JPH10240962A - Device for preparing picture and method therefor and picture preparation program storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prepare a facial picture with a high degree of recreation by collecting characteristic parts in each of an inputted plurality of facial picture data for preparing a single facial picture. SOLUTION: A storing means 3 stores facial part information constituted of facial parts and a rule for selecting, deforming, and arranging the facial parts. A picture inputting means 4 is constituted of a camera which image picks-up a face, and this can be constituted of a scanner which reads the picture of a facial photographic so that picture data can be fetched. A controlling means 2 synthetically controls the picture inputting means 4, characteristics amount extracting means 5, each kind inputting means 6, and picture outputting means 7. The controlling means 2 extracts the characteristics amounts of a plurality of facial parts from each of an inputted plurality of facial picture data, reads picture part information from the storing means 3 based on the characteristic amounts whose difference with a preliminary reference value or absolute value is the maximum among the characteristics amounts extracted from the plurality of facial picture data related with each of the plurality of facial parts, and prepares a single facial picture.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image creating apparatus, an image creating method, and an image creating program storage medium for creating a face image by combining images of a plurality of face parts stored in advance.

[0002]

2. Description of the Related Art For example, a single composite image may be created by combining images of a plurality of components, such as a montage image of a person's face. In creating such a composite image, generally, a plurality of images are stored in advance for each component, and a single image is extracted for each component from the storage means arbitrarily or based on a predetermined rule, and the extraction is performed. The synthesized image of each component is synthesized.

[0003] For this reason, as a conventional image creating apparatus for creating the above-mentioned montage image, eyes, nose, mouth, eyebrows,
It is provided with storage means for storing a plurality of images for each of face parts such as ears, face contours, and hair, and stores face parts such as eyes, nose, mouth, eyebrows, ears, face contours, and hair from input face image data. It has been proposed to extract a feature amount for each of them, select an image corresponding to the extracted feature amount from a storage unit for each face part, and combine the images of the selected face parts to create a face image.

[0004]

However, in the conventional image forming apparatus, only a portrait of the person is automatically created based on the face image data of a single person. There was a problem that could not be done.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image forming apparatus, an image forming method and an image forming method capable of forming a highly entertaining face image by forming a single face image based on face image data of a plurality of persons. An object of the present invention is to provide a creation program storage medium.

[0006]

According to the first, second and third aspects of the present invention, a plurality of face image part information having different feature amounts for each of a plurality of face parts is stored in a storage means, and a plurality of face image parts are stored in the storage means. Accepting input of face image data, extracting feature amounts of a plurality of face parts from each of the plurality of input face image data, and extracting a feature amount of each of the plurality of face parts from among the feature amounts extracted from the plurality of face image data in advance. It is characterized in that a single face image is created by reading out image part information from the storage means based on the feature amount having the maximum difference or ratio with a predetermined reference value.

Therefore, in the first, second and third aspects of the present invention, a single face image is created by collecting characteristic portions in each of a plurality of input face image data.

[0008]

FIG. 1 is a block diagram showing a configuration of an image forming apparatus according to an embodiment of the present invention. The image creation device 1 creates a single portrait with a specified style and expression based on the feature amount of the face part extracted from the input plurality of face image data. Face parts are eyes, nose, mouth, eyebrows,
Ears, contours and hair. In addition, the style of painting refers to the manner of exaggeration and deformation of each unified face part, which differs from painter to painter, and that even in the same painter, a cartoon character style, a handsome or beautiful woman style, or an accurate portrait style It depends on the style. An expression is a state of a face that expresses emotions and sorrows.

The image creating apparatus 1 includes a control means 2 for selecting, deforming and arranging each face part as an image part to create a portrait, a storage means 3 for storing a plurality of pieces of face part information, a camera for taking in image data, and the like. Image input means 4, feature amount extracting means 5 for extracting feature amounts of face parts included in image data,
It is composed of a type input unit 6 for receiving a selection input of an image type such as a style or a facial expression, and an image output unit 7 such as a printer for outputting a created image.

