JP2010061408A - Image processing program and image processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing program and an image processing system for avoiding the erroneous recognition of an image when using classification based on a color space. <P>SOLUTION: The image processing program of electronic equipment for extracting the features of a photographic image, and for processing the image is installed in a computer of the electronic equipment. The computer is made to execute a procedure for detecting the color data of pixels configuring a photographic image; a procedure for separating a value corresponding to brightness among the color data from a value corresponding to chrominance, and for plotting the color data in the two-dimensional coordinate space of only the value corresponding to the chrominance; and a procedure for classifying the colors of the photographic image on the basis of the coordinate position of the color data. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image processing program and an image processing system that extract and process features of a captured image.

  This photographed image is used for providing various services. Specifically, with the recent spread of mobile phones and digital cameras with built-in camera functions, captured images can only be recorded on people, landscapes, etc., or on content that has been recorded on paper like timetables so far. In addition, the range of use is expanding, such as enabling connection to the Internet.

  This can be achieved by an image processing technique, that is, a technique for recognizing a captured image, extracting a feature amount of the image, and processing the image. When the parameter of the recognized captured image is determined, the parameter is compared with the parameter of the master data stored in the database, and the master data that substantially matches the captured image can be retrieved from the database. .

  Here, if a large amount of master data is registered in this database, the time required for search takes a long time, and the correct answer rate also decreases. This is because, if the search parameter increases, for example, there is an increased possibility of searching for master data that happens to be close to the parameters of the photographed image even though it has completely different characteristics from the photographed image.

  In this case, even if unnecessary information such as the background is deleted from the photographed image and then the feature amount is extracted, the time required for the search is greatly reduced as long as a large amount of master data is registered in the database. It is not possible. On the other hand, when registering master data in the database, for example, if the data is narrowed down using the classification based on the color space (Lab color system, etc.) described in Patent Documents 1 and 2, the amount of master data to be registered can be reduced. become.

Japanese Patent Laid-Open No. 08-251397 JP 2002-271644 A

By the way, the Lab color system described above is expressed by lightness (L value) and chromaticity (ab value), and this lightness causes a misrecognition of an image. There is a problem that it becomes prominent in an image taken in
This is because the photographed image is easily affected by ambient light, and the camera function can be adjusted by white balance (WB). However, it is not easy to set this WB appropriately. This is because when the inferior camera function is used, the brightness of the photographed image is too high and the color cannot be accurately classified.

In other words, even if you try to narrow down the amount of master data to be registered using classification by color space, this brightness will not be able to narrow down because of the hindrance to classification. As a result, the time required for the search becomes long.
SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an image processing program and an image processing system capable of solving the above-described problems and avoiding misrecognition of images when using classification by color space.

A first invention for achieving the above object is an image processing program for an electronic device that extracts a feature of a captured image and processes the image.
This program is provided in the computer of the electronic device. The program detects the color data of the pixels constituting the captured image, and the value corresponding to the lightness among the color data is the value corresponding to the chromaticity. And a procedure for plotting the color data in a two-dimensional coordinate space having only a value corresponding to the chromaticity and a procedure for classifying the color of the photographed image based on the coordinate position of the color data.

According to the first aspect of the invention, the image processing program specifies a procedure performed by the computer in extracting the characteristics of the captured image and processing the image.
Specifically, the color data of the pixels constituting the photographed image is detected, and the value corresponding to lightness in the color data is separated from the value corresponding to chromaticity, and the color data is only 2 corresponding to the value corresponding to chromaticity. Plot in the dimensional coordinate space.

Here, this lightness causes a misrecognition of the image. In the present invention, a value corresponding to this lightness and a value corresponding to chromaticity are separated, and the value corresponding to this lightness is the color of the photographed image. Therefore, erroneous recognition can be avoided more reliably than in the past.
In addition, if the color data is classified only by the value corresponding to the chromaticity, the search parameter can be reduced compared to the case where the value corresponding to the lightness is included. In addition, the load on the searched database can be reduced.

  According to a second aspect of the present invention, in the configuration of the first aspect, the procedure for classifying the color of the photographed image is the distance from the origin of the two-dimensional coordinate space, the elevation angle, and the space around the origin. A procedure for identifying each color group belonging to the divided range, a procedure for setting the order in order from the group having the largest color data, and certifying a predetermined upper group, and each color data plotted for each upper group And a procedure for determining an area from a predetermined lower limit value to an upper limit value sandwiching the calculated average elevation angle as a color search target when viewed in the elevation angle direction.

