JPH06243179A - Method and device for picture retrieval - Google Patents

Abstract

PURPOSE:To increase the picture retrieval speed and to surely perform retrieval by extracting information indicating the complexity of a picture from a picture storage means together with the pertinent picture and controlling the visible display time of the pertinent picture based on information indicating the complexity of the picture. CONSTITUTION:When plural pictures are stored in a picture storage means 7, information indicating the complexities of pictures are stored there also. When pertinent pictures are successively extracted based on a prescribed retrieval condition and are visibly displayed on a display means 16 to retrieve a picture, information indicating the complexities of pictures are extracted together with pertinent pictures, and the visible display times of pertinent pictures are controlled based on these information. Thus, it is unnecessary to set the display time in comparison with the case that an operator sets the display time of each picture, and an optimum display time adapted to the information volume can be set for each picture. As the result, the picture retrieval speed is increased and the retrieval is surely performed as the whole.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image retrieval method and an apparatus therefor, and more specifically, it stores a plurality of images in an image storage means, and based on a predetermined retrieval condition,
The present invention relates to an image search method and apparatus for sequentially extracting corresponding images from image storage means and visually displaying them on a display means.

[0002]

2. Description of the Related Art With the recent development of image processing technology and the widespread use of large-capacity storage media represented by optical discs, research on image database creation and image search using the created image database is under way. . As a method of image retrieval using the image database, a method has been proposed in which accumulated images are sequentially read out and sequentially sequentially displayed as if turning pages of a book. Of these image retrieval methods, the simplest method is to uniformly set the display time for each image. It is also conceivable to apply the method (see Japanese Patent Laid-Open No. 3-14178) that allows the operator to set the display time, which is applied in the document search method, to the image search method.

If the former method is adopted, the setting work of the display time is easy, and the images can be sequentially and visually displayed like turning pages of a book. Further, if the latter method is adopted, the work of setting the display time becomes complicated, but since the display time can be set according to the amount of image information, the image search can be performed easily.

[0004]

If the former method is adopted, all the images are displayed for a predetermined time regardless of the amount of information contained in the images, so the amount of information is extremely small. There is an inconvenience that the display time is too long for an image and the display time is short for an image having a remarkably large amount of information, which makes it difficult to visually grasp information contained in the image.

If the latter method is adopted, the former method can be solved to some extent. However, since it is applied to a pre-classified image database and the display time is set for each image type, it is almost impossible to apply it to an image database in which random images are stored. is there. That is, since the information amount of each image stored at random has no regularity, the operator has to determine the amount of information for each image and set the display time, which is stored. If the number of the displayed images increases, it takes a significantly long time to set the display time as a whole, and it is practically impossible to set such a display time. In addition, since the operator determines the amount of information, there is an individual difference, and even if the display time is optimal for one person, the display time is inappropriate for another person. There is also.

As is clear from the above, it is possible to eliminate the need for the operator to set the display time, and to speed up and ensure the image retrieval (the desired image can be retrieved reliably).
There was no image retrieval method that can achieve both.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and regardless of whether or not images are randomly stored, the display time for each image is set in correspondence with the information amount of each image. It is an object of the present invention to provide a novel image search method and apparatus that can be automatically set and can achieve both high-speed image search and reliable search.

[0008]

In order to achieve the above-mentioned object, the image retrieval method of claim 1 stores each image in the image storage means, and also stores information indicating the complexity of the image. A method of storing the image in the storage means, extracting information indicating the complexity of the image together with the corresponding image from the image storage means, and controlling the visual display time of the corresponding image based on the information indicating the complexity of the image. is there.

The image search method of claim 2 is a method of adopting the occupancy rate of the boundary line in the image as the information indicating the complexity. The image search method of claim 3 is a method of adopting, as the information indicating the complexity, the number of regions defined by a plurality of representative colors in the image. The image search device according to claim 4 employs, as the image storage means, one that stores information indicating the complexity of the image together with the stored image,
Moreover, an extraction unit that extracts information indicating the complexity of the image together with the corresponding image from the image storage unit, and a display time control unit that controls the visible display time of the corresponding image based on the information indicating the complexity of the image. Is included.

