JP2002074341A - Device and method for extracting impressive character picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and method for clarifying the featured values of the face character, etc., of a human being, and extracting a proper character picture from information on it. SOLUTION: The subjective view of a program preparing person and emotion of a questionnaire preparing person can be excluded in extracting the featured values by a learning function and the featured values of an information originating person is clarified from all the pieces of past information. As the result of this, a character picture customized faithfully reflecting the character, the face, etc., of the information originating person can be extracted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image selection, and more particularly to a function of extracting a character image that impresses a person.

[0002]

2. Description of the Related Art With the recent development of mobile phone technology and the development of an information distribution system through a network or the like, not only voice but also image information of a partner's face are digitized and distributed through a network or the like. Among them, as a communication accent, many uses are performed, such as transferring a character image common to the public to the other party.

[0003]

Heretofore, character images common to the public have been used in the usual concept. However, considering the interactivity of the Internet, if a character image that faithfully prospers the person's character and face can be created, it may be used in various scenes.

The present invention has been made to solve the problems of such a conventional image identification information management system.

[0005]

According to a first aspect of the present invention, there is provided an apparatus for extracting an image that impresses a person from a large number of character images. Means for holding a measured amount of an impressive item, means for inputting an image of the face or body of the target person, and means for extracting a measured amount of each part of the face or body from the input image Means for determining an impressive part from each of the measured quantities, means for comparing only the impressive part with the attribute information of the character image, and extracting a character image based on the result of the comparison. Means.

According to a second aspect of the present invention, there is provided an apparatus for extracting an image which impresses a person from a large number of character images, and means for preliminarily holding a measured amount of an impressive item as attribute information of the character image. Means for inputting information relating to the personality and emotion of the target person, means for extracting a measured amount of each profile from the input content, and means for determining an impressive profile from each of the measured amounts And means for comparing the attribute information of the character image with only the impressive profile as a comparison target, and means for extracting the character image based on the result of the comparison.

Further, the invention according to claim 3 is an apparatus for extracting an image that impresses a person from a large number of character images, the means for preliminarily holding a measured amount of an impressive item as attribute information of the character image. Means for inputting an image of the face or body of the target person, means for extracting a measured amount of each part of the face or body from the input image, and inputting information related to the character and emotion of the target person Means for extracting the measured amount of each profile from the input content, means for determining an impressive part or an item such as a profile from the measured amount, and comparing only the impressive items with And means for matching the attribute information of the character image, and means for extracting the character image based on the result of the matching.

According to a fourth aspect of the present invention, in the image extracting apparatus of the first to third aspects, the method for determining an impressive item includes the step of determining each of the measured quantities by a deviation based on a general statistical distribution. It has means for converting to a value to make a determination.

According to a fifth aspect of the present invention, in the image extracting apparatus of the fourth aspect, each time a different person is used, there is provided a means for learning the measured amount.

According to a sixth aspect of the present invention, there is provided a method for extracting an image that impresses a person from a large number of character images, the method comprising: storing a measured amount of an impressive item in advance as attribute information of the character image. Having a step of inputting an image of the face or body of the target person, a step of extracting a measured amount of each part of the face or body from the input image, and inputting information related to the character and emotion of the target person Performing a step of extracting a measured amount of each profile from the input contents; a step of determining an item such as an impressive part or a profile from each measured amount; A step of collating the attribute information of the character image, and a step of extracting the character image based on the result of the collation.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS <Description of Image Processing Apparatus in FIG. 1> FIG.
FIG. 1 is a diagram showing a configuration of an image processing apparatus 1 according to one embodiment of the present invention.

As shown in FIG. 1, the image processing apparatus 1 has a configuration utilizing a normal computer system. The image processing apparatus 1 includes a CPU 11 that executes various arithmetic instructions,
A memory 12 for storing programs and the like, a dedicated image memory 13 for storing an image of a target person, a fixed disk 14 serving as a nonvolatile storage area, a display unit 15 such as a display,
An operation instructing unit 16 such as a mouse, a keyboard, a digitizer or the like for giving an operation instruction, a reading unit 17 for reading a program or various information from the storage medium 9, and a network interface 19 for enabling data communication with a wide area computer network such as the Internet. Are connected to the bus line 8 via an interface as appropriate.

The Internet 191 indicates the entire Internet, and the base station 192 is a wireless relay base station for a portable telephone. The mobile phone 193 includes an image input unit 194 which is a built-in CCD camera. A liquid crystal display 196 displays a character image and the like. Various operations are performed on the operation unit 195, including transfer of a captured image.