  According to the second invention, in addition to the operation of the first invention, the color data of the pixel is plotted in a two-dimensional coordinate space, which includes a distance from the origin, an elevation angle, and a group of colors to which the pixel belongs. Each is identified. After identifying the upper group having the most color data, in addition to this upper group, the region from the predetermined lower limit value to the upper limit value sandwiching the average elevation angle of the upper group is also determined as the search target. While reducing the number of parameters, it is possible to secure colors that may be necessary, which contributes to improving the retrieval accuracy of captured images.

  According to a third invention, in the configuration of the second invention, following the procedure of specifying the distance, the elevation angle, and the color group to which each belongs, a procedure for calculating an average distance from the origin for each plotted color data for each group; When the calculated average distance is shorter than the radius of the unidentifiable circle of the color centered on the origin, a procedure for determining a color belonging to the inside of the circle as a search target is executed.

  According to the third invention, in addition to the operation of the second invention, the average distance from the origin is calculated for each group, and when this average distance is shorter than the radius of the unspecified color circle, Even if it does not correspond to the upper group, the color belonging to this circle is also determined as the search target. Therefore, in this case as well, this is a guarantee that the value corresponding to the above-described brightness is not taken into consideration, and contributes to improvement in the retrieval accuracy of the captured image.

According to a fourth invention, in the configurations of the first to third inventions, the Lab color system is used for the color data, the L value of the color data is separated from the ab value, and each color data is arranged on the ab plane. It is characterized by only being plotted.
According to the fourth invention, in addition to the operations of the first to third inventions, the Lab color system is the most commonly used color system in all fields at present, and the Lab color system is used for the color data of pixels. Using a color system makes it possible to use a wide range of programs.

5th and 6th invention is an image processing system which extracts the characteristic of the picked-up image image | photographed with the portable apparatus, and processes an image.
The image processing apparatus includes a preprocessing unit that classifies the colors of the captured image. The preprocessing unit detects the color data of the pixels constituting the captured image, and sets a value corresponding to lightness in the color data to the chromaticity. While separating from corresponding values, the color data is plotted in a two-dimensional coordinate space of only chromaticity, and the color of the photographed image is classified based on the coordinate position of the color data. In the fifth invention, the classification result is stored in the database.

  On the other hand, the sixth invention further extracts a feature amount of the photographed image, and also includes a feature extraction unit that determines the parameter of the photographed image in consideration of the color information classified by the above-described preprocessing unit, and the determined parameter Is compared with the parameters of the master data stored in the database, and the recognition / determination unit that searches the database for master data that substantially matches the captured image, and returns the URL to which the searched master data belongs to the portable device from the database. And an output unit.

  According to the fifth and sixth inventions, in the image processing system, the value corresponding to the lightness that causes a misrecognition of the image among the color data of the pixels constituting the captured image by the preprocessing unit corresponds to the chromaticity. Since this color data is plotted in a two-dimensional coordinate space having only values corresponding to chromaticity, it is possible to avoid erroneous recognition more reliably than in the past.

  Further, if this pre-processing unit classifies the color data only by the value corresponding to the chromaticity, the number of parameters for the search can be reduced, and the system user can obtain master data that substantially matches the captured image as compared with the conventional case. You can search quickly from the database. Furthermore, from the viewpoint of the system administrator, the number of searches (including erroneous searches) is reduced, so the load on the database is reduced compared to the conventional case, or the free space of the database is increased, so that more master data can be stored. Can be registered in the database.

  According to the present invention, when using classification by color space, a value corresponding to lightness is not considered when classifying the color of a photographed image, and thus an image processing program and an image processing system that can reliably avoid erroneous recognition of an image are provided. Can be provided.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described with reference to the drawings.
FIG. 1 is an overall configuration diagram of an image processing system according to the present embodiment. The system 1 is configured using a network 6, and the network 6 includes a base station 4 and a search server (including a database 10). Electronic devices) 8 are connected to each other.

  The server 8 has a function of extracting a feature of a photographed image photographed by a portable device (for example, a portable phone) 2 according to a program command and processing the image. Although the portable device 2 of the present embodiment is described as having a built-in camera function, the portable device 2 is a digital camera or the like in addition to a personal digital assistant (PDA) and a notebook computer as long as the subject can be photographed. May be.