[0010]

According to the image retrieval method of claim 1, when storing a plurality of images in the image storing means, information indicating the complexity of the images is also stored in the image storing means, and the predetermined retrieval condition is satisfied. Based on the above, when performing an image search by sequentially extracting the corresponding images from the image storage means and visually displaying them on the display means, the information indicating the complexity of the images is extracted from the image storage means together with the corresponding images. However, since the visual display time of the corresponding image is controlled based on the information indicating the complexity of the image, unlike the case where the operator sets the display time of each image, the work of setting the display time can be eliminated. At the same time, it is possible to set the optimum display time for each image in accordance with the amount of information. As a result, as a whole, both high-speed image search and reliable image search can be achieved.

According to the image retrieval method of claim 2, since the occupancy ratio of the boundary line in the image is adopted as the information indicating the complexity, each image is stored in the image storage means. By detecting the boundary line and calculating the number of pixels of the boundary line, information indicating the complexity can be automatically obtained. In addition, the occupancy of the boundary obtained in this way is a value that is objectively determined for each image, and there is no room for arbitrary intervention, so speed up image search and ensure image search. Can be compatible.

According to the image retrieval method of the third aspect, since the number of areas defined by a plurality of representative colors in the image is used as the information indicating the complexity, each image is stored in the image storage means. When storing, information indicating complexity can be automatically obtained by performing area division and calculation of the number of divided areas. In addition, the number of regions thus obtained is a value that is objectively determined for each image, and there is no room for arbitrary intervention, so both high-speed image search and reliable image search can be achieved. .

According to another aspect of the image retrieval apparatus of the present invention, a plurality of images are stored in the image storage means, the corresponding images are sequentially extracted from the image storage means based on a predetermined retrieval condition, and are displayed by the display means. When an image search is performed by displaying the image on the screen, the image storage means that stores information indicating the complexity of the image together with the stored image is adopted, and the extraction means uses the image together with the corresponding image from the image storage means. The information indicating the complexity of the image is extracted, and the display time control means controls the visible display time of the corresponding image based on the information indicating the complexity of the image. Therefore, the operator sets the display time of each image. Unlike the above case, the work of setting the display time can be eliminated, and the optimum display time can be set for each image according to the amount of information. As a result, as a whole, both high-speed image search and reliable image search can be achieved.

[0014]

Embodiments will be described in detail below with reference to the accompanying drawings showing embodiments. FIG. 1 is a flow chart for explaining a process of registering an image in an image database in one embodiment of the image search method of the present invention, and FIG. 2 is based on the image database obtained by the process of the flow chart of FIG. It is a flow chart explaining processing which searches for an image.

First, the flow chart of FIG. 1 will be described.
In step SP1, color data (for example, RGB data) of each pixel forming the original image is converted into chromaticity data (for example, L * a * b * data), and in step SP2, smoothing is performed using a median filter or the like. The borderline is detected from the smoothed image in step SP3, the occupancy ratio of all the detected borderlines to all pixels is calculated in step SP4, and step S4 is performed.
At P5, the boundary occupancy rate is stored in the recording medium such as an optical disk together with the original image and the keyword information. still,
The above series of processing is performed for each original image.

Further details will be described. Above step SP
In the conversion process in 1, for example, the tristimulus values of the target color stimulus are X, Y, and Z, and the tristimulus values of the illumination light source are X ₀ , Y ₀ ,
Z ₀ (a unique value determined for each light source) can be achieved by performing the following arithmetic processing. That is, X = 0.607 x R + 0.174 x G + 0.201 x B Y = 0.299 x R + 0.587 x G + 0.114 x B Z = 0.066 x G + 1.117 x B obtained tristimulus values X of the stimulus, Y, and Z, then, L * = 116 (Y / Y 0) 1/3 -16 a * = 500 [(X / X 0) 1/3 - (Y / Y 0 ^{) 1/3] b * = 200 [} (X / X 0) 1/3 - (Z / Z 0) 1/3] by performing calculation of L *, a *, b * values can be obtained.