The program 21 is stored in the fixed disk 14 from the storage medium 9 via the reading unit 17, and is read into the memory 12 when the program 21 is executed. The storage medium 9 includes a magnetic disk, an optical disk,
This is a so-called removable storage medium such as a magneto-optical disk, and is mainly used for transferring the program 21 from another computer to the image processing apparatus.

The fixed disk 14 is used as a storage medium for reading character image data and the program from the storage medium 9 into the image processing apparatus 1.

The display unit 15 displays a character image,
Used to display and confirm the input attribute information.

The operation instructing section 16 is used for inputting attribute information of a character image and the like.
It is composed of an operation input device such as a keyboard, a digitizer or a stylus pen.

<Explanation of Functional Block Diagram of FIG. 2> FIG. 2 is a block diagram showing the configuration of the image processing apparatus 1 in terms of functions. That is, when the CPU 11 operates according to the program 121, it plays a role of each component shown in FIG.

The correspondence between the claims and the functional blocks will be described first in order to make the description common.

The functional blocks corresponding to claim 1 are an image input unit 21, an image measurement unit 22, and a maximum feature item extraction unit 2
5, an image attribute information matching unit 26, a character image database 27, and a character image selection / output unit 28.

On the other hand, a functional block corresponding to claim 2 includes a questionnaire input unit 23, a profile measurement unit 24,
Maximum feature item extraction unit 25, image attribute information matching unit 26, character image database 27, character image selection /
An output unit 28 is provided.

Claims 3 to 6 include all the functional blocks of FIG.

First, in the image input section 21, an image of a face or a body of a target person who is mainly an operator, which is captured by a CCD camera built in a mobile phone, is acquired. This image is subjected to various correction processes and then transferred from the mobile phone to the apparatus via the Internet according to an instruction from the operator.

The image measuring section 22 specifies the position of the face image by a correlation method using an average face image pattern. Furthermore, based on the relative distance from the face contour position, it is possible to extract parts such as hair, eyebrows, eyes, nose, mouth, and chin. The measurement amounts such as the area, shape, and inclination of the partial images are measured.

On the other hand, the questionnaire input section 23 presents various questions to the registrant and obtains answers.

The profile measuring section 24 gives a measured value such as a frequency to an item of a profile related to various personalities or a profile related to various emotions based on the answer of the person.

Next, the maximum characteristic item extracting section 25 extracts the most characteristic item from various measured values, that is, the item of the impressive part or profile.

According to a fourth aspect of the present invention, there is provided a method according to the fourth aspect, wherein a general statistical distribution prepared in advance for each measurement item is converted into a deviation value indicating how far the measured value is from the average value. Even if there is a large difference in the standard deviation of the distribution, the degree of deviation from the average can be evaluated for each item. Here, the item having the measurement value farthest from the average: 50 is the person's largest characteristic item. Alternatively, the determination may be made using a threshold such as “80 or more” or “20 or less”. Incidentally, the number of maximum feature items does not need to be one, and in the case of a plurality of items, weighting or ranking may be given according to the deviation value.

According to a fifth aspect of the present invention, such a statistical distribution is newly corrected by increasing the frequency of a newly measured user measurement value, and the current statistical distribution is learned from past data in a learning manner. To reflect. In the statistical distribution for each measurement item, the vertical axis is frequency and the horizontal axis is a histogram distribution of measured values. As a procedure, when a measurement amount of a new user is obtained, the frequency of the measurement value in the histogram is counted up by one. If this is performed for each item every time a new user is generated, the statistical distribution gradually becomes along with the entire real user.

The image attribute information collating unit 26 performs collation with the image attribute information in the character image database 27 using the plurality of maximum characteristic items obtained by the maximum characteristic item extracting unit 25 as collation target items.

The character image database 27 stores accompanying attribute information together with the character images.

The character image selection / output unit 28 selects and outputs a character image from the character image database 27 based on the collation result of the image attribute information collation unit 26.

[0033]

As described above, according to the first aspect of the present invention, an impressive portion is extracted from a face or body image even when information indicating the character or emotion of the target person cannot be obtained. Since the character image is selected based on this, the character image that is easily impressive can be easily used from among a large number of images without being influenced by personal assumptions. Thereby, communication that both the sender and the receiver of the image can enjoy can be realized.