The control means 2 controls the image input means 4, the feature quantity extracting means 5, the type input means 6, and the image output means 7 in a centralized manner. The control unit 2 includes a feature amount comparison unit 13 that compares the feature amount input from the feature amount extraction unit 5 with a predetermined reference value, and a selection rule, a deformation rule, and an arrangement rule corresponding to the input image type. A component information selecting unit 8 that reads out face parts from the storage unit 3 in accordance with the read selection rules while reading from the storage unit 3, a component deformation unit 9 that deforms the face components in accordance with the deformation rule, and arranges the face parts after the deformation. It includes a part arranging means 10 for arranging according to rules.

The storage means 3 selects a face part and a face part,
Face part information including rules for deformation and arrangement is stored. For this reason, the storage unit 3 stores the face part library 11 in which a plurality of images are classified and stored for each image type for each face part, and the rule group used when selecting, deforming, and arranging the face part. And a rule library 12 stored by classifying each image type.

As shown in FIG. 2, the face part library 11 includes image libraries 11a to 11g of face parts such as eyes, nose, mouth, eyebrows, ears, contours, and hair. One image library stores a plurality of image groups of face parts having different feature amounts for each of image types i = 1 to n. For example, the eye image library 11a stores, for each image type, an image group constituted by a plurality of eye images having different inclinations and roundnesses as eye characteristic amounts.

As shown in FIG. 3, the rule library 12 includes a selection rule library 12a, a transformation rule library 12b, and an arrangement rule library 12c. The libraries 12a to 12c store a rule group including a plurality of rules for each face part for each of the image types i = 1 to n. For example, the selection rule library 12a
Is composed of a selection rule group for each image type. One selection rule group includes, for example, a selection rule for each feature amount that specifies each image in each of the image libraries 11a to 11g of the face part library 11, such as a rule for the inclination of the eyes and a rule for the roundness. (See FIG. 5B).

Each rule can be a fuzzy inference membership function and a fuzzy rule. For example, as for the selection rule, the range in which the measurement result of the feature amount is included in the membership function is set as the antecedent label of the fuzzy rule, and each image is output as the consequent part of the fuzzy rule (see FIG. 5C). . Also, each rule can be configured by a table.

The image input means 4 is constituted by a camera for picking up a face, but may be constituted by a scanner for reading an image of a face photograph. The feature amount extraction means 5
The image processing is performed on the face image data input from the image input means 4, the position of the face part is specified in the face image data after the image processing, and the feature amount of the face part is extracted. For example, with respect to the eyes, measured values of inclination, roundness (aspect ratio) and size are extracted as feature amounts. The type input means 6 is constituted by a keyboard, a mouse, or the like, and receives a selection input of an image type such as style, facial expression, age, and gender. The image output unit 7 is a printer that outputs a synthesized image on a sheet.

FIG. 4 is a flow chart showing an image creation processing procedure in the image creation apparatus. Control means 2
Receives the selection input of the image type by the type input means 6 (s1), and when the image type is selected and input, the image input means 4 captures a plurality of face images (s2). Here, for the sake of explanation, two face images are taken as a first face image and a second face image. Next, the control means 2
Performs predetermined image processing on each of the first face image data and the second face image data captured by the image input unit 4 (s
3) The feature amount is extracted by specifying a face part in the first face image data and the second face image data after the face image processing by the feature amount extracting means 5 (s4).

Thereafter, the control means 2 controls the feature quantity comparing means 1
In 3, the feature value extracted from the first face image data and the feature value extracted from the second face image data for each face part are compared with a predetermined reference value, and the difference between the feature value and the reference value is determined. The feature value having the largest absolute value is determined as a feature value used for creating a face image (s5). This reference value is, for example, an average value of the Japanese for the feature amount of each face part. The ratio between the feature value extracted from the first face image data and the feature value extracted from the second face image data and the reference value is calculated.
The feature value having the largest absolute value of this ratio may be determined as the feature value used for creating the face image. Further, based on the input image type, the selection rule library 12
A selection rule relating to the selection of a face part image is read from a (s6), and a face part image is selected from any of the image part libraries 11 based on the read selection rule and the determined feature amount (s7). .

It should be noted that the second image type is determined only as an image type different from the first image type selected and input through the type input means 6, and is, for example, predetermined. The image type whose priority order is next to the first image type can be selected.