The smoothing process in step SP2 is performed by, for example, the one-dimensional observation window shown in (A) in FIG. 3 or (B).
This can be achieved by using a two-dimensional observation window indicated by and performing the process of replacing the density value of the central pixel with the median value of the density of the pixels belonging to the observation window for each pixel. Step SP3 above
The boundary line detection process of 1 can be achieved by, for example, performing a differential process and a threshold process to obtain a boundary region, and then thinning the boundary region. More specifically, in the process of obtaining the boundary area, the calculation window shown in FIG.
The color difference differential value D (x, y) of (x, y) is calculated.

[0018]

[Equation 1]

However, D _i = {(L _i * -L *) ² + (a _i *
-A *) ² + (bi * -b *) ² } ^1/2 , (L * a * b
*), L * a * b * values of f (x, y), ( L n * a n * b
_n *) is the L * a * b * value at i = n. Then, the magnitude of the color difference differential value D (x, y) obtained by the calculation of Equation 1 and the predetermined threshold value THR is discriminated, and D (x, y) ≧ THR
In the case of, the differential image g (x, y) = 1 is set, and D (x,
By setting differential image g (x, y) = 0 when y) <THR, a differential image g (x, y) in which only the boundary portion is "1" and other portions are "0" is obtained. be able to.

The processing for thinning the boundary area is as follows, for example. As shown in FIG. 5A, the central pixel is represented by p0, and the eight pixels adjacent to the central pixel are represented by Equation 2.

[0021]

[Equation 2]

Then, as a first step, the central pixel p
When 0 satisfies all the conditions shown in Expression 3, a flag is set, and after processing for all boundary pixels,
The value B (p) of the pixel for which the flag is set is set to "0". That is, the value of the right and lower boundary pixels is set to "0" by this processing.

[0023]

[Equation 3]

However, S (p ₀ ) is the value B of the neighboring pixels of p ₀
It is the number of times the pixel value changes from “0” to “1” when (p ₁ ), B (p ₂ ), ..., B (p ₈ ) are examined in this order. Specifically, as shown by (B) in FIG.
When B (p ₂ ) = B (p ₅ ) = B (p ₆ ) = 1, N (p ₀ ) = 3 and S (p ₀ ) = 2. Next, as a second step, a flag is set when the central pixel p0 satisfies all the conditions shown in Formula 4, and after processing for all boundary pixels, the value B of the flagged pixel is set. (P) is set to "0". That is, by this processing, the values of the left and upper boundary pixels are set to "0".

[0025]

[Equation 4]

Then, the processing of the first step and the processing of the second step are repeated until the line width becomes "1". The occupation rate calculation process of step SP4 is performed by, for example, the differential image g after the thinning process obtained in step SP3.
This can be achieved by obtaining the total number of pixels whose value is “1” in (x, y) and dividing by the total number of pixels of the differential image g (x, y).

The occupation rate obtained as described above is stored in the keyword management table together with the keyword information of the corresponding image and the pointer to the image data, for example. As is clear from the above description, by simply specifying or reading an image to be registered, conversion processing, smoothing processing, boundary line detection processing, and boundary line occupation rate calculation processing are performed automatically, Since it can be stored in the keyword management table or the like together with the information and the like, it can be stored with an objective occupation ratio obtained without being affected by the operator's will. Here, if the occupancy rate of the boundary line is high, it means that there are many objects that are divided by the boundary line in the image,
On the other hand, if the occupancy rate is low, there will be few objects that are divided by the boundary line, so that the occupancy rate can express the complexity of an image with high accuracy.

Next, the flowchart of FIG. 2 will be described. If the image search by the page turning display is instructed, in step SP1, a list of corresponding display target images is obtained based on the keyword set and input by the operator,
In step SP2, the image data is read according to the list, the occupation ratio of the boundary line is obtained, and the display time of the corresponding image is calculated based on the occupation ratios of all the boundary lines in the list in step SP3. At, the corresponding image is visually displayed for the calculated time. Then, in step SP5, it is determined whether or not the end of the page turning display is instructed. If the end is not instructed, the process of step SP4 is performed again. On the contrary, if the end of page turning display is instructed,
A series of processing is finished as it is.

