JP2021037660A - Stamp image processing device and stamp processing machine - Google Patents

Abstract

To provide a stamp image processing device and a stamp processing machine which can create a stamp of a clear face photograph image without using a color chart.SOLUTION: A stamp image processing device comprises: face image detection means 11 which detects a face image from an input face photograph; background image detection means 12 which detects a background image; gradation correction means 13 which converts these images into a gray scale; correction image storage means 21 which stores the face image and background image converted into the gray scale and the whole image; gradation conversion means 14 which decides a conversion reference value from the gradation of the stored background image and converts the gradation of the whole image on the basis of the conversion reference value; converted image storage means 23 which stores the whole image with the converted gradation; binarization arithmetic means 16 which binarizes the stored whole image; and binarized image storage means 27 which stores the binarized whole image. The binarized image is sent to a stamp processing machine 30 to create a porous stamp 40.SELECTED DRAWING: Figure 1

Description

The present invention relates to a stamp image processing device and a stamp processing machine used to create a stamp from a portrait.

A device that automatically creates a stamp based on a face photograph taken by a user has been developed and has already been put into practical use. Using this device, the user can easily create an original stamp by inputting a photo of himself or a friend's face.

This stamp can be produced by heat-processing a porous material to be an ink occlusion body, for example, using a stamp processing machine equipped with a thermal head as shown in Patent Document 1. The thermal head is provided with a large number of heat generating elements arranged in a line, and the thermal head is moved to come into contact with the surface of the porous material, and each heat generating element is selectively heat-driven to melt and solidify the porous material. Create stamps one line at a time. Since the ink cannot pass through the melt-solidified portion of the porous material, the portion becomes white when stamped, and the other portion is colored in the color of the ink because the ink exudes. In the present specification, the ink is described as black for the sake of brevity, but it goes without saying that the color of the ink is arbitrary.

In this way, the stamp surface of the stamp is configured as a set figure of fine points that are binarized, but if the face photograph is simply binarized to form the stamp, the stamp surface becomes rough and unsightly. Therefore, Patent Document 2 proposes an image processing method in which the original image is converted into a grayscale with 256 gradations, the conversion value is corrected, and then a well-known dot-dispersed dither matrix is used to form a stamp surface expressing halftone. Has been done.

However, the image processing method of Patent Document 2 is not specified in the stamp of a facial photograph, but targets a general original image. Therefore, when the image processing method of Patent Document 2 is applied to the stamp creation using the thermal head using the face photograph as the original image, there are the following two problems.

First, when the background of the facial photograph has a high brightness and is close to white, the thermal head melts and solidifies a wide portion around the face. For this reason, a large amount of heat is applied to the porous body, and the residual heat causes the opening of the porous body near the face to be slightly closed, which may cause distortion in the face image.

Secondly, it is required to make the facial expression clear especially in the facial photograph, but if the background brightness is too high or too low, the face part becomes white or dark, and a beautiful stamp surface is obtained. There were times when I couldn't.

In addition, although Patent Document 3 does not use image processing for stamping, a color tag composed of a reference color and a single color background are photographed at the same time as the subject, and the image data of the entire image is taken with reference to the image data of the color tag portion. An image processing method for correcting and adjusting the color tone is described. However, it cannot be applied to a face photograph easily taken by a general user because it requires a color tag of a standard color and a single color background.

Japanese Patent No. 6399308 Japanese Patent No. 4122669 Japanese Unexamined Patent Publication No. 7-274194

Therefore, an object of the present invention is to solve the above-mentioned conventional problems and to provide a stamp image processing device and a stamp processing machine capable of producing a stamp of a clear facial photographic image without using a color tag. is there.

The image processing apparatus for stamping of the present invention, which has been made to solve the above problems, includes a face image detecting means for detecting a face image from an input face photograph, a background image detecting means for detecting a background image, and these. A conversion reference value is determined from the gradation correction means for grayscale the image, the correction image storage means for storing the grayscaled face image, the background image and the entire image, and the gradation of the stored background image, and this conversion is performed. A gradation conversion means for converting the gradation of the entire image based on a reference value, a conversion image storage means for storing the gradation-converted entire image, and a binarization for binarizing the entire image stored in the conversion image storage means. It is characterized by including a calculation means and a binarized image storage means for storing a binarized whole image.