  Various operations of the server 8 of this embodiment are controlled by the controller 9 (FIG. 2). The controller 9 is an element that functions as a computer, and has hardware resources such as a CPU and a memory that stores a program. Then, the controller 9 executes a predetermined program using this hardware resource.

  More specifically, the controller 9 includes an input unit 12, a preprocessing unit 14, a feature extraction unit 16, a recognition / determination unit 18, and an output unit 20, as shown in FIG. The input unit 12 inputs a photographed image from the mobile device 2, and the preprocessing unit 14 deletes the background of the photographed image and identifies the color of the photographed image described later. A predetermined process before extracting the feature is executed and output to the feature extracting unit 16 so as to be converted into a converted image.

The feature extraction unit 16 of the present embodiment defines the image feature amount extraction range for the binarized image, and subsequently determines the parameters of the captured image for the extraction range in consideration of the color information described above. The result is output to the recognition / determination unit 18.
The recognition / judgment unit 18 searches the database 10 for master data that substantially matches the captured image and outputs it to the output unit 20, which outputs the URL (Universal) to which the retrieved master data belongs.
(Resource Locator) is returned to the portable device 2. As a result, the mobile device 2 browses the homepage related to the photographed image. Specifically, when the URL indicates a video, the mobile device 2 indicates the video, when the URL indicates music, the flash, and the flash. When pointing to an animation, the animation can be played back.

More specifically, FIG. 3 shows processing that the program causes the controller 9 to execute. Hereinafter, the operation according to the present invention of the image processing system 1 configured as described above will be described.
First, in step S301 in the figure, the portable device 2 captures a logo image (subject) 72 with its camera function (FIG. 4). The logo image 72 of the present embodiment is composed of a combination of characters related to beverages, graphics of contents, and colors, and is printed on the printed material 70 in color as an advertisement.

  Here, the portable device 2 displays the logo image 72 on the screen and takes a picture. The photographed image 74 of the present embodiment is photographed in a state where the logo image 72 is not enlarged or reduced, but rotated to the left. On the other hand, the mobile device 2 transmits the captured image 74 toward the base station 4 and transmits it to the search server 8 via the network 6. The URL of the server 8 is described in advance in the program. Alternatively, it may be described in advance in the printed material 70 or embedded in the printed material 70 using a digital watermark.

In step S302, the input unit 12 receives the captured image 74 and stores it in the memory, and the process proceeds to step S303.
In this step S303, the preprocessing unit 14 deletes the background of the captured image 74 and displays it while leaving the periphery. In addition, the preprocessing unit 14 identifies the color of the photographed image 74 partitioned in a dot matrix, and then converts the photographed image 74 into a white and black binary image based on the reference value.

Specifically, the pre-processing unit 14 of the present embodiment has a color identification function in addition to a background deletion function and a binarization conversion function. As shown in FIG. The image feature amount calculation unit 40 and the discrimination processing unit 60 are provided.
First, the Lab conversion grouping unit 30 includes a pixel Lab conversion unit 32, an ab data calculation unit 34, and a result return unit 36.

The Lab conversion unit 32 converts the RGB values into Lab values when the color data of valid pixels constituting the captured image 74 is detected as RGB values of, for example, red, green, and blue. Is output to the ab data calculator 34.
The ab data calculation unit 34 separates the L value representing lightness from the converted Lab values from the ab values representing chromaticity, that is, hue and saturation.

More specifically, the ab data calculation unit 34 plots the color data of each pixel on the ab plane 80 which is a two-dimensional coordinate space (FIG. 6), and specifies the a value and the b value.
The ab plane 80 includes an a-axis representing red and green hues and a b-axis representing yellow and blue hues. The positive direction of the a-axis indicates red saturation, and the negative of the a-axis. Direction indicates green saturation. On the other hand, the positive direction of the b-axis indicates yellow saturation, and the negative direction of the b-axis indicates blue saturation.

Here, the ab data calculation unit 34 partitions the ab plane 80 into m groups when plotting the color data of each pixel on the ab plane 80. In this embodiment, as shown in FIG. 6, groups of colors I to VIII divided into eight equal parts around the origin of the ab plane 80 are provided.
As a result, each pixel plotted on the ab plane 80 is identified in addition to the distance from the origin and the angle (elevation angle) from the a axis. This distance is obtained from each value of the a-axis and b-axis using the three-square theorem, and this angle is obtained from an inverse trigonometric function related to sine and cosine, and is displayed as a value of 0 ° to 360 °. The

  In this embodiment, it is assumed that the logo image 72 is composed of a light orange ground color, a bright orange character, a dull red content, a slightly bright orange color, and a bright yellow color. If a detected pixel of the captured image 74 has an orange hue, this pixel is plotted in the first quadrant of the ab plane 80 (indicated by a cross in FIG. 7). The distance s, the elevation angle t, and the group II to which it belongs are specified.