In the display time calculation processing in step SP3, for example, the display time for an image whose boundary occupancy rate is "0" is set as a reference time, and the boundary line occupancy rate is set with respect to this reference time. It can be achieved by multiplying by a constant determined based on Here, as the relationship between the occupancy rate of the boundary line and the constant, a linear function, a quadratic function or the like can be exemplified. Further, the reference time may be set as a value that cannot be changed, but it is preferable to set it as a value that can be changed in consideration of individual differences among searchers.

The determination of the page-turning display end instruction in step SP5 may be made based on, for example, whether or not the searcher has selected the image that is being visually displayed as the search target image. As is clear from the above description, each image extracted from the image database based on the search condition is displayed only for the time determined based on the occupancy ratio of the boundary line. Therefore, the required time can be shortened as a whole, and the image search can be surely achieved by not overlooking the image to be searched.

[0031]

[Embodiment 2] FIG. 6 is a flow chart for explaining the processing of registering an image in the image database in another embodiment of the image retrieval method of the present invention, and FIG. 7 is obtained by the processing of the flow chart of FIG. It is a flow chart explaining the processing which searches for an image based on the image database.

First, the flowchart of FIG. 6 will be described. In step SP1, color data (for example, RGB data) of each pixel forming the original image is converted into chromaticity data (for example, L * a * b * data), and step S1
In P2, a plurality of representative colors (colors that appear frequently in the image) are selected from the image to be registered, and step SP
In step 3, the unselected color is replaced with a representative color that can be most approximated, and in step SP4, a partial image area is obtained for each representative color. Then, in step SP5, it is determined whether or not there is a partial image area whose image quality change cannot be visually identified. If there is a partial image area whose image quality change cannot be visually identified, then in step SP6 The partial image area whose change in image quality cannot be identified is integrated with any partial image area, and the determination in step SP5 is performed again. If it is determined in step SP6 that there is no partial image area whose image quality change cannot be visually identified, the number of each partial image area is counted in step SP7, and the number of partial image areas is normalized in step SP8. In step SP9, the number of normalized areas is stored in the recording medium such as an optical disk together with the original image and the keyword information. The series of processes described above is performed for each original image.

A more detailed description will be given. Above step SP
Regarding the processing of 1, SP2 and SP3, for example, after the color data (for example, R, G, B data) of the processing target image is converted into chromaticity data (for example, L * a * b * data), the appearance frequency The corresponding chromaticity values are selected as the representative color in descending order, and the colors not selected as the representative color are replaced with the chromaticity value of the representative color having the shortest distance on the chromaticity diagram.

The determination in step SP5 is made based on the color difference between the adjacent partial image areas and the size of the partial image area to be processed. Specifically, the identification shown in FIG. 8 is based on whether or not the image quality change can be visually identified as a whole when various partial image areas are integrated and removed to other partial image areas having the smallest color difference. An unreadable area is created, and it is determined based on whether the target partial image area belongs to the unidentifiable area shown in FIG. In FIG. 8, the horizontal axis represents the area size (the number of pixels × the number of pixels), and the vertical axis represents the color difference ΔEab ′ (for example, C
1/10 of the color difference value of IE's 1976L * a * b * color system
Value). When the boundary line defining the unidentifiable area is approximated by a function, when the area size is 6 or less, ΔEa
When b ′ = ∞ and the area size is larger than 6, the area size is 7, 10, 20, 30, 40, 50, 60, 10 based on the identification results by a large number of subjects.
ΔEab ′ is 93, 83, 70, 62, 58, 52, 50, 44, 3 corresponding to 0, 200, 500 respectively.
Since the point sequence of 6,36 is applicable, Y = 64 · exp
It can be approximated by a function of (-0.026X) +36. However, X is the area size and Y is the value of ΔEab ′.

