JP2009003581A - Image storage-retrieval system and program therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable a user to easily perform image retrieval by extracting image data suitable for an amount of color senses on the basis of a color sense language. <P>SOLUTION: In an image storage-retrieval system, when image data is input, an image storage computation unit 11 stores the image data in an image data storage unit 15. A color sense analysis computation unit 10 analyzes an input signal, outputs color sense data, and stores the sense data in a color sense data storage unit 14. A name list creation computation unit 12 creates a name list of the sense data and stores the name list in a color sense data name list storage unit 16. A color sense data name list output computation unit 23 makes an output device 19 display color sense language data on the basis of the name list of the sense data. When a user selects the language data displayed on the device 19, an image retrieval computation unit 24 retrieves the image data corresponding to the selected language data from all items of the image data stored in the storage unit 15 to extract the retrieved image data. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  According to the present invention, data obtained by performing color name analysis on digital image data is scored according to the color degree of the digital image data, and the color names are displayed by rearranging the data in descending order of the score. The present invention relates to a data search system and a program used for the system.

  With the spread of digital cameras, a large amount of digital image data is accumulated in personal computers. These image data are saved in a folder in the local area and used by the author, or saved in a folder that can be uploaded and shared and used on the Internet by the author and users other than the author. Therefore, conventionally, as a system for extracting specific image data desired by the user from image data on the Internet, the uploaded image data is color-analyzed and stored in a database, and based on the color characteristics of the image data Various techniques for extracting predetermined image data are known.

For example, in Patent Document 1, the RGB values of the search source image data and the search target image data are calculated, the RGB values in at least a part of the regions are compared, and the identity is determined based on the degree of approximation. A method for extracting specific image data from a plurality of image data obtained is disclosed.
In this method, the search source image data is set by inputting the image data stored in the hard disk of the personal computer. Note that the image data to be searched from the database can be limited by inputting a keyword, a category, or the like at the time of setting.

  Patent Document 2 analyzes RGB values and CMYK values for each bit of image data, calculates an average characteristic color by calculating an average value of the entire image data, and describes colors and emotions describing the image data. A method is disclosed in which one element of image data or a word representing color or emotion is input and the input element and the other element are mechanically extracted. Further, in this method, when capturing image data to be stored in the database, an image is captured by placing a strap having a plurality of color filters in front of the imaging lens of the portable communication terminal and storing the image in the database. By making the data as a whole colored, the calculation accuracy of the average characteristic color is improved.

Further, Patent Document 3 discloses the color, color distribution, brightness, target range, and the like of image data to be searched when searching for image data stored in a database after analysis of hue, color distribution, brightness, target range, and the like. A method is disclosed in which a search condition obtained by combining a plurality of conditions is stored with an arbitrary name, the name is designated, a corresponding search condition is called, and image data is retrieved from a database. Yes.
JP 2003-233800 A JP 2005-182703 A JP 2000-207408 A

  However, in the method described in Patent Document 1 described above, search target image data similar to the search source image data is searched by inputting the search source image data. When searching, it is necessary to search for similar search source image data first, which is troublesome. In addition, when a function for limiting image data is provided in this method, the user needs to review the contents and input the above-described keywords and categories when uploading the image data. Therefore, if the captured image data is image data that is greatly different depending on the user, the keyword or category to be input is different, and the image data cannot be limited appropriately depending on the keyword or category input at the time of search. There is a possibility that a simple image search cannot be performed.

  Further, in the method described in Patent Document 2, image data varies depending on the color and size of a symmetrical object to be imaged. Therefore, the image data captured using the strap is stored in a database. Even so, the average feature color and the above word may not be associated as intended by the user. Therefore, even if a word representing a color is input, image data representing that color cannot be extracted. In addition, it is necessary for the user to manually input words representing colors and emotions as text data of e-mail with image data, and the work is complicated, and the association may not be performed properly due to an input error.

  In the method described in Patent Document 3, first, it is necessary to set a search condition by inputting a hue and a color distribution as RGB values, and it cannot be directly input with a color sense for human-owned image data. It takes time and effort. In addition, even if the name of the color indicated by the RGB value is added to the search condition and stored, the name of the color may not match the name indicating the sense of the entire image data owned by humans. There is a possibility that the image data required by the user cannot be accurately extracted even if the search is performed with the stored search conditions.

  Here, when humans perceive the features of the entire image data sensibly, color names of about 2130 words in Japanese and about 7500 words in English are recorded as languages relating to the color features. Yes. Therefore, based on the background described above, using such color names, it is possible to easily retrieve images from a huge database by extracting image data having image features quantified appropriately for the color names. It is hoped that.

  The present invention has been made to meet the above-described demand, and is based on a language (color-sensing language) used for expressing a sense size (color sense amount) that humans feel subjectively with respect to hues. An object of the present invention is to provide an image storage / retrieval system and an image storage / retrieval system program that enable a user to easily perform image retrieval by extracting image data suitable for the sense.

  In order to achieve the above object, the invention of claim 1 is an image storage / retrieval system comprising an image storage device and an image search device, wherein the image storage device analyzes an input signal from an image input means. Color sensation analysis means for quantitatively associating a predetermined color sensation language related to color with evaluation and outputting color sensation data comprising color sensation points; and the color sensation analysis means for outputting the color sensation data. A color sensation data storage unit that stores color sensation data, an image data storage unit that stores image data corresponding to the color sensation data, and all image data stored in the image data storage unit. Sorting according to the correspondence frequency, color sense data name list creating means for creating a name list of the color sense data, and the color sense data name list creating means A color sensation data name list accumulation unit for accumulating a name list of sensation data, wherein the image search device extracts the color sensation data from the color sensation data name list accumulation unit, and a name list of the extracted color sensation data The first display means for displaying the color sensation language data corresponding to the color sensation data based on the image data, and extracting the image data from the image data storage unit, and for each image data for one image, Rearrangement means for rearranging the corresponding color sensation data in descending order of the color sensation score; and second display means for displaying the color sensation language data corresponding to the color sensation data rearranged by the rearrangement means; Selection for the user to select the color sensation language data displayed by either the first display means or the second display means And stage, on the basis of the color sense language data selected by said selecting means, in which and a retrieval means for retrieving the image data stored in the image data storage unit.

  The invention of claim 2 is the invention of claim 1, wherein the color sensation analysis means has a color sensation function for calculating the number of color sensation points corresponding to each of the color sensation languages, and the color sensation function is calculated. By changing, the color degree of the color sense function is changed to the color degree corresponding to the composite color sense language in the same color sense space.

  According to a third aspect of the present invention, in the invention of the second aspect, the color sensation analysis means changes the color sensation function according to the amount and integration degree of colors included in the image data and the position of the image plane. To do.

  According to a fourth aspect of the present invention, in the invention according to any one of the first to third aspects, the image search device includes one or more of the color sensation data rearranged by the rearrangement unit. The color arrangement pattern display means for displaying the color arrangement pattern of the image data composed of the color sense data, and the image data composed of the same color arrangement pattern as the color arrangement pattern of the image data displayed by the color arrangement pattern display means are searched. Image data retrieval means is further provided.