In addition, a gradation linearizing means for linearizing the gradation of the entire image stored in the converted image storage means using a correction table based on the imprint density, and a linearized image storage means for storing the linearized image. It is preferable to further prepare.

Further, the stamp processing machine of the present invention, which has been made to solve the above problems, moves a thermal head equipped with a large number of heat generating elements arranged in a line in a state of being in contact with the surface of a porous material. A stamp processing machine for selectively heat-driving each heat generating element to melt and solidify a porous material to produce a stamp surface line by line, which is obtained by the stamp image processing apparatus according to claim 1 or 2. It is characterized in that it is provided with a control means for controlling the amount of heat generated by each heat generating element based on the binarized overall image.

According to the present invention, the conversion reference value is determined from the gradation of the background image, and the gradation of the entire image is converted based on this conversion reference value. Therefore, regardless of the brightness of the background image and without using a color tag, It is possible to create a stamp of a clear facial photograph image.

It is a block diagram which shows the whole structure of this invention. It is a block diagram which shows the flow of this invention. It is explanatory drawing of the face photograph. It is a graph which shows the relationship between the tone and the density as seen by the human eye. It is a graph explaining the linearization process of a tone. It is a principle diagram of the Tiza matrix. This is the original image of the face photo. It is an image which has been processed up to ST7. It is an image which has been processed up to ST9.

An embodiment of the present invention will be described below.
As shown in FIG. 1, the stamp image processing apparatus of the present invention includes a calculation unit 10 and a storage unit 20. The calculation unit 10 includes various calculation means, and the storage unit 10 includes various storage means and a database. The functions of each means will be described below with reference to the flow of FIG.

First, the facial photograph 50 as shown in FIG. 3 is input to the stamp image processing device by a scanner or other appropriate reading means (ST1). The face photograph 50 is generally a color photograph, and includes a face image 51 and a background image 52. The brightness of the face image 51 and the background image 52 varies, and the background image 52 may be brighter than the face image 51 as shown in FIG. 3 (A), or the opposite may occur as shown in FIG. 3 (B). The skin color of the face image 51 also varies. The face image detecting means 11 of the calculation unit 10 detects the face image 51 from the input face photograph 50, and the background image detecting means 12 detects the background image (ST2).

In order to separate the face image 51 and the background image 52 (other than the face image) from the face photograph 50, first, the face image 51 is extracted. For the extraction of the face image 51, a method of extracting only the color close to human skin can be considered according to the RGB rules of the color photograph. Specifically, when each color distributed in a color image is represented by the brightness (0 to 255) of red (R), green (G), and blue (B), R (red) is the maximum value. Areas where red (R) is 100 or more, areas where the difference between red (R) and green (G) is 20 to 100, and areas where the difference between green (G) and blue (B) is less than 100 Certified and extracted as face image 51. By certifying and extracting the area other than the area recognized as the face image as the background image 52 (other than the face image), the face image 51 and the background image 52 (below the face image) can be separated from the face photograph 50. The method of extracting the face image 51 is not limited to the above method, and many methods such as a method of identifying and extracting only the contour portion of the face from the face photograph can be considered.

These images are grayscaled by the tone correction means 13 of the calculation unit 10 (ST3). Grayscale conversion means that the pixels of the face image 51 and the background image 52 are converted into 256 steps from 0 (pure black) to 255 (pure white) in gradation according to the brightness. The grayscaled face image, background image, and overall image are stored in the corrected image storage means 21 (ST4).

Next, the conversion reference value is determined from the gradation average value of the background image stored in the corrected image storage means 21 based on the brightness correction table shown in Table 1 (ST5). The gradation average value of the background image is the sum of the individual pixel values having a gradation range of 0 to 255 distributed in the area recognized as the background image, and the total value is divided by the number of distributed pixels. Value. The target gradation average value of the face pixel in Table 1 is the conversion reference value. In this embodiment, when the brightness of the background image 52 is high (when the gradation average value exceeds 160), the gradation 160 is set as the conversion reference value, and conversely, when the brightness of the background image 52 is low (gradation average value). If is less than 140), the gradation 180 is set as the conversion reference value. In the correction table this time, the reference gradation value is used as the average value, but it goes without saying that this may be used as another reference value such as the median value.

Next, the gradation conversion means 14 converts the gradation of the entire image based on the conversion reference value. As shown in FIG. 2, first, the gradation of the grayscaled face image is converted based on the conversion reference value (ST6), and then the gradation of the entire image is converted based on the conversion of the face image (ST6). ST7). It is preferable that this conversion is performed by a method of changing the conversion rate according to the difference between the gradation value of each pixel and the conversion reference value, instead of uniformly translating the gradation of each pixel. In this embodiment, the conversion reference value for converting the gradation of the face image is determined based on the gradation average value of the background pixels. When the gradation average value of the background pixel area is 200, the conversion reference value is 160.

Subsequently, the gradation conversion rate is calculated. The gradation conversion rate is the ratio of the conversion reference value and the gradation average value in a specific area before gradation conversion, and is expressed by the following equation (1). When the gradation average value of the specific area before gradation conversion is larger than the conversion reference value, the gradation conversion rate shows a negative value.
[(Conversion reference value-Before conversion / Gradation average value of specific area) ÷ 256] × 100 ... Equation (1)

For example, when the gradation average value of the background image area is 200 as described above and the gradation average value of the face pixels before gradation conversion is 160, the gradation conversion rate is 0%. When the gradation average value of the face pixel before gradation conversion is 150, the gradation conversion rate is 3.9%, and when the gradation average value of the face pixel before gradation conversion is 140, the gradation When the conversion rate is 7.8% and the gradation average value of the face pixels before gradation conversion is 130, the gradation conversion rate is 11.7%.

Subsequently, the calculated gradation conversion rate is assigned to each pixel of the face portion to convert the gradation of the face portion image. For example, when the average value of the face pixels before gradation conversion is 130 and the conversion reference value is 160, the gradation conversion rate is 11.7%, and the gradation value of each pixel of the face is changed. The gradation value before the gradation conversion is adjusted so as to increase by 11.7%, respectively, and the gradation average value of the face portion after the gradation conversion becomes 145.

Similarly, the gradation of the background portion (other than the face portion) is converted based on the obtained gradation conversion rate. For example, when the gradation average value of the background portion (other than the face portion) before gradation conversion is 200 and the gradation conversion rate is 11.7%, the gradation value of each pixel of the face portion is set. The gradation value before the gradation conversion is adjusted so as to increase by 11.7%, respectively, and the gradation average value of the background portion (other than the face portion) after the gradation conversion becomes 223. By such a series of gradation conversion, the gradation of the entire image is converted.
The specific numerical value of the conversion reference value is not limited to this example, and can be appropriately set depending on the block copy image to be created. Further, the gradation conversion rate may not be provided, and the conversion reference value may be used as the target gradation average value of the face pixels by gradation conversion.

Further, two conversion reference values may be provided for the gradation average value of the background in one range, and the degree of gradation conversion of the face portion and the gradation conversion of the background portion may be different.

In the steps from ST5 to ST7, when the brightness of the background image is high, the gradation average value of the face image is set low to convert the gradation of the face image to a low value, and the gradation of the entire image is also converted to a low value. Means. As a result, the gradation average value of the background image is also lowered, and the brightness of the background is lowered, so that distortion of the face image due to the residual heat of the thermal head can be avoided. On the contrary, when the brightness of the background image is low, the average value of the gradation of the face image is converted to a higher value, and the gradation of the background image is also raised at the same time, so that the overall brightness is increased. In this way, clear facial photo data is created regardless of the brightness of the background. The obtained image is stored in the converted image storage means 23.

Next, this image is linearized by the gradation linearizing means 15 using the stamping correction table 24 (ST8). FIG. 4 is a graph in which the horizontal axis is the gradation and the vertical axis is the density. Theoretically, the gradation and the density are proportional to each other, so the plot should be a straight line. However, when viewed by the human eye, the gradation and the density are not directly proportional to each other, and appear almost white in the region where the gradation is close to 255, and appear black in the region where the density is close to 0. The arrows in FIG. 4 indicate this change. Therefore, if the gradation of the image is used as it is for both ends of the gradation, the contrast between black and white becomes too strong. Therefore, as shown by an arrow in FIG. 5, the values at both ends of the gradation are shifted using the marking correction table 24. With this correction, it is possible to linearize the gradation as seen by the human eye. The linearized image is stored in the linearized image storage means 25.