The processing in step SP8 can be achieved by, for example, dividing the number of regions by the total number of pixels in the image. Therefore, by performing the above-described series of processing, the original image can be divided into a plurality of partial image areas without substantially degrading the image quality of the original image, and the normalized area is divided based on the number of the divided partial image areas. The number can be calculated. Then, the calculated normalized area number can be stored in a storage medium such as an optical disk together with the original image, keyword information, etc., and an image database for search can be completed. As a result, an objective occupancy rate can be obtained and stored without being affected by the operator's will. Here, if the number of normalized regions is high, there will be many targets for each partial image region in the image. Conversely, if the number of normalized regions is low, there will be few targets for each partial image region. Therefore, the complexity of the image can be expressed with high accuracy by the number of normalized regions.

The flowchart of FIG. 7 is different from the flowchart of FIG. 2 only in that the number of normalized areas is used instead of the occupation rate of the boundary line, and the other points are the same, and therefore detailed description will be omitted. As is clear from the above description, each image extracted from the image database based on the search condition is displayed for the time determined based on the number of normalized regions, so that the display time of each image is minimized. As a result, the time required as a whole can be shortened, and oversight of the image to be searched can be eliminated to achieve reliable image search.

[0037]

[Embodiment 3] FIG. 9 is a block diagram showing a portion for creating an image database in an embodiment of the image retrieval apparatus of the present invention. An image capturing unit 1 for capturing an image that can be a search target, and a captured image. Is temporarily stored, a color data conversion unit 3 that extracts color data for each pixel from the image storage unit 2 and converts it into chromaticity data, a median filter based on the converted chromaticity data, and the like. Smoothing processing unit 4 for performing a smoothing process using, a differential processing unit 5a for calculating a color difference differential value and a threshold value process based on the color difference differential value, and a thinning process for performing a thinning process for a boundary line having a line width of "1". Boundary line detection unit 5 having thinning processing unit 5b for obtaining
And a boundary line occupancy calculation unit having a boundary line pixel counting unit 6a that counts all the pixels that form the boundary line and a calculation unit 6b that calculates the boundary line occupancy by dividing the number of boundary line pixels by the total number of pixels in the image. 6 and an image storage unit 7a for storing the image held in the image holding unit 2 and a keyword table storage unit 7b for storing the keyword input by the keyword input unit 8, the boundary occupancy ratio, and the image pointer indicating the image storage address. And an image database 7 having

The operation of each component is the same as the process of the corresponding step of the flowchart of FIG. 1, and therefore detailed description will be omitted. Therefore, by simply specifying or reading an image to be registered, conversion processing, smoothing processing, boundary line detection processing, and boundary occupancy ratio calculation processing are performed automatically, and keyword management tables, etc., along with keyword information, etc. Since it can be stored in, it is possible to store it by obtaining an objective occupancy rate without being affected by the operator's will.

FIG. 10 is a block diagram showing a portion for performing image retrieval based on the image database in the embodiment of the image retrieval apparatus of the present invention. The retrieval condition input unit 11 for inputting a keyword and the like, and an input By accessing the keyword table storage unit 7b based on the retrieved search conditions, the image retrieval unit 12 that extracts the image pointer and the boundary occupancy of the corresponding image; and the image retrieval unit 1
Search result holding unit 1 that holds the search results obtained in 2
3 and sequentially retrieve the search results from the search result holding unit,
An image display control unit 14 that supplies an image pointer as read instruction data to the image storage unit 7a and outputs the boundary occupancy ratio, and an image display time calculation unit that calculates the image display time by inputting the boundary occupancy ratio. 15 and an image display unit 16 that visually displays the search target image for the calculated image display time based on the image data read from the image storage unit 7a. It should be noted that a synchronization signal is exchanged between the image display control unit 14 and the image display unit 16 to synchronize the operations of both.

The operation of each component is the same as the process of the corresponding step of the flowchart of FIG. 2, and therefore detailed description will be omitted. As is clear from the above, each image extracted from the image database based on the search condition is
Since it is displayed only for the time determined based on the occupancy ratio of the boundary line, the display time of each image can be minimized, the required time can be shortened as a whole, and the image to be searched can be searched without overlooking. The certainty of can be achieved.

[0041]

[Embodiment 3] FIG. 11 is a block diagram showing a portion for creating an image database in another embodiment of the image retrieval apparatus of the present invention. The difference from the embodiment of FIG. 5 and the boundary line occupation ratio calculation unit 6 instead of the region division unit 25 and the normalized region number calculation unit 26.