Further, a deformation rule relating to the deformation of the face part image is read from the deformation rule library 12b based on the input image type (s8), and the face part image is deformed based on the read deformation rule and the determined feature amount. (S
9) Read out an arrangement rule related to the arrangement position of the face part image from the arrangement rule library 12c based on the input image type (s10), and arrange the position of the face part image based on the read arrangement rule and the determined feature amount. Is determined (s11).

The control means 2 executes the processing of s4 to s11 for each of the face parts (s12). When the selection, deformation, and determination of the arrangement position of the face part images are completed for all the face parts, the selected and deformed face part images are arranged at the determined arrangement positions and the face images are synthesized (s1).
3) The synthesized face image is output by the image output means 7 (s14).

The control means 2 can be constituted by a microcomputer, and the storage means 3 can be constituted by a hard disk, for example. Control means 2
The processing contents described above and the storage contents of the storage means 3 can be externally supplied via a storage medium.

FIG. 5 is a diagram specifically showing processing relating to the determination of the feature quantity, the selection of the face part image, the deformation of the face part image, and the determination of the arrangement position in the image creating apparatus. For example, when the feature amount is determined for the eye image as the face part by the process of s5 shown in FIG.
Specifies an eye in each of the first face image data and the second face image data on which predetermined image processing has been performed, an angle θ between the center line of the eye and a horizontal line, an eye width x, and an eye height y The distance d from the center line of the eye to the chin is measured (FIG. 5A).

The feature amount comparing means 13 stores in advance the average values θ0 and y0 / x0 of the Japanese eye inclination θ and the roundness of the eyes y / x, and extracts the eye inclination extracted from the first face image data. θ1 and the inclination θ2 of the eyes extracted from the second face image data are compared with the average value θ0, and the value of θ1 and θ2 having the larger absolute value of the difference from the average value θ0 is determined as the feature amount θ. . Similarly, the feature value comparison unit 13 calculates the average value y0 of the roundness of the eyes y1 / x1 extracted from the first face image data and the roundness of the eyes y2 / x2 extracted from the second face image data.
/ X0, and the value of y1 / x1 and y2 / x2, which has the larger absolute value of the difference from the average value y0 / x0, is determined as the feature value y / x.

The feature quantity comparing means also calculates the average value d0 of the Japanese for the distance d between the eyes and the distance h from the center line of the eyes to the chin.
And h0 are stored in advance, and an eye interval d1 extracted from the first face image data and an eye interval d2 extracted from the second face image data are compared with an average value d0, and d1 and d2 are compared.
The value of the one with the larger absolute value of the difference from the average value d0 is determined as the feature amount d. Similarly, the feature amount comparison means 13
A distance h1 from the chin extracted from the first face image data, and
The distance d2 from the chin extracted from the second face image data is compared with the average value d0, and the value of d1 and d2 having the larger absolute value of the difference from the average value d0 is determined as the feature amount h.

The characteristic amount comparing means 13 determines the characteristic amount θ
And y / x to the component information selecting means 8. The part information selecting means 8 reads out a selection rule group to be used based on the type i previously selected and inputted, and adds the angle θ and the angle θ to the eye inclination rule and the eye roundness rule included in the selection rule group. Applying each value of x / y,
A label in a range that includes the measured values of the inclination and roundness of the eyes is determined (FIG. 5B). Further, using the labels determined for the inclination and roundness of the eyes, the corresponding eye images are selected from the group of eye images of the image type selected and input in the eye parts library 11a (FIG. 5).
(C)).

The feature quantity comparing means 13 outputs the determined feature quantities y and x to the component deformation means 9 and outputs the determined feature quantities d and h to the component arranging means 10. The fuzzy article deformation means 9 reads a group of deformation rules to be used based on the type i selected and input in advance, and applies the value of the feature quantity x or y to the eye size rule included in the group of deformation rules. , Determine the magnification or reduction of the eye image. Further, the component placement means 10 reads a placement rule group to be used based on the type i selected and input in advance,
The feature amount d is added to the eye arrangement rule included in this arrangement rule group.
And h, the position of the eyes within the contour of the face is determined.