Next, the linearized image is converted into a binarized image in which the binarization calculation means 16 expresses a halftone using the binarization table 26 (ST9). This is a well-known technique called dot-dispersed dither matrix, which uses the principle of halftone dot printing that looks gray to the human eye when minute dots are dispersed. The principle is shown in FIG. In this example, integers from 0 to 15 are randomly arranged in a 4 × 4 matrix, and this is multiplied by 16 to create a dither matrix. The image is also divided into minute matrices consisting of 4x4 pixels, and the gradation of each cell in the matrix is compared with the numerical value of the corresponding position in the dither matrix. Black is used, and if the gradation of the image is large, the square is white. By repeating this calculation, the image is converted into a halftone image in which minute dots are dispersed. The binarized whole image thus obtained is stored in the binarized image storage means 27. The method of converting the image into a binarized image expressing a halftone is not limited to the method using the dot-dispersed dither matrix, and other methods such as a well-known error diffusion method may be used.

As shown in FIG. 1, a stamp processing machine 30 is connected to the stamp image processing apparatus of the present invention. The stamp processing machine 30 includes a thermal head 32 provided with a large number of heat generating elements 31 arranged in a line, an elevating control means 33 thereof, a heat generation drive control means 34, an overall control means 35, a touch panel 36, a numeric keypad 37, and the like. It has.

The binarized whole image obtained by the stamp image processing apparatus of the present invention is input from the binarized image storage means 27 to the stamp processing machine 30, and the elevating control means 33 and the heat generation drive control means 34 are the thermal head 32. Is moved in contact with the surface of the porous material, and each heat generating element 31 is selectively heat-driven to melt and solidify the porous material, and a stamp surface is created line by line. This makes it possible to create a porous stamp 40 of a clear facial photographic image. If the porous stamp 40 is impregnated with ink, it can be used as a stamp for a clear facial photograph.

Instead of the embodiment, FIG. 7 shows the original image of the facial photograph, FIG. 8 shows the image processed up to ST7, and FIG. 9 shows the image processed up to ST9. In the original image, the average gradation value of the face image is 140, and the average gradation value of the background image is 60, and the face portion is darkened. In FIG. 8, the conversion reference value is set to 180, the gradation average value of the face image is converted to 180, and the gradation average value of the background image is also converted to 100, so that the face is bright, but the contrast is strong. However, in FIG. 9, the face image is clear and the contrast is in a moderately good state. It has been confirmed that even if each image is input to the stamp processing machine 30 to produce the porous stamp 40, the same tendency as described above is obtained.

As described above, according to the present invention, it is possible to create a clear stamp of a facial photographic image without using a color tag.

10 Calculation unit 11 Face image detection means 12 Background image detection means 13 Gradation correction means 14 Gradation conversion means 15 Gradation alignment means 16 Binarization calculation means 20 Storage unit 21 Correction image storage means 22 Brightness correction table 23 Conversion image storage means 24 Stamping correction table 27 Binary image storage means 30 Stamping machine 31 Heat generation element 32 Thermal head 33 Elevation control means 34 Heat generation drive control means 35 Control means 36 Touch panel 37 Tenkey 40 Porous stamp 50 Face photo 51 Face image 52 Background image

Claims (3)

A face image detecting means for detecting a face image from an input face photograph, a background image detecting means for detecting a background image, a gradation correction means for grayscale these images, a grayscaled face image, and a background. A correction image storage means for storing an image and an entire image, a gradation conversion means for determining a conversion reference value from the gradation of the stored background image, and a gradation conversion means for converting the gradation of the entire image based on the conversion reference value, and a gradation conversion are performed. A converted image storage means for storing the entire image, a binarization calculation means for binarizing the entire image stored in the converted image storage means, and a binarized image storage means for storing the binarized overall image. An image processing device for stamps, which is characterized by being equipped with. It is provided with a gradation linearizing means for linearizing the gradation of the entire image stored in the converted image storage means using a correction table based on the imprint density, and a linearized image storage means for storing the linearized image. The image processing apparatus for stamping according to claim 1. While moving a thermal head equipped with a large number of heat generating elements arranged in a line in contact with the surface of the porous material, each heat generating element is selectively heat-driven to melt and solidify the porous material. A stamp processing machine that creates a stamp surface one line at a time, and the amount of heat generated by each heat generating element is based on the binarized overall image obtained by the stamp image processing apparatus according to claim 1 or 2. A stamp processing machine characterized in that it is equipped with a control means for controlling.