Contour detection device, printing device, contour detection method, and recording medium

Abstract

Provided is a contour detection device, printing device, contour detection method, and recording medium that have a high success rate in detecting the contour of a detected object from an image. The contour detection device includes: a contour detection unit that detects a first nail contour that outlines a nail from a finger image that contains the nail by performing matching with a nail contour model; and a reception unit that receives input of a second nail contour from a user that corresponds to the first nail contour. The contour detection unit groups nail contours for which input was received by the reception unit, based on information related to the size of the first nail contour, and derives difference information that indicates the difference between the first nail contour and the second nail contour that was input by the user via the reception unit.

Description

TECHNICAL FIELD

[0001] The present application relates to a contour detection device, a printing device, a contour detection method, and a recording medium. BACKGROUND

[0002] Conventionally, a method of detecting a contour of a detection target from an image obtained by photographing the detection target is known.

[0003] For example, in a case where a person is photographed with a camera, by using a contour detection technique, it is possible to detect a contour of a face and contours of eyes, a nose, a mouth, and the like of the person from a photographed image.

[0004] Further, the object of the contour detection is not limited to the face and the parts of the face, and the contour detection technique can be used in detection of various contours such as a contour shape of a nail.

[0005] As such a contour detection technique, there is a method of collecting a plurality of learning samples to perform learning, generating learning data as a result of the learning, and performing contour detection using the learning data.

[0006] For example, as a method conventionally used in contour detection, there are AAM (Active Appearance Model) and ASM (Active Shape Model). These methods express a configuration of feature points of a contour of a face and contours of parts by a model called a shape model. Further, by fitting the shape model to an image including a detection target, contour detection of the detection target is performed.

[0007] Further, in Non-Patent Literature 1, a technique of detecting a contour of a detection target by an algorithm called ESR (Explicit Shape Regression) is disclosed.

[0008] In the ESR, a shape model (initial shape) in which feature points are arranged around a center of gravity is also generated, and fitting thereof to an image including a detection target is performed. At this time, in the ESR, as described in Non-Patent Literature 1, two-stage weak regressors (weak recognizers) are combined and applied, and contour detection is performed as a regression problem of causing the shape model (initial shape) to gradually converge toward a contour of the detection target as a correct position.

[0009] PRIOR ART DOCUMENTS

[0010] NON-PATENT LITERATURE

[0011] Non-Patent Literature 1: Xudong Cao, Yichen Wei, Fang Wen, Jian Sun "Face alignment by Explicit Shape Regression." CVPR 2012: 2887-2894 SUMMARY

[0012] PROBLEMS TO BE SOLVED BY THE INVENTION

[0013] However, as described above, the contour detection method such as ESR detects the contour based on learning. Therefore, in a case where a contour of a detection object of a shape for which a small number of learning samples are available (for example, a nail in a rare shape or the like) or a contour of a detection object of which a contour portion is unclear (for example, a nail of which a boundary of a nail epithelial portion is blurred or the like) is detected, there is a problem that a success rate of detection significantly decreases.

[0014] The present invention has been made in view of the above-described circumstances, and is advantageous in that a contour detection device, a printing device, a contour detection method, and a recording medium in which a success rate of detecting a contour of a detection object from an image is high are provided.

[0015] MEANS FOR SOLVING THE PROBLEMS

[0016] To solve the above-described problems, a contour detection device of the present invention is characterized by comprising: a contour detection mechanism that detects a first nail contour that outlines a region of a nail from a finger image containing the nail by performing matching with a nail contour model; and a reception mechanism that receives input of a second nail contour from a user corresponding to the first nail contour; the contour detection mechanism groups nail contours in the first nail contour for which input has been received by the reception mechanism based on information about a size of the first nail contour, and derives difference information that indicates a difference between the first nail contour and the second nail contour input by the user through the reception mechanism.

[0017] EFFECT OF THE INVENTION

[0018] According to the present invention, a contour of a detection object can be detected from an image with a high success rate. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a perspective view that shows an external structure of the contour detection device of the present embodiment.

[0020] Figure 2A is a perspective view that shows a finger arrangement section, Figure 2B is a plan view that shows the finger arrangement section from above.

[0021] Figure 3 is a main part block diagram that shows a functional structure of the contour detection device of the present embodiment.

[0022] Figure 4 is a diagram schematically showing a general nail contour process.

[0023] Figure 5A is a diagram showing an example of a finger image, Figure 5B is a diagram showing an example of a finger region and a nail region detected from Figure 5A Figure 5C is a diagram showing an example of a display in which a first nail contour is superimposed and displayed on Figure 5A

[0024] Figure 6 is a flowchart showing a contour detection process of the present embodiment.

[0025] Figure 7 is a flowchart showing an automatic-manual combined correction process of a nail contour of the present embodiment.

[0026] Figure 8 is an explanatory diagram showing a first nail contour and a second nail contour.

[0027] Figure 9 is a flowchart showing a sample making classification process of the present embodiment.

[0028] Figures 10A to 10F is a diagram showing an example of an erroneous first nail contour.

[0029] Figure 11 is a diagram showing an example of a result after grouping the erroneous first nail contours of Figures 10A to 10F

[0030] Figure 12 is a flowchart showing an automatic correction process of a nail contour of the present embodiment.

[0031] Figure 13 is a diagram showing a case in which a feature point of an erroneous first nail contour is forcibly moved to a correct feature point position.

[0032] Figure 14 is a diagram showing a case in which a feature point of an erroneous first nail contour is forcibly moved to a correct feature point position.

[0033] Figure 15 is a block diagram showing a control structure of a printing device of the present embodiment. DETAILED DESCRIPTION

[0034] [1st Embodiment]

[0035] Reference is made to Figures 1 to 14 ​​​A first embodiment of the profile detection apparatus, the profile detection method, and the recording medium of the present application will be described.

[0036] In the following embodiments, the case where the object to be detected, i.e., the object of profile detection, is a fingernail will be described as an example.

[0037] Various limitations that are technically preferable for implementing the present application will be added below, but the scope of the present application is not limited to the following embodiments and illustrated examples.

[0038] Figure 1 is a perspective view showing the external appearance of the profile detection apparatus of the present embodiment.

[0039] As shown in Figure 1 , the profile detection apparatus 1 of the present embodiment has a box body 11 that is generally formed in a box shape.

[0040] An input portion 12 is provided on the upper surface (top plate) or the like of the box body 11.

[0041] The input portion 12 is a functional portion in which the user makes various inputs.

[0042] On the input portion 12, for example, an operation button for turning on the power of the profile detection apparatus 1, a stop switch button for stopping the operation, a detection start button for instructing the start of profile detection of the fingernail T, and the like are arranged to make various inputs.

[0043] In the present embodiment, as will be described later, a touch panel 121 is integrally formed on the surface of the display portion 13, and the touch panel 121 also functions as the input portion 12.

[0044] For example, a fingernail profile (detection fingernail profile, which will be referred to as "first fingernail profile To1" hereinafter) that is automatically detected by the apparatus side from a finger image FI obtained by photographing the fingers U (see Figure 5A , Figure 5C , and the like) is displayed on the display portion 13, and the user confirms the first fingernail profile To1 on the screen, and can confirm whether the first fingernail profile To1 represents an appropriate fingernail profile To.

[0045] Also, in a case where the first nail profile To1 indicates a line in error (i.e., in a case where the first nail profile To1 does not indicate an appropriate profile), the user can input a second nail profile To2 by delineating a correct profile of the nail T (a line of which an inner side region is intended by the user to be recognized as "nail T", a correct nail profile, hereinafter referred to as "second nail profile To2") on the touch panel 121. Note that the input of the second nail profile To2 is not limited to the case using the touch panel 121. Various pointing devices not shown such as a mouse or a trackball can also be used for the input.

[0046] Thus, in the present embodiment, the input section 12 such as the touch panel 121 functions as a reception mechanism that receives the input of the second nail profile To2 from the user.

[0047] Further, a display section 13 is provided on the upper surface (top plate) of the housing 11.

[0048] The display section 13 is, for example, a display mechanism constituted by a liquid crystal display (LCD), an organic electroluminescence display, or another flat panel display.

[0049] In the present embodiment, on the display section 13, for example, various indications and the like are appropriately displayed, a nail image (a finger image including an image of the nail T) obtained by photographing the finger, an image of a profile line of the nail T included in the nail image, an indication screen displaying various indications, and the like are displayed.

[0050] Further, on the surface of the display section 13 of the present embodiment, the touch panel 121 functioning as the input section 12 as described above is integrally constituted.

[0051] Further, on the front side (in the Figure 1 , the front side) of the housing 11, an opening section 14 for inserting the finger U corresponding to the nail T as a detection object at the time of photographing by the profile detection apparatus 1 and placing it on a photographable position that can be photographed by the photographing section 5 is formed.

[0052] Inside the opening section 14, a finger arrangement section 3 that fixes the nail T (including the finger U of the nail T) as a detection object in the present embodiment is arranged.

[0053] Figure 2A is a perspective view of the finger arrangement section 3, Figure 2B is a plan view of the finger arrangement section 3 viewed from above.

[0054] In Figure 2A , a case where the finger U including the nail T is arranged inside the finger arrangement section 3 is indicated by a double-dotted line.

[0055] like Figure 2A As shown, the finger arrangement part 3 is a box-shaped component with an opening 31 near the front side of the device, and a finger support member 32 for supporting the finger U is disposed inside the finger arrangement part 3. The finger arrangement part 3 is disposed inside the housing 11 at a position where the opening 31 corresponds to the opening 14 of the housing 11.

[0056] The finger support member 32 pushes and supports the finger U from below, and is made of, for example, a soft resin. Furthermore, the structure of the finger support member 32 is not particularly limited as long as it can support the finger U from below. For example, it can be supported from below by an elastic member such as a spring. Additionally, for example, the finger support member 32 can be configured to expand and contract by changing the internal pressure, and in the expanded state, push the finger U up and fix its position.

[0057] The top surface of the finger arrangement section 3 has an open window 33. The fingernail T of the finger U inserted into the finger arrangement section 3 protrudes from the window 33.

[0058] Furthermore, the top surface of the finger placement part 3 near the front becomes a finger pressing part 34 to prevent the finger U from lifting and to limit the upper position of the finger U. The finger U and its nail T are supported from the bottom by the finger support member 32, and the upper side of the finger U is pressed by the finger pressing part 34, thereby positioning the height position at a predetermined position.

[0059] Furthermore, in this embodiment, a nail-holding portion 35 for holding a nail T is provided on the inner side in the finger insertion direction. The nail-holding portion 35 is a positioning mechanism for positioning the nail T, which is the object to be detected.

[0060] By placing the front part of the nail T on the nail mounting portion 35, the horizontal position (i.e., the X and Y directions) of the nail T is defined, and its vertical position is also restricted.

[0061] Figure 3 This is a block diagram showing the main functional structure of the contour detection device in this embodiment.

[0062] like Figure 3 As shown, the contour detection device 1 includes a camera unit 5.

[0063] The camera unit 5 includes a camera device 51 and a lighting device 52, and is positioned above the position where the fingernail T is positioned when the finger U is inserted into the finger placement unit 3.

[0064] The shooting device 51 is, for example, a small camera equipped with a solid-state imaging element and a lens having approximately 2 million pixels or more.

[0065] The lighting device 52 is, for example, a white LED light. In this embodiment, multiple lighting devices 52 are arranged to surround the shooting device 51.

[0066] Furthermore, the positions of the imaging device 51 and the lighting device 52 are not limited to those shown in the figure. For example, the imaging device 51 and the lighting device 52 of the imaging unit 5 may be fixedly positioned above the nail T. If the imaging unit 5 is configured to be movable by a moving mechanism, it is acceptable as long as it can be moved to a position above the nail T.

[0067] In this embodiment, the imaging unit 5 is a finger image acquisition mechanism that acquires a finger image FI including the fingernail T by capturing an image of the finger U. The imaging unit 5 captures an image of the finger U, which is the detection object, i.e., the fingernail T, positioned in the finger placement unit 3, and acquires the finger image FI.

[0068] Specifically, the imaging device 51 can capture the area corresponding to the window 33 of the finger placement unit 3, and can include the finger U and nail T exposed from the window 33 within the field of view. Since the finger U is positioned by the finger placement unit 3, by capturing the area corresponding to the window 33, the positional relationship of the finger U and nail T presented in the finger image (e.g., the front end of the first joint of the finger is included within the field of view) is approximately constant.

[0069] Furthermore, the contour detection device 1 of this embodiment includes a control device 80.

[0070] The control device 80 may be a substrate (not shown) disposed on the lower surface side of the top surface of the housing 11.

[0071] The control device 80 functions as a processor and is a computer comprising a control unit 81 including a CPU (Central Processing Unit) not shown, and a storage unit 82 including ROM (Read Only Memory) and RAM (Random Access Memory) (both not shown).

[0072] The storage unit 82 is provided with a program storage area 820 that stores various programs such as the contour detection program 821 that can be executed by the processor to operate the contour detection device 1.

[0073] Furthermore, in this embodiment, the storage unit 82 is provided with a learning data storage area 822, a contour information storage area 823, etc. The learning data storage area 822 stores the learning data of contour detection used by the contour detection unit 813 to detect the nail contour from the image, and the contour information storage area 823 stores the detection results detected by the contour detection unit 813, etc.

[0074] The learning data storage area 822 is a learning data storage mechanism that stores the initial shape To0 (refer to) which is composed of multiple feature points N and is used for the detection of the nail contour To, as a learning result generated based on the image of the learning object. Figure 4 (Hereinafter referred to as "nail contour model") data.

[0075] In the learning data storage area 822, the coordinate values ​​(Xn0, Yn0, Xn1, Yn1, ... Xnn, Ynn) of each feature point N0 to Nn of the contour that constitutes the initial shape To0 (nail contour model), the position of the centroid (not shown) of the initial shape To0, and the regressor (recognizer) composed of the displacement function (regression function) are stored, which are obtained through prior learning.

[0076] The training data is obtained by feature extraction, which extracts features useful for contour detection from the training images. These features include, for example, the brightness value of each pixel, its color, the brightness values ​​between adjacent pixels, and the amount of color variation. After feature extraction, X and Y coordinate values ​​of feature points representing the shape of the nail region are generated for each training image, and a correspondence is established between these X and Y coordinate values ​​and each training image.

[0077] Furthermore, for each learning image, pattern learning is performed using data containing the X and Y coordinate values ​​of corresponding feature points. This generates learning data consisting of an initial shape To0 (the shape similar to the detected object, i.e., the fingernail T), the coordinate values ​​of the feature points N constituting the contour of the initial shape To0, the centroid of the initial shape To0, and a recognizer, which serves as the learning result. Additionally, learning is typically performed using machine learning.

[0078] The initial shape To0 and other learning data stored in the learning data storage area 822 are generated by repeatedly performing contour detection in advance and are stored in the storage unit 82 (learning data storage area 822 within the storage unit 82) of the contour detection device 1 in its initial state when the product leaves the factory.

[0079] As will be described later, in this embodiment, based on the learning data as a learning result, the contour detection unit 813 detects the contour of the nail T, which is the object of detection.

[0080] Functionally, the control unit 81 includes a shooting control unit 811, a display control unit 812, and a contour detection unit 813. These functions, such as shooting control unit 811, display control unit 812, and contour detection unit 813, are achieved through cooperation between the CPU of the control unit 81 and the program stored in the ROM of the storage unit 82.

[0081] The shooting control unit 811 controls the shooting device 51 and the lighting device 52 of the shooting unit 5, so that the shooting device 51 captures an image of the finger U arranged on the finger arrangement unit 3, including the image of the nail T (finger image FI).

[0082] The display control unit 812 controls the display screen displayed by the display unit 13, causing it to perform various displays. In this embodiment, as described above, the display screen displays a finger image FI, or an image with a finger outline Uo, nail outline To, etc. superimposed on the finger image FI, or displays a screen requesting the user's approval of the first nail outline To1.

[0083] The contour detection unit 813 is a contour detection mechanism that detects the nail contour To of the region depicting the nail T from the finger image FI by performing matching (fitting) with the initial shape To0.

[0084] Furthermore, in this embodiment, the contour detection unit 813 also detects the finger contour Uo of the region depicting the finger U from the finger image FI. The detection of the finger contour Uo is performed, for example, using a background subtraction method. That is, for example, a preliminary image is obtained by pre-capturing the window 33 of the finger placement unit 3, which is in a state where no finger U is set, in a way that the entire area is included in the field of view by the imaging unit 5. And when the finger image FI is captured, the finger image FI is compared with the preliminary image, and the part that is not in the preliminary image is detected as the finger U.

[0085] The contour detection unit 813 detects the contour of the fingernail T (coordinate values ​​of feature points constituting the contour) based on the image of the fingernail T of the finger U captured by the imaging device 51 (finger image FI).

[0086] In this embodiment, the contour detection unit 813 uses the ESR (Explicit Shape Regression) method to detect the contour of the nail T.

[0087] That is, the contour detection unit 813 performs a match between the initial shape To0 (nail contour model) stored in the learning data storage area 822 of the storage unit 82 and the image containing the detection object.

[0088] In contour detection using ESR, as described in Non-Patent Document 1, a two-stage weak regressor (weak recognizer) is combined and applied to perform contour detection as a regression problem that gradually converges the initial shape To0 (fingernail contour model) toward the contour of the detection object as the correct position.

[0089] Figure 4This is a schematic diagram illustrating the detection process of nail contour To using the ESR algorithm (i.e., the process of deforming the initial shape To0 (nail contour model) based on the learning data to approach the nail contour To).

[0090] Typically, in the detection of nail contour To using the ESR algorithm, by repeatedly performing detection loops by gradually moving the feature points N of the initial shape To0 (nail contour model), it is possible to gradually move towards the second nail contour To2 (reference). Figure 8 )near.

[0091] Specifically, when the contour detection unit 813 receives a finger image FI containing the nail region as input, such as Figure 4 As shown on the left, the initial shape To0 is configured to the appropriate position (initial position) within the image FI. Additionally, [the following will be done / will be done / etc.]. Figure 4 The nail epithelium located between the skin of the nail T and the finger U is defined as the nail epithelium region nt. The nail epithelium region nt is an area with unclear and irregular boundaries, which is the part that is prone to false detection of the nail outline To.

[0092] In addition, contour detection methods that perform contour detection as a regression problem, such as ESR, are more robust than methods such as AAM that perform contour detection by matching shape models. Therefore, regardless of what shape is used as the initial shape To0, or even if the initial position of the initial shape To0 is not set so strictly, the impact on the accuracy of the detection results is also small.

[0093] Therefore, even when the input image FI is different, a common initial shape To0 can still be applied.

[0094] However, when using methods such as ESR, higher accuracy in contour detection can be expected if the initial shape To0 is configured to coincide with the nail area as much as possible.

[0095] Therefore, in this embodiment, it is preferable to configure the initial shape To0 such that the center of gravity of the initial shape To0 is the position of the average center of gravity of the learning data.

[0096] If an initial shape To0 is configured, the contour detection unit 813 uses a regressor (a two-stage weak recognizer in the case of ESR) composed of a displacement function (regression function) generated through prior learning and pre-stored in the contour information storage area 823 of the storage unit 82, to repeatedly perform displacement of the initial shape To0, as the orientation... Figure 4 On the right side, it gradually converges toward the outline of the nail T (nail outline To, i.e., the correct boundary of the nail area) which is the correct position of the nail.

[0097] Next, refer to Figure 5 to... Figure 14 The contour detection method performed by the contour detection device 1 of this embodiment will be described.

[0098] like Figure 6 As shown, when performing contour detection of the nail T, firstly, the finger U arranged on the finger arrangement unit 3 is captured by the imaging unit 5 to obtain the finger image FI (step S1). Figure 5A This is a diagram representing an example of image FI. For example... Figure 5A As shown, image FI is the image that will be compared with... Figure 2B The image shown is obtained by taking a picture of the portion corresponding to the window 33 of the reference configuration unit 3.

[0099] As the contour detection unit 813, the control unit 81 reads a previously acquired image of a state in which no finger U is mounted from the storage unit 82, and detects the part that is not in the previous image as a finger U by comparing it with the finger image FI using a background difference method. Then, it detects the finger contour Uo (hereinafter also referred to as "finger region") of the region depicting the detected finger U from the finger image FI (step S2). Figure 5B The left side indicates from Figure 5A The image shown is an example of a finger region detected in the image FI.

[0100] Furthermore, as the contour detection unit 813, the control unit 81 detects the nail contour To from the finger image FI (step S3, in which the nail contour in this case is also referred to as "the first nail contour To1").

[0101] The detection of the nail outline To here refers to applying the initial shape To0, which is stored as learning data in the learning data storage area 822, to the finger image FI, and using the ESR algorithm to detect the part that is similar to the nail outline To.

[0102] Figure 5B The right side indicates from Figure 5A The example shown is where the nail outline To is correctly detected in the image FI.

[0103] like Figure 5A As shown in the shaded area, on the underside of the T-zone of the nail (hereinafter also referred to as the "nail area") Figure 5A The lower side (the side at the root of the nail T) contains the nail plate area (nt). For example... Figure 5B As shown, the nail region refers to the area of ​​the nail T outside of the nail epithelium region nt.

[0104] Nail outline To1 as shown Figure 5C As shown, it is a collection of multiple feature points N (N0~Nn).

[0105] For example, the point located on the nail T closest to the nail tip is designated as feature point N0 (coordinates: Xn0, Yn0), and from there counterclockwise, feature points N1 (coordinates: Xn1, Yn1), N2 (coordinates: Xn2, Yn2), ..., Nn (coordinates: Xnn, Ynn). Furthermore, the number of feature points N constituting the first nail contour To1 is not particularly limited.

[0106] Upon detecting the first nail outline To1, the control unit 81, as the outline detection unit 813, prompts the user with the detection result of the first nail outline To1 and requests approval (step S4). The method of prompting is not particularly limited. For example, ... Figure 5C As shown, the first nail outline To1 is superimposed on the finger image FI and displayed. Alternatively, it is preferable to display a message on the display screen of the display unit 13 asking whether the user approves the detection result, i.e., the first nail outline To1. However, the method for requesting user approval is not limited to this. For example, it could also be a voice-based message.

[0107] As the contour detection unit 813, the control unit 81 determines whether the user has approved the detection result, that is, whether the user has made an input from the input unit 12 or the like indicating approval of the first nail contour To1 as the detection result (e.g., pressing the OK button, etc.) (step S5).

[0108] If the user approves the first nail outline To1 (in step S5), the approved detection result (the first nail outline To1) is determined as the nail outline To of the nail T (in step S6), and the outline detection process ends.

[0109] On the other hand, if the user does not approve the first nail outline To1 (No in step S5), the control unit 81, as the outline detection unit 813, further determines whether a certain number or more error samples (error samples generated based on the erroneous first nail outline To1 (i.e., the first nail outline To1 not approved by the user) and the second nail outline To2, as described later) have been stored (step S7). Furthermore, if a certain number or more error samples have not been stored (No in step S7), an automatic-manual correction process is performed, that is, a process that automatically corrects the first nail outline To1 and a process that requests the user to manually correct it is performed (step S8).

[0110] Here, refer to Figures 7 to 11 Automatic and manual correction processing of nail outline ( Figure 6 The details of step S8) will be explained.

[0111] In automatic and manual combined correction processing, such as Figure 7As shown, firstly, the display unit 13 displays a message urging the user to manually draw what they consider to be the correct nail outline (referred to as "second nail outline To2") (step S11).

[0112] The user accepts the message and inputs the second nail outline To2 from the touch panel 121, etc. Specifically, the user draws a line using a pen or fingertip at the position they deem correct.

[0113] When the user inputs the second nail outline To2, the control unit 81, which is the outline detection unit 813, obtains the information of the second nail outline To2 (step S12).

[0114] Figure 8 This is a schematic diagram illustrating user input of the second nail outline To2. When the second nail outline To2 is input, even if the user inputs the second nail outline To2 as a line, the outline detection unit 813 will treat it as a line. Figure 8 The set of characteristic points R that constitute the line, as shown, is used to grasp the concept.

[0115] Regarding the feature points R of the second nail contour To2, for example, the point located on the nail T closest to the nail tip can be designated as feature point R0 (coordinates: Xr0, Yr0), and from there counterclockwise designated as feature points R1 (coordinates: Xr1, Yr1), R2 (coordinates: Xr2, Yr2), ..., Rn (coordinates: Xrn, Yrn). The number of feature points R constituting the second nail contour To2 is not particularly limited, just as the number of feature points N constituting the first nail contour To1 is not.

[0116] In addition, Figure 8 In the diagram, feature point N constituting the first nail outline To1 is represented by a black dot, and feature point R constituting the second nail outline To2 is represented by a dot with a diagonal shading. Furthermore, for ease of explanation, feature points N and R are represented by dots of a larger size than their actual dimensions.

[0117] If information about the second nail contour To2 is obtained, the contour detection unit 813 generates error samples based on the erroneous first nail contour To1 (i.e., the first nail contour To1 that is not approved by the user) and the second nail contour To2 (step S13), and appropriately groups the error samples (step S14).

[0118] Next, the obtained second nail contour To2 (the nail contour input by the user) is determined as the nail contour To of the nail T (step S15), and the contour detection process ends.

[0119] Next, refer to Figure 9 For the creation of error samples ( Figure 7Step S13) and error sample grouping (classification) processing ( Figure 7 Step S14) explains its details.

[0120] First, the contour detection unit 813 calculates the area of ​​the nail region (the area depicted by the first nail contour To1) in the erroneous detection results and the area of ​​the finger region (the area depicted by the first nail contour To1). Figure 6 The ratio of the area of ​​the region depicted by the finger contour Uo detected in step S2 (step S21). That is, to calculate Figure 5B The area ratio shown is the ratio of the area of ​​finger U to the area of ​​nail T. Here, the "area ratio" of the area of ​​finger U to the area of ​​nail T is information related to the size of the area depicted by the first nail outline To1.

[0121] Furthermore, as information related to the size of the nail region (the region depicted by the first nail contour To1) in the erroneous detection results, the contour detection unit 813 calculates the aspect ratio (step S22). That is, it calculates... Figure 5A The ratio of the longitudinal dimension H to the transverse dimension W of the inner region of the nail outline To, i.e., the nail area.

[0122] Furthermore, the contour detection unit 813 compares the incorrect nail contour (the first nail contour To1 as the detection result) with the correct nail contour (the second nail contour To2 input by the user) to obtain difference information. This difference information indicates the degree of difference (error, i.e. the correction range or adjustment range used to make the incorrect first nail contour To1 become the second nail contour To2), and represents the difference (error) between the two (step S23).

[0123] Here, the difference information includes, for example, the information indicating the feature point with the largest error relative to the feature point constituting the first nail contour To1 and the magnitude of the difference (error).

[0124] For example, in Figure 8 In the example shown, the feature points near the root of the nail T (in) Figure 8 In the middle is the point (in Nq (coordinate values; Xnq, Ynq)) that is furthest from the feature point R that constitutes the second nail contour To2 and corresponds to it (in Figure 8 In this context, Rq (coordinates; Xrq, Yrq) represents the feature point with the largest error. Furthermore, the magnitude (correction margin, the degree of separation between points) of the difference (error) between Nq (coordinates; Xnq, Ynq) and Rq (coordinates; Xrq, Yrq) is the "distance d".

[0125] Furthermore, regarding which feature point is set as the feature point with the largest error, it can be automatically set by the control unit 81, which is the contour detection unit 813, or it can be set by the user through touch panel operation or by specifying the area considered to be the farthest.

[0126] Furthermore, the difference information is not limited to the magnitude of the difference (error). For example, the difference information can also include information such as the direction of deviation (vector).

[0127] Furthermore, the order of steps S21 (obtaining the area ratio), S22 (obtaining the aspect ratio), and S23 (obtaining the difference information) is not limited to the example shown here. For example, the aspect ratio can be obtained before the area ratio, thus reversing the processing order.

[0128] Furthermore, the process of calculating the area ratio of the finger region to the nail region (step S21) is not necessary. Alternatively, the aspect ratio and difference information of the nail region can be used only for the purpose of creating error samples and grouping (classifying) error samples.

[0129] In this embodiment, the contour detection unit 813 establishes a correspondence between the area ratio, aspect ratio, and difference information to create error samples (step S24). By establishing a correspondence between the area ratio, aspect ratio, and difference information, it is possible to generate samples that show the tendency of different sizes and shapes of nails T to produce different types of errors in the detection results.

[0130] Furthermore, in this embodiment, as differential information, information representing the maximum error feature point and the magnitude of its error is stored in correspondence with the area ratio and aspect ratio. This facilitates comparisons and other processing in subsequent steps and reduces the amount of data stored in the storage unit 82. Additionally, the differential information is not limited to information about the maximum error feature point; it may also include lines representing the outline of the nail T input by the user as the second nail outline To2, or the coordinate values ​​of feature points R that constitute part or all of the second nail outline To2.

[0131] Furthermore, the contour detection unit 813 can also attempt to group error samples based on area ratio and aspect ratio (step S25).

[0132] Figures 10A to 10F This is an example of a detection sample of a nail profile collected as an error sample. Furthermore, Figure 11 It is Figures 10A to 10F A classification example with groups of 3.

[0133] The result of classifying error samples based on area ratio and aspect ratio is: Figure 11 The outline of group A shown is the first nail profile To1, located near the boundary between the nail epithelium region nt of nail T and the skin portion of the finger. Figure 10A ,Figure 10C It is classified here. Furthermore, group B consists of the first nail profile To1, whose outline lies within the nail epithelial region nt of the nail T. Figure 10B , Figure 10E They are classified here. Furthermore, group C is characterized by the first nail outline (To1) being located lower than the nail epithelial region of the nail (T) due to factors such as wrinkles in the fingernail skin. Figure 10D , Figure 10F It is categorized here.

[0134] While area ratio and aspect ratio do not directly affect the position of the outline, the shape characteristics of the fingernail U and nail T are often similar when their area ratio and aspect ratio are close. Figure 11 As shown in the example, error samples that are similar in terms of error are ultimately classified into the same group. Furthermore, in grouping, not only area ratio and aspect ratio can be considered, but also difference information. For example, by grouping based on the magnitude of the error (difference), the mode of deviation (vector information of the direction of deviation), etc., more detailed classification can be achieved.

[0135] In addition, error samples are preferably grouped according to each index type.

[0136] Even among the same user, nail shapes vary considerably depending on the type of finger, such as the thumb and little finger. Therefore, grouping nails by finger type allows for more accurate detection of the nail contour (To).

[0137] In the grouping, the contour detection unit 813 determines whether the newly generated error sample can be classified into an existing group, that is, determines whether there is a corresponding group with similar tendencies (step S26).

[0138] Furthermore, if a corresponding group exists (as in step S26), the error sample is classified into the corresponding group (step S27), and the error (difference) is averaged within that group and stored in the storage unit 82 along with the area ratio and aspect ratio (step S28). For example, the maximum error feature points of the first nail contour To1 and the second nail contour To2 are also averaged. For example, the average error (difference, correction magnitude) between the averaged maximum error feature point of the first nail contour To1 and the averaged maximum error feature point C of the second nail contour To2 is the average error ad (refer to...). Figure 13 , Figure 14 In the case of averaged differential information, the contour detection unit 813 establishes and stores the coordinate values ​​(Xc, Yc) of the maximum error feature point C and the value of the average error ad with the area ratio and aspect ratio.

[0139] On the other hand, if there is no corresponding group (No in step S26), the contour detection unit 813 creates a new group (step S29) and classifies the error sample into the new group (step S30).

[0140] In addition, return Figure 4 When a certain number of error samples have been stored (in step S7), that is, when the above error samples have been repeatedly created and classified during the multiple detection processes of the nail contour To, and the grouped error samples have been stored, the process of automatically correcting the first nail contour To1 is performed (in step S9).

[0141] Alternatively, before determining whether a certain number of error samples have been stored, the area ratio and aspect ratio of the nail T newly added to the nail contour To can be calculated to determine which group the nail T belongs to. The processing can then be differentiated based on whether a certain number of error samples belonging to that group have been stored. There are also cases where, although error samples are stored overall, the number of samples varies depending on the group. In such cases, automatic and manual correction processing of the nail contour can be considered. Figure 6 The detection accuracy increases when the process is carried out in step S8).

[0142] Here, refer to Figures 12 to 14 Automatic correction of nail outline ( Figure 6 The details of step S9 will be explained.

[0143] like Figure 12 As shown, when the new first nail contour To1 is subjected to automatic correction processing, firstly, the contour detection unit 813 searches for feature point A among the feature points constituting the first nail contour To1, which corresponds to the coordinate value (Xc, Yc) of the maximum error feature point C of the averaged second nail contour To2 (step S31).

[0144] With feature point A set, the position of feature point A for each detected tap is set to (Xa(n), Ya(n)). Each time the detection loop is completed once, the loop counter n is incremented (step S32).

[0145] The contour detection unit 813 determines whether the loop counter n is less than or equal to the preset total number of detection loops N (step S33). If the loop counter n exceeds the preset total number of detection loops N (no in step S33), the automatic correction process for the nail contour ends and returns to normal. Figure 4 Step S4 involves displaying the nail contour detection results to the user and requesting their approval, repeating this process for subsequent steps. The number of detection cycles is arbitrary; for example, the total number of detection cycles N = 1000.

[0146] On the other hand, if the loop counter n is less than or equal to the preset total number of detection loops N (as in step S33), the contour detection unit 813 further determines whether the loop counter n is between the number of loops M (e.g., M = 500) in the initial stage (step S34). If the loop counter n is less than or equal to M (as in step S34), it further determines whether the loop counter n is an arbitrary number of loops L (e.g., 100 loops each time, such as L = 100, 200, 300...) (step S35). And, if it is exactly the Lth loop (as in step S35), the nail contour model is forcibly deformed so that point A (coordinates: Xa, Ya) becomes point C (coordinates: Xc, Yc) (step S36).

[0147] That is, the contour detection unit 813 forcibly changes the coordinates (Xa, Ya) of the maximum error feature point A of the first nail contour To1 corresponding to the averaged maximum error feature point C to the coordinates (Xc, Yc) of the maximum error feature point C using the averaged difference information, i.e., the average error ad, which represents the magnitude of the error.

[0148] Figure 13 It is a diagram that schematically represents the situation where point A (coordinates: Xa, Ya) moves to the coordinates of point C (coordinates: Xc, Yc) based on the average error ad.

[0149] like Figure 13 As shown, if point A (coordinates: Xa, Ya) is dragged into point C (coordinates: Xc, Yc), then as follows: Figure 14 As shown, points near point A are also subjected to stretching deformation with the amount of change adjusted proportionally to their distance from point A. Therefore, the nail contour model applied to the first nail contour To1 dynamically deforms, making it easier and faster to approach the second nail contour To2 appropriately.

[0150] In contrast, if the loop counter n exceeds M times (No in step S34), and if it is less than M times (Yes in step S34) but the loop counter n is not an arbitrary number of loops L times (No in step S35), the nail contour To is detected based on the usual ESR algorithm (i.e., the nail contour model is deformed based on the learning data, step S37).

[0151] Based on this learning data, the nail contour model is deformed and processed. Figure 4 The situation shown is the same, so the explanation is omitted here.

[0152] Furthermore, following steps S36 and S37, the process returns to step S33 and repeats until the loop counter n exceeds the predetermined number N. This allows the first nail profile To1 to approximately converge to the correct position along the profile of the nail region.

[0153] If the loop counter n exceeds the specified number of times N, then return to the previous state. Figure 6 Step S4 involves displaying the test results to the user and requesting their approval, repeating this process until approval is obtained.

[0154] By using this method, the outline of the nail T can be automatically detected with high precision without the user having to cut or define the nail area themselves.

[0155] As described above, according to this embodiment, the system includes: an imaging unit 5, which is a finger image acquisition mechanism, for acquiring a finger image FI that captures the finger U and includes the nail T; a learning data storage area 822, which stores data of an initial shape To0, which is generated based on a learning object image and is composed of multiple feature points N and used for detecting the nail contour To; a contour detection unit 813, which detects the nail contour To of the region depicting the nail T from the finger image FI by matching it with the initial shape To0; and an input unit 12, which accepts the input of a correct nail contour To2 from a user who does not approve of the detected nail contour To; the contour detection unit 813, at least based on the aspect ratio of the nail region based on the detected nail contour To1, groups the first nail contour To1 that is not approved by the user, compares the first nail contour To1 with the second nail contour To2 input by the user, stores the difference information representing the difference between the two together with the grouping result, and performs correction processing on the first nail contour To1 using the difference information in subsequent detections.

[0156] Therefore, even if the shape is characteristic, or the nail T is difficult to detect due to the difficulty in identifying the nail outline To in the nail epithelium region nt, it can gradually approach the second nail outline To2 that the user desires.

[0157] Furthermore, the detection results (such as grouped error samples) are stored each time a detection process is performed, thereby gradually customizing the contour detection device 1 to match the characteristics of the user's nail T, and also reducing the time the user has to manually input the second nail contour To2.

[0158] Furthermore, the contour detection unit 813 of this embodiment also detects the finger contour Uo of the region depicting the finger U from the finger image FI, and performs a first nail contour To1 grouping based on the area ratio of the finger region based on the finger contour Uo to the nail region based on the nail contour To.

[0159] Therefore, it is possible to more appropriately classify the first nail profile To1, which corresponds to the characteristic shape of the nail T.

[0160] Furthermore, in this embodiment, the difference information is stored by averaging each group.

[0161] Therefore, it easily reflects differential information.

[0162] Furthermore, the differential information in this embodiment includes information indicating the maximum error feature point A among the feature points N constituting the first nail contour To1 that has the largest difference (error) relative to the second nail contour To2, and the magnitude of the difference (error).

[0163] Therefore, the first nail profile To1 can be effectively corrected based on the maximum error feature point A and its difference (error).

[0164] Furthermore, the contour detection unit 813 of this embodiment repeatedly performs correction processing on the first nail contour To1 that is not approved by the user. After a certain number of the grouped first nail contours To1 are stored, the contour detection unit 813, in the correction processing, forcibly changes the coordinate value of the maximum error feature point A using information representing the magnitude of the difference (error) according to any number of times.

[0165] After a certain number of error samples have been stored, appropriate corrections reflecting the user's intent can be made even without input of the second nail profile To2 by the user.

[0166] [Second Implementation]

[0167] Next, refer to Figure 15 As a second embodiment, an example is described in which the contour detection device of the present invention is applied to a printing device (nail printing device) for nail art design.

[0168] Furthermore, in this embodiment, the structure, function, and effect of the contour detection device are the same as those described in the first embodiment, so detailed descriptions of the structure, function, and effect of the contour detection device are omitted below.

[0169] Figure 15 This is a block diagram showing the main part of a functional structure example of a printing apparatus 100 that uses a contour detection device 1.

[0170] like Figure 15 As shown, the printing apparatus 100 includes a printing unit 4 in addition to the structural parts of the contour detection device 1 described in the first embodiment.

[0171] The printing section 4 includes a printing head 41 and a head moving mechanism 42 for moving the printing head 41.

[0172] In this embodiment, the printhead 41 is, for example, a printhead-integrated ink cartridge that has a built-in ink storage section (not shown) and an inkjet ejection mechanism that sprays ink as fine droplets to apply printing to the nail T, which is the object to be printed.

[0173] The print head 41 performs printing by appropriately ejecting the prescribed ink under the control of the printing control unit 814.

[0174] The ink storage section is provided, for example, to correspond to cyan (C); magenta (M); and yellow (Y). Furthermore, the ink that the printhead 41 can eject is not limited to the example shown here; it may also have an ink storage section for storing other colors of ink. Additionally, it may be configured to eject liquids for substrates or coatings.

[0175] The head movement mechanism 42 includes, for example, a stepper motor (not shown), and is configured to move the print head 41 accurately in predetermined steps each time. The movement of the print head 41 by the head movement mechanism 42 is controlled by the printing control unit 814.

[0176] Control unit 81, in addition to the contour detection device 1, Figure 3 In addition to the structure shown, it also includes a printing control unit 814 for controlling the printing unit 4.

[0177] In addition, the storage unit 82, besides relating to the contour detection device 1, Figure 3 In addition to the structure shown, it also features a design storage area of ​​824, etc.

[0178] The design storage area 824 stores image data of nail art designs to be printed on nail T.

[0179] In addition, the storage unit 82 may also store information such as the contour of the nail T (nail shape, XY coordinates of the horizontal position of the nail T, etc.), the tilt angle of the nail T, the nail curvature, and various correction information detected by the contour detection unit 813.

[0180] Furthermore, the other structures are the same as in the first embodiment, so the same reference numerals are assigned to the same components, and their descriptions are omitted.

[0181] Next, the printing control process performed by the printing apparatus 100 of this embodiment will be described.

[0182] In this embodiment, when printing onto a nail T by the printing apparatus 100, the user turns on the power to the apparatus and operates the input section 12, such as the printing start button. This inputs a printing start instruction corresponding to the operation to the control device 80 of the printing apparatus 100.

[0183] When the printing start instruction is entered, the display control unit 812 causes the display unit 13 to display a nail design selection screen, urging the user to make a selection.

[0184] When the user selects a desired nail design from the input section 12, the user then selects the nail design to be printed on the nail according to the selection instructions.

[0185] When a nail design is selected, the display control unit 812 causes the display unit 13 to display an instruction screen that instructs the finger U of the nail T (which is to be printed) to be positioned at a predetermined position on the finger placement unit 3, urging the user to fix the nail T (and its finger U).

[0186] The user inserts the finger U into the finger placement part 3 as instructed, and positions the fingernail by placing the tip of the fingernail T onto the fingernail placement part 35, which serves as a positioning mechanism.

[0187] Furthermore, with the nail T positioned by the nail carrier 35, the shooting control unit 811 activates the shooting unit 5 to capture the nail T and obtain a finger image FI containing the area of ​​the nail T.

[0188] When the finger image FI is acquired, the contour detection unit 813 performs contour detection processing to detect the nail contour To that depicts the nail region from the image. Furthermore, the content of this contour detection processing is the same as that described in the first embodiment, so its description is omitted.

[0189] When the nail outline To of nail T is detected, the selected nail design is matched with the detected nail outline To. Then, appropriate corrections such as surface correction are made to generate printing data (printing data for the nail design).

[0190] Next, the printing control unit 814 outputs printing data to the printing unit 4, and while actuating the head moving mechanism 42 to move the printing head 41 appropriately, performs printing processing based on the printing data. Thus, the nail art design based on the printing data is printed onto the nail T.

[0191] Furthermore, other points are the same as in the first embodiment, so their description is omitted.

[0192] As described above, according to this embodiment, in addition to obtaining the same effects as in the first embodiment, the following effects can also be obtained.

[0193] That is, in this embodiment, the printing apparatus 100, which is a printing device, includes the contour detection device 1 shown in the first embodiment. Therefore, for objects such as nails T, which are difficult to detect and have boundaries with the fingernail, it takes less time and allows for precise determination of the nail area as the printing range without tiring the user's hands, enabling aesthetically pleasing nail printing without overflow or other defects.

[0194] Furthermore, the user's nail T features are stored as error samples each time the device is reused. Therefore, the more it is used, the less effort the user needs to put into the printing device 100.

[0195] Furthermore, the embodiments of the present invention have been described above, but the present invention is not limited to these embodiments, and various modifications are possible without departing from its spirit.

[0196] For example, in the above embodiments, the contour detection unit 813 uses the ESR method to match the region of the initial shape To0 with the region of the nail T. However, the algorithm that can be used by the contour detection unit 813 when matching the region of the initial shape To0 with the region of the nail T is not limited to ESR.