  According to a fifth aspect of the invention, in the invention according to any one of the first to third aspects of the invention, the image search device is configured so that each of the color sensation data is the number of the color sensation points for all image data. Regardless of the color retention rate measuring means for measuring the color retention rate indicating how frequently it exists in all image data, and the color retention rate measured by the color retention rate measurement means. Accordingly, a size setting means for setting a character size of the color sensation language data displayed by the first display means, and a character color of the color sensation data displayed by the first display means, It further includes a hue setting means for setting a hue according to the color sensation data.

According to a sixth aspect of the present invention, there is provided a program for an image storage / retrieval system comprising an image storage device constituted by a computer and an image search device. A color sensation analysis step for quantitatively associating and relating to an associated predetermined color sensation language to output color sensation data composed of the number of color sensation points, and the color sensation data output in the color sensation analysis step Color sensory data storage step for storing in the color sensory data storage unit, image data storage step for storing image data corresponding to the color sensory data in the image data storage unit, and storage of the color sensory data in the image data storage unit A color sense data name list creation step for rearranging according to the frequency of correspondence with all the image data and creating a name list of the color sense data; Performing a color sensation data name list accumulation step of accumulating the name list of the color sensation data, and causing the image search device to extract the name list of the color sensation data from the color sensation data name list accumulation unit, A first display step of displaying the color sensation language data corresponding to the color sensation data based on a list;
A rearrangement step of extracting image data from the image data storage unit, rearranging the color sensation data corresponding to the image data for each image data in descending order of the number of color sensation points, and the rearrangement A second display step for displaying the color sensation language data corresponding to the color sensation data rearranged in steps; and the color sensation language data displayed in either the first display step or the second display step. A search step for searching the image data stored in the image data storage unit using the color sensation language data selected by the user from among the above.

  According to the first aspect of the present invention, a language that can naturally imagine the characteristics of the hue of an image, such as a commonly used color that has been used for a long time, is determined in advance as the color sensation language, and is based on the RGB values of the input signal. Since this color sense language is quantitatively associated with evaluation, the user can easily and accurately acquire image data suitable for his / her sense. In addition, since the color sensation language data is displayed based on the name list of the color sensation data, the user can intuitively and quickly select the name of the required image data, and the image data can be acquired more easily. Can do. Further, since the color sensation data in which the color sensation points are rearranged in the descending order is displayed for each image data of one image, image data similar to the acquired image data can be easily obtained. In addition, since the color sensation analysis unit analyzes the input signal and quantitatively associates and evaluates the color sensation language to output the color sensation data, an image input from a digital camera or a mobile phone as the image input unit Data characteristics are converted into physical quantities, and various data are converted based on the physical quantities. Therefore, it is possible to accurately construct a database storing various data associated with the user's color sensation amount. Furthermore, since various data are automatically accumulated by inputting image data, it is possible to construct or manage a database at low cost without requiring artificial processing.

  According to the invention of claim 2, when the color degree of the color sense function is changed to the color degree corresponding to the composite color sense language, the number of color sense points corresponding to the composite color sense language is calculated, Since more detailed color sensation data is output, the user can select in detail the color sensation language data suitable for his / her sense, and more easily obtain the required image data.

  According to the third aspect of the present invention, the color sense data is obtained by calculating the number of color sense points reflecting the amount of color and the degree of integration included in the image data and the effect of the color sense amount due to the position of the image plane, and further subdividing the color sense data. Therefore, the user can accurately acquire image data suitable for his / her feeling.

  According to the invention of claim 4, the image retrieval device can retrieve image data composed of a color arrangement pattern having the same order as the color arrangement pattern of the acquired image data, so that the user is similar to the acquired image data ( Image data having a similar color configuration can be obtained more easily.

  According to the invention of claim 5, since the character size of the color sensation language data displayed by the first display means is changed according to the color holding ratio, the user can change the predetermined color sensation data. Since the color sensation language data can be selected while checking the capacity of the corresponding image data, the efficiency of image search can be increased. In addition, since the character color of the color sensation language data is set to a hue corresponding to the color sensation data, the user can more intuitively and quickly select the color sensation language data corresponding to the required image data.

  According to the sixth aspect of the present invention, a program for an image storage / retrieval system for allowing a user to easily acquire image data suitable for his / her sense is obtained.

  An image storage / retrieval system according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows a schematic configuration of an image storage / retrieval system according to the present embodiment. The image storage / retrieval system 1 is based on a language (color sensation language) used to express a sense of human feeling (color sensation amount) subjectively with respect to hues. An image storage device 2 that extracts data and executes an image storage process for storing image data, and an image search device 3 that executes an image search process for searching for image data. The The image storage device 2 and the image search device 3 are connected to be able to communicate with each other.

  The color sensation language is a language used when expressing the color sensation amount as described above, in other words, a language related to color. As the color sense language, the color name defined as “object color name” in JIS Z8102 is used, and ten basic chromatic color names such as “red”, “blue”, “green”, etc. , Achromatic names such as “black” and “white”, and such chromatic names and achromatic color names with lightness / saturation modifiers such as “bright”, “dull”, “dull”, etc. There are 269 kinds of conventional color names which are colors that cannot be handled systematically, such as “Azalea Color”, “Cherry Blossom Color”, “Rose Red”.

(Configuration of image storage device)
The image storage device 2 includes a computer 4, an image input device 5 that inputs image data to the computer 4, and an output device 6 that displays data output from the computer 4. As the computer 4, a personal computer, a workstation, or the like is applicable, and these may or may not be configured as a server. Examples of the image input device 5 include a digital camera, a memory stick, an SD card, a digital scanner, and the like, and these are connected to the computer 4 by USB. Examples of the output device 6 include a CRT display and a TFT liquid crystal display.

  The computer 4 stores the program and image data in a hard disk, an optical disk, a floppy (registered trademark) disk, or the like, and processes them during the calculation. A storage device 8 that temporarily stores data in a memory, and a bus 9 that is a path for exchanging data in the central processing unit 7, the storage device 8, the image input device 5, and the output device 6.

  The central processing unit 7 analyzes the input signal from the image input device 5 and quantitatively associates it with a predetermined color sensation language so as to output color sensation data consisting of the number of color sensation points. Unit 10, image data corresponding to color sensation data, that is, image storage operation unit 11 for storing image data input from image input device 5 in image data storage unit 15 described later, and color sensation data as image data storage unit 15 includes a name list creation calculation unit 12 that creates a name list of color sensation data by rearranging according to the frequency of correspondence with all image data stored in 15, and outputs the name list.

  The storage device 8 is output from an accumulation calculation program storage unit 13 that stores programs used in the color sensation analysis calculation unit 10, the image accumulation calculation unit 11, the name list creation calculation unit 12, and the like, and the color sensation analysis calculation unit 10. A color sensation data storage unit 14 for storing color sensation data, an image data storage unit 15 for storing image data, and a color sensation data name list storage unit 16 for storing a name list output from the name list creation calculation unit 12. With.

(Configuration of image search device)
The image search device 3 includes a computer 17, a search language input device 18 for inputting a color sensation language as a search language, a color arrangement pattern to be described later, and the like, and an output device 19 for displaying data output from the computer 17. Consists of. Examples of the computer 17 and the output device 19 include those similar to the computer 4 of the image storage device 2 described above. Examples of the search language input device 18 include a mouse, a trackball, and a keyboard for selecting a color sensation language displayed on the output device 19.

  The computer 17 includes a central processing unit 20, a storage device 21, and a bus 22 like the computer 4 of the image storage device 2 described above.

  The central processing unit 20 extracts color sensation data from the color sensation data name list storage unit 16 and causes the output device 19 to display color sensation language data corresponding to the color sensation data based on the name list of the extracted color sensation data. An image from which the image data stored in the image data storage unit 15 is searched and the image data is extracted based on the color sense data name list output calculation unit 23 and the content selected by the user using the search language input device 18 The search calculation unit 24, the image data output calculation unit 25 that causes the output device 19 to display the image data extracted by the image search calculation unit 24, and color sensation data corresponding to the image data extracted by the image search calculation unit 24 The color sensory data rearrangement calculation unit 26 rearranges the color sense data in descending order, and the color sensory data rearrangement calculation unit 26 performs the rearrangement. One of the color sensation language data output calculation unit 27 that causes the output device 19 to display color sensation language data corresponding to the color sensation data and the color sensation data rearranged by the color sensation language data output calculation unit 27 The character size of the color sensation language data displayed by the color arrangement language data output calculation unit 27 and the color arrangement language output calculation unit 27 for displaying the color arrangement pattern of the image data composed of the color sensation data described above on the output device 19 is predetermined. A character size setting calculator 29 for setting the character color of the color sensation language data displayed by the color sensation language data output calculation unit 27 to a predetermined color.

  The storage device 21 is used in the color sensation language data output calculation unit 27, the image data output calculation unit 25, the color arrangement pattern output calculation unit 28, the character size setting calculation unit 29, the character size setting calculation unit 29, and the image search calculation unit 24. A search calculation program storage unit 31 for storing a program is provided.

  By the way, the color sensation is not represented by RGB values used in camera output images, computer monitors, or the like, but may be represented by HSI values based on three attributes of hue (Hue), saturation (Saturation), and brightness (Intensity). Are known. Therefore, the color name defined in the above-described JIS Z8102 is expressed along the modified Munsell color system based on the HSI value as a language for expressing color sensation.

  Of the color names defined in JIS Z8102, the area of the color space representing the color name is ambiguously defined for the system color name, and the representative color symbol for the color name is defined for the conventional color name. Only the hue, brightness, and saturation are defined, and the area of the color space is not defined. In addition, regarding both the system color name and the conventional color name, the degree of feeling corresponding to the color name in the area of the color space is not defined.

  Therefore, in this system using such a color name as the color sensation language, in order to analyze the input signal from the image input device 5 and to quantitatively evaluate and associate the color sensation language with the color sensation language, the color sensation amount is set to HSI. In the color sensation space expressed in color space, the range of color sensation language and the degree of color sensation corresponding to the color sensation language within the range are set, and the amount of color sensation is quantified by the color sensation language and the number of color sensation points. It is necessary to keep.

  Here, quantification of the color sensation amount (color sensation amount related to hue) expressed by the color sensation language related to hue will be described. Examples of color sensation languages related to hue include “red”, “green”, and “blue”. FIG. 2 shows the color sensation space described above, and the color sensation space (“green” color sensation space) for the color sensation amount expressed by the “green” color sensation language. The shading in the drawing depicting the color sensation space indicates the degree (large or small) of the color sensation amount. In the color sensation space in this figure, the amount of color sensation at the center is the largest, and the amount of color sensation decreases as the distance from the center increases.

  To set the range of the color sensation language and the degree of color sensation amount corresponding to the color sensation language within the color sensation space, which is a three-dimensional space like this, the saturation in the color sensation space is fixed. In addition, setting is performed for each color plane by using a plurality of two-dimensional color planes composed of hues. This color plane is a cross-section obtained when the lightness is divided into 10 levels at equal intervals, and the color sense space is cut horizontally at the divided lightness levels, and hue and saturation are defined as vertical and horizontal axes. .

  FIG. 3 shows a part of a color plane having a lightness level of 5 for the “green” color sensation space shown in FIG. A method for setting a color sensation area that is a range of a color sensation language will be described with reference to FIG. First, the maximum boundary threshold values H2 and H1 that determine areas that are strongly felt as “green” in the hue, and the minimum boundary threshold values H2 ′ and H1 ′ that determine areas that are not felt as “green” are determined. Similarly, in the saturation, the maximum boundary threshold values S2 and S1 that define a region that is strongly felt as “green” and the minimum boundary threshold values S2 ′ and S1 ′ that are not felt as “green” are determined. The boundary lines H2 ', H2, H1, H1', S2 ', S2, S1, and S1' determined in this way determine the area lines that divide the color plane, and the color sensation area can be set.

  By the way, such a region line has not been defined so far, and the degree of the color sensation amount varies depending on the position in the region. Therefore, determination of the maximum boundary threshold and the minimum boundary threshold is performed by measuring the color displayed on a calibrated monitor or the like using a visual colorimetry method under a certain condition. Table 1 shows boundary threshold values H2 ', H2, H1, H1', S2 ', S2, S1, and S1' for each brightness rank in the "green" color sensation space.

  Further, a method for setting the degree of color sensation amount corresponding to the color sensation language in the color sensation region and scoring the degree of color sensation amount as the number of color sensation points will be described. This scoring is performed by associating boundary threshold values H2 ′, H2, H1, H1 ′, S2 ′, S2, S1, and S1 ′ with reference color sensation points to calculate the color sensation points corresponding to the color sensation language. This is done by determining the color sensory function. At this time, a saturation color sensation function that calculates the saturation color score that evaluates the degree of the color sensation amount by the saturation value, and a hue color sensation function that calculates the hue color score that evaluates the degree of the color sensation amount by the hue value To decide. Here, the color sense score is calculated by multiplying the saturation color score and the hue color score, as will be described later. In the color plane of FIG. 3, an area that is strongly felt as “green” in both hue and saturation is an area having a color sensation score of 1.0.

  FIG. 4 shows a saturation color sensation function and a hue color sensation function at a certain lightness level. The saturation color sensation function is connected with a curve between H2 'and H2, with the saturation color point number at H2', H1 'being 0 value, and the saturation color point number at H2, H1 being 1.0 value. H2 to H1 are connected by a straight line, and H1 to H1 ′ are connected by a curve. As a curve connecting H2 ′ to H2 and between H1 to H1 ′, a quadratic curve may be mentioned, and a saturation color between H2 ′ and H2 depending on what value is set for the coefficient of the quadratic curve. The accuracy of the score varies.

  The hue color sensation function also corresponds to the case of the saturation color sensation function, and the hue color score at S2 ′ and S1 ′ is 0, and the hue color score at S2 and S1 is 1.0. , S2 ′ to S2 are connected by a curve, S2 to S1 are connected by a straight line, and S1 to S1 ′ are connected by a curve. These saturation color sense functions and hue color sense functions are determined for all lightness levels from 0 to 10. Thereby, the degree of the color sensation amount can be scored as the color sensation score, and the color sensation amount relating to the hue is quantified.

  Next, quantification of the color sensation amount (saturation and lightness color sensation amount) expressed by the color sensation language relating to saturation and lightness will be described. Examples of color sense languages relating to saturation and lightness include “brilliant”, “bright”, “koi”, “dull”, “soft”, and the like. In quantifying the color sensation amount, when setting the range of color sensation language and the degree of color sensation amount corresponding to the color sensation language in the color sensation space, quantification of the color sensation amount expressed by the color sensation language related to hue Similarly to the conversion, a boundary threshold value is determined for each color plane.

  FIG. 5 shows part of a color plane with a lightness level of 5 for a “dull” color sensation space. Here, the color sensation amount related to saturation and lightness is the size of the sensation that humans feel in common with all hues, so that the color sensation region is compared with the color sensation region of the color sensation amount related to hue, It becomes a wider area. In the case of a color sensation amount expressed mainly by a color sensation language relating to saturation, such as “brilliant”, “good”, “dull”, etc., a color sensation area is set on a color plane of the same lightness level. . However, if the hue is different, even if the saturation is the same, the magnitude of the color perception amount of the saturation is different, so that the area line in the color plane is not a straight line. Further, in the case of a color sensation amount expressed mainly by a color sensation language relating to lightness, such as “light” and “dark”, only one color sensation area is set under the same lightness level.

  After setting the color sensation area, the degree of color sensation amount corresponding to the color sensation language is set in the color sensation area, and the degree of color sensation amount is scored as a color sensation score. Thereby, the amount of color sensation relating to saturation and lightness is quantified.

  As described above, by setting the range of the color sensation language and the degree of the color sensation amount corresponding to the color sensation language within the range, the color sensation space corresponding to the color sensation language related to hue, and the saturation and lightness A color sense space corresponding to the color sense language is defined in the HSI color space. A new color sensation space can be defined by combining such defined color sensation spaces. For example, a color sense space such as “bright green” or “brilliant green” can be defined.

  By the way, when the color sensation language is modified into a compound word, the color sensation function is changed according to the color sensation language, and the color sensation points are changed by quantifying the color sensation amount described above. In other words, the changed color sensory function is not a color sensory space in which a color threshold region is set with a clearly recognizable boundary threshold, such as “green” color sensory space. This is used for quantifying the amount of color sensation in the case of a color sensation space in which a color sensation area is set with a boundary threshold around the center, such as a “colorful” color sensation space.

  FIG. 6 shows a method for changing the color sensory function. For example, in the color sensation point curve a with a weight of “A color”, when the color sensation point of “A color” with a weight of “A color” is 0 to 1.0, the number of color sensation points is By moving the part a in the 1.0 value around the 0.8 color sense point, and setting the color sense point at which the color sense point defined by “A color” is 1.0 value to be low, The number of color sensation points in the color sensation area around the center increases, and the weighting of the color sensation points is changed. The curve set in this way is a color sensation point curve b with a weight of “A-like”. By changing the color sensation function in this way, the number of color sensation points can be changed. Further, by making such a change, an “A-color-like” color sensation space is defined, so that “A color” is similar to a “deep A color” or “thin A color” color sensation space. On the other hand, a fine color sensation space can also be defined by changing the color sensation function.

  Here, in order to correct the number of color sensation points according to the amount and integration degree of colors included in the image data and the position of the image plane, a position correction function for correcting the number of color sensation points is set. In order to correct the color sensation score according to the amount of color included in the image data, it is considered how much the calculated color sensation score of each pixel contributes to the color sensation amount of the entire image data. FIG. 7 shows a position correction function set in that case. If this position correction function is used, a pixel whose “A” color sensation score is lower than 0.5 is considered not to contribute to the color sensation of the entire image data, and the “A” color sensation score is corrected to zero. Is done.

  In order to correct the number of color sensation points according to the position of the image plane, the degree of influence of the position of the subject on the image data on the color sensation amount is considered. FIG. 8 shows the degree of influence on the image data. Usually, when an image is taken with a digital camera or the like, the image is often taken with the aim of the digital camera positioned at the center of the subject. For example, when a whole image of a plant in which small “red flowers” are covered with many “green leaves” is captured, the “red flowers” are often located at the center of the image data. When the color sensation points of the entire image data are calculated by the above-described method, the “green” color sensation points are larger than the “red” color sensation points. However, when a person views the image data, the image data In order to pay attention to the “red flower” located in the center of the “red”, the degree of impression of “red” increases and the amount of color sense of “red” increases. Therefore, the degree of influence of the position of the subject on the image data on the color sensation amount is dealt with by changing the weighting of the color sensation points in the central part and the peripheral part of the image data. At this time, a position correction function is set as in the case of correcting the number of color sensation points according to the amount of color included in the image data.

  Next, the operation of the entire system in the present embodiment configured as described above will be described. In the image storage / retrieval system 1, an image storage process for storing image data and an image search process for searching for image data are executed. The image search device 3 operates.

(Operation of image storage device)
FIG. 9 shows an operation procedure of the image storage device 2 in the image storage process. First, when image data is input from the image input device 5, the image accumulation calculation unit 11 accumulates the image data in the image data accumulation unit 15, and executes a process of saving the accumulated image data (# 1). Next, the color sensation analysis calculation unit 10 analyzes the input signal from the image input device 5, outputs the above-described color sensation data, and stores the color sensation data in the color sensation data storage unit 14 (color sensation analysis). Process) (# 2). Next, the name list creation calculation unit 12 executes a process (name list creation process) for creating the name list and storing the name list in the color sensation data name list storage unit 16 (# 3). The color sensation analysis process and the name list creation process will be described in detail later. Each process of # 1, # 2, and # 3 may be executed by inputting image data from the image input device 5, and may be executed simultaneously in parallel.

  Here, the color sensation analysis processing will be described. FIG. 10 shows an operation procedure of the color sensation analysis calculation unit 10 in the main color sensation analysis processing.

  First, the RGB value of one pixel in the image data is acquired from the input signal of the image input device 5 (# 11). Next, a color sensation score A based on brightness and hue is calculated from the acquired RGB value of one pixel (# 12). The color sense score A is the hue color score described above.

  Here, calculation of the color sensation score A will be described. The obtained RGB value of one pixel is converted into an HSI value, and the hue sensation function described above is used for the converted HSI value to calculate the color sensation point number A. In the present embodiment, as a conversion equation for converting RGB values into HSI values, a “conversion equation based on the HSI hexagonal pyramid color model” of a hexagonal pyramid based on the Ostwald color system as shown below is used.

Shark (pie): Circumference ratio (3.1415 ...)
max = MAX (R, G, B): maximum value of R, G, B mid = intermediate value of R, G, B min = MIN (R, G, B): minimum value H of R, G, B Range: 0.0-2Π Range of S: 0.0-1.0 Range of I: 0.0-1.0
The calculation formula of H varies depending on the RGB values.
When R>G> B, H = (mid-min) / (max-min) * Π / 3
When G>R> B, H = − (mid−min) / (max−min) * Π / 3 + (2Π / 3)
When G>B> R, H = (mid-min) / (max-min) *） / 3+ (2Π / 3)
In the case of B>G> R, H = − (mid−min) / (max−min) * （/ 3+ (4Π / 3)
When B>R> G, H = (mid-min) / (max-min) * ｍｉｎ / 3 + (4Π / 3)
When R>B> G, H = − (mid−min) / (max−min) * ｍｉｎ / 3 + (6Π / 3)
The calculation formula of S is S = max−min / max.
The formula for calculating I is I = max / 255.

  Note that several methods for converting RGB values into HSI values have been proposed in the “Digital Image Processing CG-ART Association”, and any of them may be adopted.

  Next, from the acquired RGB value of 1 pixel, the number B of color sensation points by brightness and saturation is calculated (# 13). The color sensation score B is the above-described saturation color score.

  Next, the calculated color sensation point number A is multiplied by the color sensation point number B to calculate the color sensation point number C (# 14). The color sensation score C is the color sensation score described above. In other words, when the lightness level is N, when the color sensation point number A is Pnh, the color sensation point number B is Pns, and the color sensation point number C is Pn, it can be expressed as Pn = Pnh × Pns.

  Next, the number of color sensation points C ′ is calculated by correcting the number of points of color sensation points C by using the “ish color” correction coefficient (# 15). That is, the color sensation score C ′ is the color sensation score for the compound language color sensation language, and the “ish color” correction coefficient is a curve set when changing the weighting of the color sensation score described above. , Corresponding to a color sensation point curve b with a weight of “A-like”.

  Next, position information on the image data in the target pixel is acquired, and (# 16) the color sensation point C is subjected to point correction using the position information and the position correction function (# 17), and the color sensation point C ′. Is corrected using the position information and the position correction function (# 17).

  Next, it is determined whether or not the processing from # 11 to # 18 has been completed for all target pixels in the image data (# 19). If the processing is complete (YES in # 19), one image data Calculate the color sensation score at. (# 20). If the processing has not been completed, the processing of # 11 to # 18 is executed for the other one pixel (NO in # 19).

  Here, calculation of the number of color sense points in one image data will be described. The number of color sensation points in one image data is an average value of the number of color sensation points in all target pixels with one image data as a collection of pixels. For example, in the case of image data of X pixels × Y pixels, there are (X * Y) pixels, and the color sense point PA of “A color” can be calculated for each individual pixel, so that the total (X * Y) Pieces of PA are obtained. The number of color sensation points in one image data of “A color” is calculated by the calculation formula {PA1 + PA2 +... + PA (X * Y)} / (X * Y). In this way, the color sense points in one image data for all color sense languages such as “red”, “red”, and “green” are calculated.

  Next, the number of color sensation points in one image data for all the color sensation languages is stored as color sensation data in the color sensation data storage unit 14 (# 21), and the color sensation analysis processing ends.

  Here, the name list creation process will be described. FIG. 11 shows an operation procedure of the name list creation calculation unit 12 in the name list creation process.

  First, one piece of image data is extracted from all the image data stored in the image data storage unit 15 (# 31), and color sense data of the extracted image data is extracted from the color sense data storage unit 14 (# 32). Next, the extracted color sensation data is sorted in descending order by color sensation score (# 33), and the top N color sensation data are extracted from the sorted color sensation data (# 34). This N can be set arbitrarily.

  Next, for each of the top N color sensation data, the number of appearances p obtained by counting the frequency of correspondence with all image data is acquired (# 35), and the number of appearances p is incremented for each color sensation data (# 36). . The number of appearances p is a value representing the degree of presence of each color sensation data with respect to all the image data stored in the image data storage unit 15. Therefore, the value of the number of appearances p is proportional to the number of image data having the color sensation data. .

  Next, it is determined whether or not the processing from # 31 to # 36 has been completed for all the image data stored in the image data storage unit 15 (# 37). YES), the appearance count p for each color sensation data is stored in the color sensation data name list storage unit 16 as a name list of the color sensation data (# 38). If the process has not been completed, the processes of # 31 to # 36 are executed for the other image data (NO in # 37).

(Operation of image search device)
FIG. 12 shows an operation procedure of the image search device 3 in the image search process. First, the color sensation data name list output calculation unit 23 performs a process (color sensation data name list analysis process) for causing the output device 19 to display color sensation language data corresponding to the color sensation data based on the name list of the color sensation data. Execute (# 41). The color sensation data name list analysis process will be described in detail later. Next, when the user selects the color sensation language data displayed on the output device 19 (# 42), the image search calculation unit 24 stores image data corresponding to the selected color sensation language data in the image data storage unit 15. Search is performed from all stored image data, and the searched image data is extracted (# 43).

  Next, the image data output calculation unit 25 causes the output device 19 to display the extracted image data (# 44). Next, the color sensation data rearrangement calculation unit 26 rearranges the color sensation data corresponding to the extracted image data in descending order of the number of color sensation points (# 45). Next, the color sensation language data output calculation unit 27 extracts the top N color sensation data from the rearranged color sensation data, and outputs the color sensation language data corresponding to the color sensation data to the output device 19. Display (# 46). This N can be set arbitrarily. Next, the color arrangement pattern output calculation unit 28 causes the output device 19 to display the color arrangement pattern of the image data composed of the extracted top N color sense data (# 47). The color arrangement pattern of image data is the distribution of shades constituting the image data such as 30% black, 25% white, 18% yellow-green, 12% brown, and 8% yellow-red (in this case, N = 5).

  FIG. 13 shows an example of a state in which the image data displayed in # 44, the color sense language data displayed in # 46, and the color arrangement pattern displayed in # 47 are displayed on the output device 19. Thus, the image data 32, the color sensation language data 33, and the color arrangement pattern 34 are displayed on the same screen, the image data 32 is located at the upper left of the screen, and the color sensation language data 33 and the color arrangement pattern 34 are at the lower right of the screen. positioned. Such a positional relationship can be arbitrarily set.

  If the user selects the color sensation language data displayed in # 46 or the color arrangement pattern displayed in # 47, the processes from # 43 to # 47 are executed again (YES in # 48). Here, when the color arrangement pattern displayed in # 47 is selected, the image search calculation unit 24 stores the image data composed of the same color arrangement pattern as the color arrangement pattern of the selected image data in the image data storage unit 15. Search is performed from all the stored image data, and the searched image data is extracted. If these are not selected by the user, the image search process ends (NO in # 48).

  Here, the color sensation data name list analysis process will be described. FIG. 14 shows an operation procedure of each calculation unit in the color sensation data name list analysis process. First, the color sensation data name list output calculation unit 23 extracts one color sensation data from the color sensation data name list storage unit 16 (# 51), and acquires the number of appearances p from the name list of the extracted color sensation data. (# 52). FIG. 15 shows an example in which color sense data to be extracted is output for each image data. Although not shown in the figure, the appearance count p is associated with the color sensation data.

  Next, the character size setting calculation unit 29 sets the character size of the color sensation language data corresponding to the extracted color sensation data to a predetermined size according to the acquired number of appearances p (# 53). Next, the character color setting calculation unit 30 sets the character color of the color sensation language data to a hue corresponding to the extracted color sensation data (# 54). The color sensation data name list output calculation unit 23 causes the output device 19 to display color sensation language data in which the character size and hue are set as described above (# 55).

  Next, it is determined whether or not the processing from # 51 to # 55 has been completed for all the color sensation data stored in the color sensation data name list storage unit 16 (# 56). (YES in # 56), the color sensation data name list analysis process is terminated, and if the process is not completed, the processes of # 51 to # 55 are executed for the other one of the color sensation data (in # 56) NO). FIG. 16 shows an example of a state in which the color sensation data name list analysis process is executed and the color sensation language data is displayed on the output device 19. The user selects predetermined color sensation language data from the list of color sensation language data.

  According to the image storage / retrieval system of the present embodiment having the above-described configuration, a language that can naturally imagine the characteristics of color tone of image data, such as conventional colors that have been used for a long time, is preliminarily used as a color sensation language. The color sense language is quantitatively associated with the evaluation based on the RGB values of the input signal from the image input device 5, and the user can easily and accurately acquire image data suitable for his / her sense. can do. Further, since the color sensation language data is displayed based on the name list of the color sensation data, the user can intuitively and quickly select the name of the required image data, and the image data can be acquired more easily. it can. Further, since the color sensation data in which the color sensation points are rearranged in the descending order is displayed for each image data of one image, image data similar to the acquired image data can be easily obtained. Further, the color sensation analysis calculation unit 10 analyzes the input signal and quantitatively associates the color sensation language with the evaluation and outputs the color sensation data. Therefore, input from a digital camera or a mobile phone as the image input device 5 The feature of the image data is converted into a physical quantity, and various data are converted based on the physical quantity. Therefore, it is possible to accurately construct a database storing various data associated with the user's color sensation amount. Furthermore, since various data are automatically accumulated by inputting image data, it is possible to construct or manage a database at low cost without requiring artificial processing.

  In addition, when the color degree of the color sense function is changed to a color degree corresponding to the composite color sense language, the color sense points corresponding to the composite color sense language are calculated, and a more detailed color sense is obtained. Since the data is output, the user can select in detail the color sensation language data suitable for his / her sense and can more easily obtain the required image data.

  In addition, the number of color sensation points reflecting the amount of color and the degree of integration included in the image data and the effect of the amount of color sensation due to the position of the image plane are calculated, and more detailed color sensation data is output. Can accurately acquire image data suitable for its own sense.

  In addition, since the image search device 3 can search for image data configured with a color arrangement pattern having the same order as the color arrangement pattern of the acquired image data, the user is similar to the acquired image data (the color configuration is similar). Image data can be acquired more easily.

  In addition, since the character size of the color sensation language data displayed on the output device 19 is set according to the number of appearances p acquired by the color sensation data name list output calculation unit 23, the user can set predetermined color sensation data. Since the color sensation language data can be selected while checking the capacity of the corresponding image data, the efficiency of image search can be increased. In addition, since the character color of the color sensation language data is set to the hue according to the color sensation data extracted by the color sensation data name list output calculation unit 23, the user can select the color sensation language corresponding to the required image data. Data can be selected more intuitively and quickly.

(Network image storage / retrieval system)
Here, a network-type image storage / retrieval system according to a modification of the image storage / retrieval system of the present embodiment will be described. FIG. 17 shows a schematic configuration of a network-type image storage / retrieval system according to this modification. The network-type image storage / retrieval system 100 is a system in which the image storage / retrieval system 1 of the present embodiment is configured as a network, and includes an image storage server 40, an image search server 41, and a distribution terminal 42. The image storage server 40 includes a communication device 44 that is connected to a network 43 and communicates various data, and a transmission / reception control unit 45 that controls the communication device. Similar to the image storage server 40, the image search server 41 is provided with a communication device 46 and a transmission / reception control unit 47 in the above-described image search device 3. The distribution / reception terminal 42 is equivalent to the image storage server 40 or the image search server 41. According to the network-type image storage / retrieval system of this variation having the above-described configuration, a huge database can be constructed by transmitting and receiving various data via the network, so that the user is suitable for his / her own sense. Image data can be acquired with higher accuracy.

(Specific examples of image storage processing and image search processing)
Here, the image storage process and the image search process will be described based on specific examples. As the computer 4 of the image storage device 2, OS: Windows XP (registered trademark), CPU: 64-bit Intel Xeon processor 2.80 GHz, memory: 1 GB, monitor: 20-inch TFT monitor, hard disk capacity: 250 GB Use. Further, as the computer 17 of the image search device 3, a computer with a hard disk capacity of 160 GB in the above specifications is used. A single-lens reflex Nikon D2X digital camera is used as a device for capturing images. With this digital camera, a natural landscape is photographed so that a large tree grows in the center of the frame. The captured image size is 2000 × 3008 pixels. A digital camera is connected to the image storage device 2 via USB, and the captured image data is input to the image storage device 2. Note that the input method to the image storage device 2 may be by mail or upload to a shared site from a remote location.

(About image storage processing)
After the image data is input from the digital camera (image input device 5), the image storage process is executed by starting the program stored in the storage operation program storage unit 13 of the storage device 8. The image data is stored by the image accumulation process, and the color sensation analysis process is executed. In the color sensation analysis processing, the color sensation analysis calculation unit 10 obtains RGB values for each pixel in order from the upper left end of the image data to the lower right end of the image data, and calculates the number of color sensation points. The calculation result regarding the pixel at the coordinates (300, −200) will be described in detail with the coordinates of the upper left corner of the image data as (0, 0).

Table 2 shows a list of color sense languages. The calculation of the color sensation score is performed for all the color sensation languages described in this table.

First, the RGB values (R = 18, G = 108, B = 84) of the pixel at coordinates (300, -200) were converted into HSI values. The RGB value of this pixel corresponds to the case of G>B> R. Because max = 108, mid = 84, min = 18,
H = (mid-min) / (max-min) * Π / 3 + (2Π / 3)
S = max-min / max
From the calculation formula of I = max / 255,
It is calculated that H = 2.862255554, S = 0.833333333, and I = 0.423529411.
Referring to Table 1, since the lightness level is between 4 and 5, it is determined that this pixel is located in the color sensation region between H2 ′ and H2 and between S2 ′ and S2.

Next, the color sensation score was calculated from the HIS value. Since the lightness level is between 4 and 5, the required saturation color sensory function and hue color sensory function are those with lightness levels of 4 and 5. First, it is applied to the “green” saturation color sensation function and the hue color sensation function to calculate the “green” color sensation score.
Hue color point number P4h = 0.12521896 with a lightness level of 4,
Since the saturation color point number P4s = 1.814058956 with a lightness level of 4,
Number of color sense points with a lightness level of 4 P4 = P4h × P4s = 0.385527248
Hue color point number P5h = 0.637427626 with a lightness level of 5,
Since the saturation color point number P5s = 1.147842054 with a lightness level of 5,
The number of color sensation points with a lightness level of 5 is P5 = P5h × P5s = 0.73666235.
When the “green” color sensation point d in the color sensation space is calculated based on the color sensation points of brightness levels 4 and 5, d = (P (N + 1) h × P (N + 1) s) × d1 + (PNh × PNs) × d2 = (0.731666235) × (0.423529411−0.4) × 10 + (0.385527248) × (0.5−0.423529411) × 10≈0.4670.
Further, if the “green” color sensation point d is corrected using the “ish color” correction coefficient, the “greenish” color sensation point d ′ becomes zero. Further, when the “green” color sensation point d and the “greenish” color sensation point number d ′ are corrected according to the amount of colors included in the image data, the “green” color sensation point d and “green” Both the color sense points d ′ of “Poi” are 0 values. When the point correction is performed using the position information on the image data in the pixel and the position correction function, the “green” color sensation point d and the “greenish” color sensation point d ′ both have zero values.

  In this way, the color sensation points are calculated for all the color sensation languages listed in Table 2. Table 3 shows a list of calculation results for the pixel at coordinates (300, -200).

  The numerical value shown in the item “number of color points” in the table is the calculated number of color sense points. In this pixel, the color sensation points of “Tokiwa Color”, “Dark Green”, “Moe Blue”, “Malachite Green”, “Forest Green”, “Viridian” and “Billiard Green” are larger than 0, "Reddish", "Yellow reddish", "Skinish", "Brownish," Yellowish "," Yellowish "," Greenish " , “Blueish”, “Blue”, “Blue-purple”, “Purple”, “Red-purple”, “White”, “Gray” and “Black” are all 0 Value.

The numerical value shown in the item “place conversion” in the table is the color sensation score obtained by correcting the calculated color sensation score according to the position of the image plane. Since this pixel is located at coordinates (300, -200) with the coordinates of the base point at the upper left corner as (0, 0), the coefficient for changing the weighting to the number of color sense points is 0.66. Therefore, by multiplying 2/3 (= 0.66666666) by the color sense score of “Tokiwa Color”, “Dark Green”, “Moe Blue”, “Malachite Green”, “Forest Green”, “Bilidian” and “Billiard Green” Then, the number of color sense points corrected according to the position of the image plane is calculated.
Color sense points before correction x Coefficient = Color sense points after correction
0.896722 × 2/3 = 0.597815 (“Tokiwa Color”)
0.813508 × 2/3 = 0.542339 (“fresh green”)
0.80604 × 2/3 = 0.537363 (“Moe-iro”)
0.868884 × 2/3 = 0.579256 (“Malachite Green”)
0.087777 × 2/3 = 0.058518 (“Forest Green”)
0.874187 × 2/3 = 0.582791 (“Viridian”)
0.016286 × 2/3 = 0.010857 (“Billiard Green”)

  In this way, the number of color sensation points for one pixel in the image data is calculated. Further, the color sensation points are calculated for all the target pixels in the image data, the average value of the color sensation points in all the target pixels is calculated, and this is set as the color sensation point number in one image data. The number of color sensation points is stored in the color sensation data storage unit 14 as color sensation data, and the color sensation analysis processing ends.

  When the color sensation analysis process ends, a name list creation process is executed. In the name list creation process, the name list creation calculation unit 12 extracts the color sensation data and sorts the color sensation data in descending order by the number of color sensation points. Further, the top N color sensation data are extracted from the sorted color sensation data. In this specific example, N = 5. As shown in Table 3, the top five color sensation data are “Tokiwa Color” (0.597815), “Viridian” (0.582791), “Malachite Green” (0.579256), “Dark Green” (0.542339), and “Moe Color” (0.537363). For each color sensation data, the number of appearances p obtained by counting the frequency of correspondence with all image data is acquired. Table 4 shows the data of the number of appearances p stored in the color sensation data name list storage unit.

  The appearance counts p of “Tokiwa Color”, “Viridian”, “Malachite Green”, “Dark Green”, and “Moeiro” are 898, 255, 46, 553, and 602, respectively. These appearance counts p are incremented individually, and the appearance counts p of “Tokiwa Color”, “Bilidian”, “Malachite Green”, “Dark Green”, and “Moe Blue” are set to 899, 256, 47, 554, and 603, respectively. The incremented appearance count p is stored in the color sensation data name list storage unit 16 as a name list of the color sensation data, and the name list creation process ends. As a result, the image accumulation process ends.

(About image search processing)
The image search process is executed by starting a program stored in the search calculation program storage unit 31 of the storage device 21. A color sensation data name list analysis process is executed by the image search process. In the color sensation data name list analysis process, the color sensation data name list output calculation unit 23 extracts one color sensation data from the color sensation data name list storage unit 16, and the above-mentioned appearance from the name list of the extracted color sensation data. The number of times p is acquired. The character size setting calculation unit 29 sets the character size of the color sensation language data to be output to the output device 19 such as a monitor according to the number of appearances p corresponding to the acquired color sensation data. In order to perform such setting, for example, when the appearance count p is 900 to 1000, the character size is 20 pt, and when the appearance count p is 800 to 899, the appearance size p is 18 pt. The correspondence with the character size is determined in advance. The character color setting calculation unit 30 sets the character color of the color sensation language data to a hue corresponding to the extracted color sensation data. For example, when the color sense language data is “malachite green”, “malachite green” is set as the character color. When the setting of the character size and the character color is completed for all the color sensation data, the color sensation data name list output calculation unit 23 causes the output device 19 to display the set color sensation language data. The order of extracting the color sensation data from the color sensation data name list storage unit 16 may be random or may be in the order of “aiueo” in the color sensation language data. Thus, the user can select color sense language data corresponding to the required image data from the output color sense language data with a mouse or the like, and search for the required image data.

  Next, the user selects the color sense language data of “Tokiwa Color”, and from the image data that uses a lot of “Tokiwa Color”, a natural landscape image is photographed so that a large tree grows in the center. When the image data is extracted, the image data output calculation unit 25 causes the output device 19 to display the image data. The color sensation data rearrangement calculation unit 26 rearranges the color sensation data corresponding to the image data in descending order of the number of color sensation points. In the specific example, the top five color sensation data are extracted, and the color sensation language data corresponding to the color sensation data is displayed on the output device 19. Further, the color arrangement pattern output calculation unit 28 acquires all the color sensation data included in the selected image data, and based on the color sensation point data, calculates the color arrangement pattern of the image data composed of the top five color sensation data. The color arrangement pattern is created and displayed on the output device 19. The user can acquire another image data having the same combination and ratio (color arrangement pattern) of color sense data as the displayed image data by selecting the color arrangement pattern. In other words, the image data that can be acquired by the user is not image data having an object similar to the object shown in the image data, but, for example, an image showing a house having a roof of “Tokiwa-color” or “Viridian” The image data is similar in color scheme such as an image of a woman wearing a dress, an image of the sea that looks like “emerald green”, and the like.

  In addition, this invention is not restricted to the structure of the said embodiment, A various deformation | transformation is possible in the range which does not change the meaning of invention. For example, the color sensation language data that is selected less frequently by the user may be configured not to be automatically displayed on the output device 19.

1 is a block diagram showing an image storage / retrieval system according to an embodiment of the present invention. The conceptual diagram which shows the color sense space of "green". The figure which shows the color plane about the color sense space of "green". (A) is a figure which shows the saturation color sense function which calculates a saturation color point number, (b) is a figure which shows the hue color sense function which calculates a hue color point number. The figure which shows the color plane about the color sense space of "dullness". The figure which shows the method of changing a color sense function according to a color sense language. The figure which shows the position correction function for correct | amending a color sense point number according to the quantity of the color contained in image data. The figure which shows the influence degree which the position of the to-be-photographed object on image data has on the amount of color sensations. 5 is a flowchart showing image storage processing in the image storage device. The flowchart which shows the color sense analysis process in an image storage process. The flowchart which shows the name list creation process in an image storage process. 6 is a flowchart showing image accumulation processing in the image search apparatus. The figure which shows the state by which the image data, the color sense language data, and the color arrangement pattern were displayed on the output device. The flowchart which shows the color sensation data name list analysis process in an image search process. The figure which shows the color sense data which a color sense data name list output calculating part extracts. The figure which shows the state by which the color sense language data was displayed on the output device. The block diagram which shows the network type image storage / retrieval system which concerns on the deformation | transformation form of the image storage / retrieval system.

Explanation of symbols

1 image storage / retrieval system 2 image storage device 3 image search device 4 computer (for storage)
5 Image input device 6 Output device (for storage)
7 Central processing unit (for storage)
8 Storage device (for storage)
9 bus (for storage)
DESCRIPTION OF SYMBOLS 10 Color sense analysis calculation part 11 Image accumulation calculation part 12 Name list creation calculation part 13 Storage calculation program storage part 14 Color sense data storage part 15 Image data storage part 16 Color sense data name list storage part 17 Computer (for search)
18 Search language input device 19 Output device (for search)
20 Central processing unit (for search)
21 Storage device (for search)
22 Bus (for search)
23 Color sense data name list output calculation unit 24 Image search calculation unit 25 Image data output calculation unit 26 Color sense data rearrangement calculation unit 27 Color sense language data output calculation unit 28 Color arrangement pattern output calculation unit 29 Character size setting calculation unit 30 Characters Color setting calculation section 31 Calculation calculation program storage section 32 Image data 33 Color sense language data 34 Color arrangement pattern 40 Image storage server 41 Image search server 42 Distribution receiver 43 Network 44 Communication device (for storage)
45 Transmission / reception controller (for storage)
46 Communication device (for search)
47 Transmission / reception controller (for search)
100 Network image storage and retrieval system

Claims (6)

An image storage / retrieval system comprising an image storage device and an image search device,
The image storage device includes:
A color sensation analysis unit that analyzes an input signal from the image input unit, quantitatively associates and evaluates a predetermined color sensation language related to color, and outputs color sensation data including color sensation points;
A color sensation data storage unit for storing the color sensation data output from the color sensation analysis means;
An image data storage unit for storing image data corresponding to the color sensation data;
Sorting the color sensation data according to the frequency of correspondence with all the image data stored in the image data storage unit, color sensation data name list creating means for creating a name list of the color sensation data,
A color sensation data name list storage unit for storing a name list of the color sensation data created by the color sensation data name list creation means;
The image search device includes:
First display means for extracting the color sensation data from the color sensation data name list storage unit and displaying the color sensation language data corresponding to the color sensation data based on a name list of the extracted color sensation data;
Reordering means for extracting image data from the image data storage unit, and reordering the color sensation data corresponding to the image data for each piece of image data in descending order of the number of color sensation points;
Second display means for displaying the color sensation language data corresponding to the color sensation data rearranged by the rearrangement means;
A selection means for a user to select the color sensation language data displayed by either the first display means or the second display means;
An image storage / retrieval system comprising: search means for searching image data stored in the image data storage section based on the color sensation language data selected by the selection means. The color sensation analysis means has a color sensation function for calculating the number of color sensation points corresponding to each of the color sensation languages,
2. The color sensation function is changed to change the color degree of the color sensation function to a color degree corresponding to the composite color sensation language in the same color sensation space. The image storage / analysis search system described in 1.   3. The image storage / analysis according to claim 2, wherein the color sensation analysis unit changes the color sensation function in accordance with the amount and integration degree of colors included in the image data and the position of the image plane. Search system. The image search device includes:
Color arrangement pattern display means for displaying a color arrangement pattern of image data composed of one or more of the color sensation data among the color sensation data rearranged by the rearrangement means;
4. The image according to claim 1, further comprising image data search means for searching image data composed of the same color pattern as the color pattern of the image data displayed by the color pattern display means. Accumulation / analysis search system. The image search device includes:
Color retention rate measuring means for measuring the color retention rate indicating how frequently the individual color sensation data is present in all image data regardless of the number of color sensation points for all image data; ,
A size setting means for setting a character size of the color sensation language data displayed by the first display means according to the height of the color holding ratio measured by the color holding ratio measuring means;
The hue setting means for setting the character color of the color sensation data displayed by the first display means to a hue corresponding to the color sensation data is further provided. Image storage / analysis search system. A program for an image storage / retrieval system comprising an image storage device and an image search device configured by a computer,
In the image storage device,
A color sensation analysis step of analyzing an input image signal and quantitatively associating a predetermined color sensation language related to color with an evaluation association and outputting color sensation data composed of color sensation points;
A color sensation data storage step for storing the color sensation data output in the color sensation analysis step in a color sensation data storage unit;
An image data storage step of storing image data corresponding to the color sensation data in an image data storage unit;
Reordering the color sensation data according to the frequency of correspondence with all image data stored in the image data storage unit, and creating a color sensation data name list creating step for creating a name list of the color sensation data;
Performing a color sensation data name list accumulation step of accumulating a name list of the color sensation data,
In the image search device,
A first display step of extracting a name list of the color sensation data from the color sensation data name list storage unit and displaying the color sensation language data corresponding to the color sensation data based on the name list;
Reordering step of extracting image data from the image data storage unit, and rearranging the color sensation data corresponding to the image data for each image data in descending order of the number of color sensation points;
A second display step of displaying the color sensation language data corresponding to the color sensation data rearranged in the rearrangement step;
An image stored in the image data storage unit using the color sensation language data selected by the user from the color sensation language data displayed in either the first display step or the second display step. A program for an image storage / retrieval system, characterized by executing a retrieval step for retrieving data.