Similarly, all the pixels of the captured image 74 are plotted on the ab plane 80 (indicated by x in FIG. 8), and the result is output from the result return unit 36 to the image feature amount calculation unit 40.
Returning again to FIG. 5, the image feature amount calculation unit 40 includes an ab data storage unit 42, a group classification unit 44, an upper group authorization unit 46, a distance / angle calculation unit 48, and a result return unit 50.

The ab data storage unit 42 stores the results from the Lab conversion grouping unit 30, and the group classification unit 44 rearranges the groups I to VIII in descending order of the number of stored pixels.
In the example of the logo image 72 described above, the groups I to VIII are assigned in the order of many X marks shown in FIG. 8, for example, group II, group III, group I, group V, group VIII, group IV, group VI. , Rearrange to group VII.

First, the upper group recognition unit 46 recognizes the group storing the largest number of pixels, that is, the group II in the above example, first, that is, the first color.
Next, the higher-level group recognition unit 46 places the second to eighth ranks in the descending order of the number of stored pixels for the other groups except for the group II, and recognizes the second and third colors.

  However, in the present embodiment, the second color is a group having 70% or more (for example, 2800 or more) of the number of stored pixels (for example, 4000) of the group corresponding to the first color. Similarly, it is a condition that the third color is a group having 70% or more (for example, 1960 or more) of the number of stored pixels (for example, 2800) of the group corresponding to the second color. This is because a group with a small number of storage is not recognized as a higher rank.

  Subsequently, the distance / angle calculation unit 48 first calculates the distances of all the pixels stored in the same group, and calculates the average distance of the group. Further, the angles of all the pixels stored in the same group are obtained, and the average angle (average elevation angle) of the group is calculated. These average distances and average angles are calculated for all groups, and the results are output to the result return unit 50.

The result returning unit 50 outputs the specified group name, average distance, and average angle to the database 10 when registering the color data of the captured image 74 in the database 10. Thereby, this color data is stored in the database 10.
On the other hand, when the result returning unit 50 searches the database 10 for master data that substantially matches the captured image 74, the result returning unit 50 outputs the identified group name, average distance, and average angle to the discrimination processing unit 60.

As illustrated in FIG. 5, the determination processing unit 60 includes an angle comparison determination unit 62, a distance comparison determination unit 64, and a result return unit 66.
The angle comparison / determination unit 62 determines the groups adjacent to each other across the groups corresponding to the first to third colors as the search targets of the database 10. Specifically, for the group corresponding to the first color, an upper limit value and a lower limit value are obtained by adding or subtracting a predetermined angle to the average angle of this group, and are included in the elevation angle direction from the upper limit value to the lower limit value. A part of the group is determined as a color search target.

In the above example, as shown in FIG. 9, if the average distance s1 and the average angle t1 of the group II corresponding to the first color are set, the upper limit value is t1 + u1 and the lower limit value is t1-u1, and these upper limit values The region 91 from the lower limit value (t1−u1) to the upper limit value (t1 + u1) when viewed in the elevation angle direction as in the triangle formed in this figure is not the entire group III belonging to or the entire group I belonging to the lower limit value. It becomes a search object.
Note that all the groups I and III corresponding to both sides of the group II may be determined as color search targets. However, as described above, the region 91 from the upper limit value to the lower limit value is selected as the color search target. If determined, the search accuracy is improved as compared with the case where the color search target is determined in units of groups. This is because the range apart from the average angle t1 among these groups I and III need not be included in the search target.

  Further, if the group corresponding to the second color is group I, the upper limit value is t2 + u1 and the lower limit value is t2-u1 when the average distance s2 and average angle t2 of this group I are also shown in FIG. As shown in a region 92 indicated by a triangle at, a part of the group VIII to a part of the group II are searched for colors when viewed in the elevation direction.

Further, for the group corresponding to the third color, similarly, a range is determined using an upper limit value and a lower limit value within a range of 360 °, and all colors included in this range are determined as search targets.
Subsequently, the distance comparison / determination unit 64 determines all colors near the origin on the ab plane as search targets according to the obtained average distance.