The same processing as described above is executed for other face parts, and all face parts are selected, deformed, and their arrangement positions are determined. However, the deformation processing is executed only for the face outline, eyes and mouth, and the deformation processing for the other face parts is not executed. In addition, about eyes and mouth,
By storing images having different sizes in the face part library 11, the images can be included in the selection processing.

FIG. 6 is a diagram showing an example in which a single portrait is created by combining two pieces of face image data by the above processing. When the processing of s5 is not performed in the processing procedure illustrated in FIG. 4, the portrait illustrated in FIG. 6B is created only by the feature amounts extracted from the first face image data illustrated in FIG. A portrait shown in FIG. 4D is created only from the features extracted from the second face image data shown in FIG. 4C. On the other hand, by executing the processing procedure shown in FIG. 4, the feature amount extracted from the first face image data and the feature amount extracted from the second face image data are compared, and the first face A more individual face component is selected from the face component of the image data and the face component of the second face image data, and a composite face image shown in FIG. 6E is created.

Therefore, a unique face image that does not actually exist is created using the two face image data, and a face image with high entertainment value can be created.

Note that each of the selection rules, transformation rules, and arrangement rules stored in the rule library 12 is a rule using the average value of the Japanese as an intermediate value, and by selecting an output other than the intermediate value in each rule, A single face image can be created from a plurality of pieces of face image data based on characteristic information without using the feature amount comparison unit 13.

The number of face image data to be input is not limited to two, and the face part designated and input in the created composite face image is replaced with the feature amount which is not determined by the feature amount comparing means 13. Based on the selected, deformed and arranged face parts, the face parts may be output.

[0032]

According to the first, second and third aspects of the present invention, a feature amount of a plurality of face parts is extracted from each of a plurality of pieces of input face image data, and a plurality of feature parts are extracted for each of the plurality of face parts. The image component information is read out from the storage unit based on the feature amount having the largest difference or ratio from a predetermined reference value among the feature amounts extracted from the face image data, and a single face image is created. As a result, a unique face image can be created by collecting characteristic portions in each of the plurality of pieces of input face image data, and a highly entertaining face image can be created.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an image creating apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing storage contents of a face part library constituting a storage unit of the image creating apparatus.

FIG. 3 is a diagram showing a storage state of a contour image and a hair image in the face part library.

FIG. 4 is a flowchart illustrating an image creation processing procedure in the image creation apparatus.

FIG. 5 is a diagram specifically showing processing relating to determination of a feature amount, selection of a face part image, deformation of the face part image, and determination of an arrangement position in the image creating apparatus.

FIG. 6 is a diagram showing an example in which a single portrait is created by combining two face image data by the above processing.

[Explanation of symbols]

 Reference Signs List 1-Image creation device 2-Control unit 3-Storage unit 4-Image input unit 5-Feature amount extraction unit 6-Type input unit 7-Image output unit 8-Part information selection unit 9-Part transformation unit 10-Part arrangement unit 11-Parts library 12-Rule library 13-Feature comparison means

Claims (3)

[Claims] 1. A storage means for storing a plurality of pieces of face image part information having different feature amounts for each of a plurality of face parts; an image data input means for receiving an input of a plurality of face image data; The feature amount of a plurality of face parts is extracted from each of the image data, and the absolute value of the difference or ratio between the feature amount extracted from the plurality of face image data and the predetermined reference value for each of the plurality of face parts is determined. A control unit for reading out the image part information from the storage unit based on the maximum feature amount to create a single face image. 2. A storage means for storing a plurality of pieces of face image part information having different feature amounts for each of a plurality of face parts, receiving an input of a plurality of face image data, and A feature amount of a plurality of face parts is extracted from each of the plurality of face parts, and for each of the plurality of face parts, a feature having a maximum difference or absolute value of a ratio from a predetermined reference value among feature amounts extracted from the plurality of face image data. An image creating method comprising: reading out image part information from a storage unit based on an amount to create a single face image. 3. A process for receiving an input of a plurality of face image data, a process for extracting feature amounts of a plurality of face parts from each of the plurality of input face image data, and a plurality of face parts for each of the plurality of face parts. A process of creating a single face image using image component information selected based on a feature amount having a maximum difference or ratio from a predetermined reference value among feature amounts extracted from the face image data; and An image creation program storage medium characterized by storing a program comprising: