (FIGS. 8, 13, 14, 18 to 20) described above can be selectively presented to a subject by a beauty counselor or the like, and the face classification device 10 is used. A beauty counselor or the like may be displayed on the display step by step while explaining.

<2-8. Face classification and cheek makeup procedure according to the present invention>
When the present invention is applied and a cheek makeup is actually performed on a customer (subject) at a store, that is, the above-described classification determination map (FIGS. 8, 13, 14, 18 to 20) in the present invention is used. An example of a series of face classification processes from the classification of the subject's face executed by the face classification device 10 to the display of the face adjustment method by cheek makeup based on the classification result will be described with reference to a flowchart.

  FIG. 21 is a flowchart for explaining a face classification and cheek makeup procedure using a classification determination map according to the present invention. First, a face image of a subject to be makeup is acquired (S01). In S01, for example, the front face image of the subject and a face image facing sideways at a predetermined angle (for example, 45 degrees oblique) are acquired by the imaging device 18, for example.

  Next, the cheek characteristics of the subject are analyzed from the acquired image data (S02). Specifically, the standard cheek length (balance) as shown in FIG. 8 above, and whether the cheek bone is conspicuous, whether it is slender / linear, plump / curve, etc. I do.

  Further, in the analysis result of S02, the classification to which the subject belongs is analyzed using the classification determination map shown in FIG. 8 (S03), and the classification result is evaluated. In addition, a classification | category is positioned in the predetermined | prescribed classification | category on a corresponding map based on the result and the value set beforehand based on the analysis information of a test subject's face. The classification result is plotted at a predetermined position on the map and displayed on a display or the like, or displayed as a printed matter on a subject or the like (S04).

  Next, based on the classification result, the corresponding cheek makeup method is displayed by a classification determination map as shown in FIGS. 13, 14, and 18 to 20 (S05). At this time, only the makeup adjustment method corresponding to the type corresponding to the subject among the cheek types classified may be displayed. Finally, a corresponding makeup is applied by the subject or a counselor or beauty technician in charge of the makeup (S06).

  Specifically, the method of placing the brush on the starting point for cheeking is determined, and the brush is placed so that the starting point is at the center of the brush. At this time, according to the cheek characteristics of the subject already analyzed, the cheek is placed horizontally if it is short and vertically if it is long. For standard cheek lengths, place them diagonally.

  Further, a cheek makeup is applied in consideration of the characteristics of the cheeks analyzed. Specifically, as described above, cheek makeup is applied using the parallelogram that is the range of blurring as a guide. At this time, if the balance is a standard cheek, an ellipse inscribed in the parallelogram is imaged, and the light is naturally blurred from the starting point to the outside. If the cheek is short, the lower side is made shorter than the inscribed ellipse, and if the cheek is longer, the lower side is made slightly longer than the inscribed ellipse.

  When the cheek bone is conspicuous, the upper part near the center of the face has a larger arc than the standard inscribed ellipse, and the upper part near the contour of the face has a smaller arc than the standard inscribed ellipse. If the cheekbone is inconspicuous, the upper part near the center of the face is lowered from the standard inscribed ellipse and incurved.

  Furthermore, if the cheekbone is conspicuous, move the brush in a curved manner to draw a U-shape, and if the cheekbone is inconspicuous, move the brush in a straight line to draw a V-shape, so It becomes easy to express the above-mentioned shape corresponding to the feature of. Thereby, the cheek makeup which shows the blood color well according to the cheek characteristic for each subject and realizes a beautiful finish can be simply and accurately performed.

  As described above, according to the present invention, cheeks can be easily and effectively classified according to morphological features. In addition, it is possible to provide a face adjustment method using an effective cheek makeup for each morphological feature. Specifically, cheeks are classified according to morphological features, and effective cheek makeup can be performed for each face shape according to the classification result. That is, according to the present invention, a cheek makeup method can be easily performed for beautifully rendering the cheek color and simultaneously adjusting the cheek length and the appearance of the skeleton and fleshiness.

  In addition, the present invention performs face analysis using a tool that maps individual cheek length, skeleton, and fleshiness, and classifies them into preset face types, and maps them according to the classified types. By presenting a suitable cheek makeup according to the characteristics of the individual cheeks by showing the shape of the cheek makeup depending on the position, anyone can perform cheek makeup easily and efficiently.

  The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to such specific embodiments, and various modifications, within the scope of the gist of the present invention described in the claims, It can be changed.