Specifically, in the above example, the average distance s2 of the group I is located in the vicinity of the origin as indicated by a cross in FIG.
Here, since the group I includes many truly dull red pixels, the average distance is located in the vicinity of the origin, or the L value described above is not considered, It is unclear whether the average distance is located near the origin. On the other hand, it is not preferable to exclude this unknown case from the search target.

  Therefore, in this embodiment, an unspecified circle of color centered at the origin is registered in the database 10 as a criterion for determining an unclear region of color, and the distance comparison / discriminating unit 64 determines that the average distance s2 is When the radius is shorter than the radius of the circle, in addition to the above-mentioned group VIII to group II (or the region indicated by a triangle in FIG. 10), a region 92 including the circle with the radius centered on the origin is included. The colors of other groups belonging to the inside of the circle are also determined as search targets.

Next, the result return unit 66 outputs the color information thus identified to the feature extraction unit 16, and proceeds to step S304 in FIG.
In step S304, the feature extraction unit 16 extracts the features of the above-described binarized image, and adds the identified color information to determine the parameters of the captured image 74.

  In step S 305, the recognition / determination unit 18 compares the parameter determined for the captured image 74 with the master data parameter stored in the database 10, and has a value closest to the feature amount of the captured image 74. The master data is retrieved from the database 10. Thereafter, when the logo mark can be acquired from the determined parameters, the process proceeds to step S306, and the output unit 20 returns the URL to which the logo mark belongs to the portable device 2 and proceeds to step S307 to exit the series of routines. As a result, the mobile device 2 can browse the detailed information of the advertisement from this URL.

As described above, according to the present embodiment, the image processing program specifies the procedure performed by the controller 9 of the server 8 in extracting the features of the captured image 74 and processing the image.
Specifically, the RGB values of the dot matrix pixels constituting the captured image 74 are detected, the RGB values are converted into Lab values, and the L values are separated from the ab values. Plotting.

Here, in this embodiment, since the L value is not considered when classifying the color of the captured image 74, erroneous recognition can be reliably avoided as compared with the conventional case.
Further, if the pixel color data is classified only by the ab value, the search parameter can be smaller than the case where the L parameter is included. Therefore, master data that substantially matches the captured image 74 as compared with the conventional case is stored in the database 10. Thus, it is possible to search quickly, and the load on the searched database 10 can be reduced. Alternatively, more master data can be registered in the database 10.

  Further, the pixel color data is plotted on the ab plane, which specifies the distance from the origin, the elevation angle, and the group of colors to which it belongs. Then, after identifying the upper group having the most color data, areas 91 and 92 from the predetermined lower limit value to the upper limit value sandwiching the average angle of the upper group are determined as search targets in addition to the upper group. Thus, while reducing the number of parameters for search, it is possible to secure colors that may be necessary, which contributes to improvement in search accuracy of master data that substantially matches the captured image 74.

  On the other hand, the average distance from the origin is calculated for all groups, and if this average distance is shorter than the radius of the unspecified color circle, even if it does not fall under the upper group, the color belonging to this circle Is also determined as a search target. Therefore, in this case as well, this is a guarantee that the above-described L value is not taken into consideration, and contributes to improvement in the search accuracy of the master data.

Furthermore, the Lab color system described above is the most commonly used color system in all fields, and if the Lab color system is used for pixel color data, a wide range of programs can be used.
Further, according to the image processing system 1 using this program, the system user can quickly search the master data for the master data that substantially matches the photographed image 74 as compared with the conventional case, which gives stress to the user. Absent. Further, from the viewpoint of the system administrator, since the number of searches (including erroneous searches) is reduced, the load on the database 10 is reduced as compared with the conventional system, or the free space of the database 10 is increased, so that more masters are available. Data can be registered in the database 10.

The present invention is not limited to the above embodiments, and various modifications can be made without departing from the scope of the claims. For example, a part of the configurations of the above embodiments can be omitted or arbitrarily combined so as to be different from the above.
Further, the program may be provided in the portable device 2 described above in addition to the controller 9 of the search server 8.

Furthermore, although the said Example was demonstrated in the example embodied in the system which searches a logo mark, it is not necessarily limited to this form, Various systems which classify the color of a picked-up image numerically, Applicable to programs.
Furthermore, in the above embodiment, the detected RGB value is converted into a Lab value, but the color of the pixel may be directly detected by the Lab value.