According to the second aspect of the present invention, even when an image of a target person's face or body cannot be obtained, an impressive character or emotion is extracted from a questionnaire result or the like, and a character image is selected based on the extracted character or emotion. Therefore, it is possible to use an impressive character image that is not influenced by the superficial appearance of the person and easily reflects the character and emotion from a large number of images. Thereby, communication that both the sender and the receiver of the image can enjoy can be realized.

According to the third and sixth aspects of the present invention, an impressive portion is extracted from an image of the face or body of the target person, and an impressive character or emotion is further extracted from a questionnaire result or the like. Therefore, since the impression depth is comprehensively determined and the character image is selected based on the impression depth, the strongest personality of the person can be reflected, and the impressive character can be easily selected from among a large number of images. Images can be used. Thereby, communication that both the sender and the receiver of the image can enjoy can be realized.

Further, in the invention according to the fourth aspect, since each of the measured quantities is converted into a deviation value based on a general statistical distribution and determined, a measurement value of an item having a different statistical distribution can be obtained. , An appropriate impression degree can be determined.

According to the fifth aspect of the present invention, each time a different person is used, the measured amount can be learned and reflected in the statistical distribution, so that even if the impression of the actual user is biased, In this way, the impression of the entire user can be clearly expressed. Also, in realizing this patent,
It is possible to automatically obtain a more realistic statistical distribution without having to investigate the statistical distribution of each item in advance, and to perform a labor-saving and practical impression judgment.

[0038]

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image extraction device according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating an image extracting apparatus according to an embodiment of the present invention in terms of functions.

[Explanation of symbols]

 Reference Signs List 1 image processing device 8 data bus 9 removable storage medium such as FD 11 CPU 12 memory 13 image memory 14 image processing unit 15 fixed disk 16 display unit 17 operation instruction unit 18 reading unit 19 network interface unit 121 program 191 internet 192 base station 193 Mobile phone 194 Image input unit 195 Operation unit 196 Display unit 21 Image input unit 22 Image measurement unit 23 Questionnaire input unit 24 Profile measurement unit 25 Maximum feature item extraction unit 26 Image attribute information collation unit 27 Character image database 28 Character image selection / output Department

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 AA20 BA02 CA01 CA08 CA12 CB01 CB08 CB12 CC03 CH07 DA12 DA20 DB02 DB06 DB09 DC02 5L096 AA13 BA20 CA02 FA70 FA81 HA07

Claims (6)

[Claims] An apparatus for extracting an image that impresses a person from a large number of character images, the apparatus including means for holding a measured amount of an impressive item as attribute information of the character image in advance, and Means for inputting an image of a face or body, means for extracting a measured amount of each part of the face or body from the input image, means for determining an impressive part from each of the measured amounts, An image extracting apparatus, comprising: means for matching attribute information of a character image with only a part as a matching target; and means for extracting a character image based on the result of the matching. 2. An apparatus for extracting an image that impresses a person from a large number of character images, comprising means for storing in advance attributed information of the character image, a measurement amount of an impressive item, Means for inputting information related to personality and emotion; means for extracting a measured amount of each profile from the input content; means for determining an impressive profile from each of the measured amounts; and the impressive profile An image extracting apparatus, comprising: means for matching attribute information of a character image with only the object to be matched; and means for extracting a character image based on the result of the matching. 3. An apparatus for extracting an image that impresses a person from a large number of character images, the apparatus including means for holding a measured amount of an impressive item as attribute information of the character image in advance. Means for inputting an image of a face or body, means for extracting a measured amount of each part of the face or body from the input image, means for inputting information relating to the character or emotion of the target person, and Means for extracting the measured amount of each profile from the contents obtained, means for determining items such as impressive parts and profiles from the measured amounts, and attribute information of the character image with only the impressive items as collation targets. And a means for extracting a character image based on the result of the comparison. 4. A method for determining an impressive item in an image extracting apparatus according to claim 1, wherein each of the measured quantities is converted into a deviation value based on a general statistical distribution. An image extraction device characterized by the above-mentioned. 5. The image extracting apparatus according to claim 4, further comprising means for learning a measured amount each time a different person is used. 6. A method for extracting an image that impresses a person from a large number of character images, the method including a step of retaining a measured amount of an impressive item as attribute information of the character image in advance, and A step of inputting an image of a face or a body, a step of extracting a measured amount of each part of the face or the body from the input image, a step of inputting information related to the character or emotion of the target person, and Extracting the measured amount of each profile from the extracted contents, determining an item such as an impressive part or a profile from the measured amount, and comparing only the impressive items with the attribute information of the character image. And extracting a character image based on the result of the comparison.