[0197] For example, this embodiment can also be applied when algorithms such as AAM (Active Appearance Model), ASM (Active Shape Model), and ACM (Active Contour Model) are used when the contour detection unit 813 matches the region of the initial shape To0 with the area of ​​the nail T.

[0198] Furthermore, in the above embodiments, the contour detection device 1 is exemplified as performing contour detection processing on its own. However, for example, the contour detection device 1 may also cooperate with external devices such as various terminal devices to perform operations according to instructions input from the external devices, or perform part or all of the computational processing on the external device side.

[0199] External devices could include, for example, portable terminal devices such as smartphones or tablets, or servers, but are not limited to these. The external device need only be able to communicate with the contour detection device 1 (in the second embodiment, the printing device 100), and could also be, for example, a laptop or stationary personal computer, or a gaming terminal.

[0200] When the contour detection device 1 (in the second embodiment, the printing device 100) cooperates with an external device to perform various processes, the program for performing contour detection processing and various data are stored in the storage unit of the control device of the external device.

[0201] Furthermore, in the above embodiments, the case where the detection object is a fingernail T is illustrated, but the detection object that can be contour detected by the contour detection device is not limited to the fingernail T.

[0202] For example, the outline of the face or the outline of facial features such as the eyes, nose, and mouth can also be used as the detection object.

[0203] Furthermore, in the second embodiment, a printing apparatus 100 is exemplified as the printing apparatus that applies printing to a nail T. However, the printing apparatus is not limited to printing apparatus 100, and printing can also be applied to objects other than the nail T. When the printing object is other than the nail T, a contour detection device is used to detect the contour of the area depicting the printing object.

[0204] Furthermore, in the case where the contour detection device 1 is a device for detecting the contour of the nail T, the second embodiment illustrates an example of applying printing within the detected contour after the contour of the nail T is detected. However, the post-contour detection processing is not limited to printing. For example, automatic nail buffing or automatic nail care, which automatically trims the surface of the nail T, can also be performed as post-processing.

[0205] Furthermore, when the object of detection is a medical image captured by a medical imaging device containing various internal organs, as a result of contour detection processing, image diagnosis can be performed by doctors or others, or information about the patient's health status can be provided.

Claims

1. A contour detection device, characterized in that, have: A contour detection mechanism detects a first nail contour depicting the region of the nail from a finger image containing the nail by matching it with a nail contour model; and The receiving agency accepts input of a second nail profile corresponding to the first nail profile mentioned above from a user who does not approve of the first nail profile mentioned above. The first nail profile that has been input with the second nail profile includes multiple first nail profiles, and each of these multiple first nail profiles has been input with a second nail profile. The aforementioned contour detection mechanism derives information representing the feature point with the maximum error and differential information representing the magnitude of the difference. The feature point with the maximum error is the feature point among the feature points of the first nail contour that constitutes the second nail contour and generates the maximum difference relative to the second nail contour. The aforementioned contour detection mechanism groups the multiple first nail contours separately based on information related to the size of the first nail contour. When a certain number of first nail contours that have been grouped as described above are stored, the contour detection mechanism, for other first nail contours that are not accepted by the user and are newly detected, for any number of executions of the detection cycle, forcibly changes the coordinate values ​​of the feature points that produce the largest difference relative to the average of the second nail contours of the other first nail contours, using information representing the magnitude of the difference.

2. The contour detection device as described in claim 1, characterized in that, The aforementioned contour detection mechanism groups the data based on the aspect ratio of the first nail contour.

3. The contour detection device as described in claim 1 or 2, characterized in that, The aforementioned contour detection mechanism also detects the finger contour of the region depicting the finger from the aforementioned finger image, and further groups the finger regions based on the aforementioned finger contours with respect to the area ratio of the finger regions based on the aforementioned finger contours to the nail regions based on the aforementioned first nail contour.

4. The contour detection device as described in claim 1 or 2, characterized in that, The above difference information is derived by averaging each group.

5. A printing apparatus, characterized in that, The device includes a printing mechanism that prints the nail area depicted by the first nail contour detected by the contour detection device according to any one of claims 1 to 4.

6. A contour detection method, characterized in that, The process includes the following steps: The contour detection process involves matching the contour with a nail contour model to detect a first nail contour depicting the region of the nail from a finger image containing the nail; and The acceptance process accepts input of a second nail profile corresponding to the first nail profile mentioned above from a user who does not approve of the first nail profile. The first nail profile that has been input with the second nail profile includes multiple first nail profiles, and each of these multiple first nail profiles has been input with a second nail profile. The aforementioned contour detection process derives information representing the feature point with the maximum error and difference information representing the magnitude of the difference. The feature point with the maximum error is the feature point among the feature points of the first nail contour that constitutes the second nail contour and generates the maximum difference relative to the second nail contour. The aforementioned contour detection process groups the plurality of first nail contours into groups based on information related to the size of the first nail contour. When a certain number of first nail contours that have been grouped as described above are stored, the contour detection process, for other first nail contours that are not approved by the user and are newly detected, for any number of executions of the detection loop, forcibly changes the coordinate values ​​of the feature points that produce the largest difference relative to the average of the second nail contours of the other first nail contours using information representing the magnitude of the difference.

7. A computer-readable, non-transitory recording medium, characterized in that, A program is recorded that causes at least one processor to perform the following functions: The contour detection function detects the first nail contour, representing the region depicting the nail, from a finger image containing the nail by matching it with a nail contour model; and The acceptance function accepts input of a second nail profile corresponding to the first nail profile mentioned above from a user who does not approve of the first nail profile. The first nail profile that has been input with the second nail profile includes multiple first nail profiles, and each of these multiple first nail profiles has been input with a second nail profile. The aforementioned contour detection function derives information representing the feature point with the maximum error and difference information representing the magnitude of the difference. The feature point with the maximum error is the feature point among the feature points of the first nail contour that constitutes the second nail contour and generates the maximum difference relative to the second nail contour. The aforementioned contour detection function groups the multiple first nail contours separately based on information related to the size of the first nail contour. When a certain number of first nail contours that have been grouped as described above are stored, the contour detection function, for any other first nail contours that are not approved by the user and are newly detected, forcibly changes the coordinate values ​​of the feature points that produce the largest difference between the average of the other first nail contours and the second nail contours, using information representing the magnitude of the difference, according to any number of executions of the detection loop.

Images