  Further, the present invention is not limited to this Lab color system. That is, as long as the color system is capable of completely separating lightness and luminance with values different from chromaticity and can plot this chromaticity in a two-dimensional coordinate space, the present invention is not limited to the XYZ color system (Y Is lightness, Z indicating the degree of blue, and X indicating other elements correspond to chromaticity), xyY color system (Y is lightness, x and y indicating chromaticity coordinates are chromaticity) In addition to the Luv color system, the present invention is also applicable to known color systems such as the HSV color system and the YUV color system.

  In any of these cases, similarly to the above, there is an effect that misrecognition of an image can be avoided when using classification by color space.

It is a figure which shows the image processing system of a present Example. It is a schematic block diagram of the search server of FIG. It is a flowchart of the logo mark search by the system of FIG. It is explanatory drawing of the imaging | photography process of FIG. FIG. 4 is a control block diagram of the pretreatment process of FIG. 3. It is explanatory drawing of ab plane coordinate by the pre-processing part of FIG. It is the figure which plotted the color data of 1 pixel on ab plane coordinate of FIG. It is the figure which plotted several color data on ab plane coordinate of FIG. It is explanatory drawing of the search range in ab plane coordinate of FIG. It is explanatory drawing of the search range in ab plane coordinate of FIG.

Explanation of symbols

1 Image processing system 2 Mobile device (electronic device)
8 Search server (electronic equipment)
9 Controller (computer)
DESCRIPTION OF SYMBOLS 10 Database 14 Pre-processing part 16 Feature extraction part 18 Recognition / judgment part 20 Output part 74 Photographed image 80 Ab plane 91,92 area | region

Claims (6)

An image processing program for an electronic device that extracts a feature of a captured image and processes the image,
The program is provided in a computer of the electronic device,
A procedure for detecting color data of pixels constituting the captured image;
Separating a value corresponding to lightness from the value corresponding to chromaticity in the color data, and plotting the color data in a two-dimensional coordinate space of only a value corresponding to the chromaticity;
An image processing program for executing a procedure for classifying the color of the photographed image based on the coordinate position of the color data. An image processing program according to claim 1,
The procedure for classifying the color of the captured image is as follows:
For the plotted color data, a procedure for identifying a distance from the origin of the two-dimensional coordinate space, an elevation angle, and a group of colors belonging to a range obtained by equally dividing the space around the origin;
A procedure for setting a rank in order from the group having the most color data and certifying a predetermined upper group;
For each upper group, a procedure for calculating an average elevation angle for each plotted color data;
An image processing program comprising: a procedure for determining an area from a predetermined lower limit value to an upper limit value sandwiching the calculated average elevation angle as a color search target when viewed in the elevation angle direction. An image processing program according to claim 2,
Following the steps of identifying the distance, elevation angle, and color group to which they belong,
For each group, a procedure for calculating an average distance from the origin for each plotted color data;
And executing a procedure for determining a color belonging to the inside of the circle as a search target when the calculated average distance is shorter than a radius of an unspecified circle of color centered on the origin. program. An image processing program according to any one of claims 1 to 3,
Lab color system is used for the color data,
An image processing program characterized in that an L value is separated from an ab value in the color data, and each color data is plotted only on an ab plane. An image processing system that extracts features of a captured image captured by a mobile device and processes the image,
A pre-processing unit for classifying the color of the captured image;
The preprocessing unit
Detects color data of pixels constituting the photographed image, separates a value corresponding to lightness from the color data from a value corresponding to chromaticity, and plots the color data in a two-dimensional coordinate space of only chromaticity An image processing system, wherein the color of the captured image is classified based on the coordinate position of the color data, and the result is stored in a database. An image processing system that extracts features of a captured image captured by a mobile device and processes the image,
A pre-processing unit for classifying the color of the captured image;
A feature extraction unit that extracts an image feature amount of the photographed image and determines parameters of the photographed image in consideration of the classified color information;
A recognition / determination unit that compares the determined parameter with a parameter of master data stored in a database, and searches the database for master data that substantially matches the captured image;
An output unit that returns a URL to which the retrieved master data belongs from the database to the portable device;
The pre-processing unit described above
Detects color data of pixels constituting the photographed image, separates a value corresponding to lightness from the color data from a value corresponding to chromaticity, and plots the color data in a two-dimensional coordinate space of only chromaticity An image processing system, wherein the color of the captured image is classified based on the coordinate position of the color data.