It is a figure which shows an example of an apparatus structure of the face classification | category apparatus in this invention. It is a figure which shows an example of the keyword recalled about the feature of the cheek. It is a figure which shows an example of the evaluation paper in which the evaluation term used for the preliminary examination was described. It is a figure which shows an example of the evaluation item used for cheek analysis index extraction. It is a figure which shows the factor load obtained from a factor analysis result. It is a figure which shows an example which plotted each object based on the score of the extracted 1st and 2nd factor. It is a figure which shows an example of the image analysis of a face with a high skeleton and fleshiness. It is a figure which shows an example of the cheek classification | category determination map. It is a figure of an example for demonstrating the point of determination in cheek analysis. It is a figure which shows an example of the total result of the center and the starting point which performs the cheek makeup which produces a blood color. It is a figure which shows an example of the total result of the range which blurs the cheek makeup which produces a blood color. It is a figure which shows an example of the finishing evaluation result of a cheek makeup. It is a figure which shows an example of the map which has arrange | positioned the shape of the cheek makeup according to the morphological characteristic which the beauty engineer assumed in the predetermined position. It is a figure which shows an example of the map which has arrange | positioned the shape and range of the cheek makeup according to the morphological feature of cheeks in a predetermined position. It is a figure of an example for demonstrating an inscribed ellipse. It is a figure of an example for demonstrating the blurring method of a cheek makeup. It is a figure which shows an example of the measured value of the brightness, the hue, and the saturation of the skin of the part which adjoins each area divided into each said 12 steps | paragraph and the outermost area. It is a figure which shows an example of the map which described the directionality of adjustment by a cheek makeup. It is a figure which shows an example of the map which has arrange | positioned the direction of the brush corresponding to the length of a cheek in the predetermined position. It is a figure which shows an example of the map which has arrange | positioned how to move the brush corresponding to cheek skeleton and flesh at a predetermined position. It is a flowchart for demonstrating the face classification | category and cheek makeup procedure using the classification | category determination map in this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Face classification device 11 Input device 12 Output device 13 Drive device 14 Auxiliary storage device 15 Memory 16 CPU
17 Network connection device 18 Imaging device 19 Recording medium 20 Memory 21, 30 Face image 22, 23, 31, 32, 33 Line 24 Area 34 Front face image 35 Horizontal line 36 Vertical line 41-47 Line 51 Teak shape 52 Inscribed ellipse 53, 54 Midpoint 61 Makeup Brush
62 Cut surface

Claims (8)

In the face classification device for cheek makeup that classifies the cheek length, skeleton and fleshiness of the subject,
Photographing means for photographing a face image of the subject;
Control means for performing control for positioning to any one of at least four types classified in advance based on at least two indices of cheek length and skeleton and fleshiness of the whole cheek face from the captured face image; ,
A face classification apparatus comprising: a display unit configured to display a classification result classified by the control unit in association with a map in which the four types are positioned in each quadrant of the plane coordinate space. The control means includes
The face classification apparatus according to claim 1 , wherein an average face set in advance and each of the four types of representative faces are used to evaluate the face of the subject. The control means includes
Based on the average face, the vertical line drawn from just below the pupil to the outline of the face is the vertical width and the horizontal line from the side of the nose to the outline of the face is the horizontal width, and the ratio of the vertical width to the horizontal width of the cheek 3. The face classification apparatus according to claim 2 , wherein the skeleton / feeling of flesh is evaluated based on a degree to which the cheek bone is conspicuously slender and linear or a degree to which the cheek bone is not noticeable and fluffy. The display means includes
In order to approximate the cheek length, skeleton, and fleshiness of the average face based on the classified face shape, the cheek makeup is used to determine how the cheek length, skeleton, and fleshiness look for each of the four types. in the face classification apparatus according to any one of claims 1 to 3, characterized in that to display an adjustment method for adjusting. The cheek makeup is
5. The face according to claim 4 , wherein the starting point for inserting a cheek is an intersection of a line connecting anticular from the tip of the nose, a horizontal line drawn from the center of the eye, and a line drawn from the point intersecting the outline of the face to the mouth corner Classification device. The range of the cheek blur is a first line drawn to the tip of the nose from a point where a horizontal line passing through the center of the eye intersects the outline of the face, a second line parallel to the mouth angle from the first line, A parallelogram surrounded by a third line drawn vertically from the corner of the mouth and a fourth line obtained by moving the third line in parallel until the horizontal line passing through the center of the eye intersects the contour of the face The face classification device according to claim 5 , wherein the face classification device is set based on Computer, face classification program to function as each means having a face classification apparatus according to any one of claims 1 to 6. A computer-readable recording medium on which the face classification program according to claim 7 is recorded.