More specifically, the area dividing section 25 determines, based on the chromaticity values of each pixel converted by the color data converting section 23, a plurality of chromaticity values having a high appearance frequency, the chromaticity values of the representative color. A representative color selecting unit 25a, a replacing unit 25b that replaces a chromaticity value that is not selected as a representative color with a chromaticity value that can be most approximated, and a partial image for each representative color based on the replacement result by the replacing unit 25b. An image dividing unit 25c that obtains a region, an area discriminating unit 25d that discriminates whether or not an image quality change can be visually identified for each partial image region obtained by the image dividing unit 25c, and an image quality change visually. In response to the discrimination result of the area discriminating unit 25d indicating that the area cannot be discriminated, the corresponding area is integrated with the adjacent area having the smallest chromaticity difference among the adjacent areas. The operation of each component is the same as the process of the corresponding step of the flowchart of FIG. 6, and thus detailed description will be omitted.

The normalized area number calculation section 26 counts the number of partial image areas divided by the area dividing section 25 and an area counting section 26a that divides the number of partial image areas by the total number of pixels to normalize. It has a region number normalization unit 26b for performing conversion. Therefore, the original image can be divided into a plurality of partial image regions without substantially degrading the image quality of the original image, and the number of normalized regions can be calculated based on the number of divided partial image regions. Then, the calculated normalized area number can be stored in a storage medium such as an optical disk together with the original image, keyword information, etc., and an image database for search can be completed. As a result, an objective occupancy rate can be obtained and stored without being affected by the operator's will. here,
If the number of normalized regions is high, there will be many targets for each partial image region in the image. Conversely, if the number of normalized regions is low, there will be few targets for each partial image region. Therefore, the complexity of the image can be expressed with high accuracy by the number of normalized regions.

FIG. 12 is a block diagram showing a portion for performing image retrieval based on an image database in another embodiment of the image retrieval apparatus of the present invention. The point different from the block diagram of FIG. 27b holds the number of normalized regions instead of the occupancy ratio of the boundary line, and only the difference is that the image display time calculation unit 15 calculates the image display time based on the number of normalized regions. is there.

As is clear from the above description, each image extracted from the image database based on the search condition is
Since it is displayed only for the time determined based on the number of normalized areas, the display time for each image can be minimized and the required time can be shortened as a whole. Certainty can be achieved. still,
The present invention is not limited to the above-described embodiment. For example, as a value indicating the complexity of an image, the number of colors used in the image (generally, an image having a larger number of colors is A), the spatial frequency distribution of the image (generally, an image containing many high-frequency components is an image in which the brightness on the image changes drastically and is a complicated image), and the number of parts with a high curvature on the boundary line ( On the boundary line, the observer's viewpoint tends to gather in a portion with a high curvature, so an image with many portions with a high curvature has to be watched by the observer in many cases, which is a complicated image), and the variance of the luminance change ( In general, it is possible to adopt an image in which the brightness changes drastically as it is difficult to see, and it is possible to make various design changes within the scope of the present invention. Is.

[0046]

As described above, according to the first aspect of the invention, unlike the case where the operator sets the display time of each image, the work of setting the display time can be eliminated and the optimum amount of information can be optimized for each image. The display time can be set, and as a whole, a unique effect that both high-speed image search and reliable image search can be achieved is achieved.

According to the second aspect of the present invention, when each image is stored in the image storage means, the boundary line is detected and the number of pixels of the boundary line is calculated. Therefore, it is possible to obtain a unique effect that both high-speed image search and reliable image search can be achieved. According to the third aspect of the present invention, when storing each image in the image storage means, it is possible to automatically obtain objective information indicating complexity by performing area division and calculation of the number of divided areas. As a result, there is a unique effect that both high-speed image search and reliable image search can be achieved.

According to the invention of claim 4, unlike the case where the operator sets the display time of each image, the work of setting the display time can be eliminated, and the optimum display time corresponding to the information amount can be set for each image. As a result, as a whole, there is a unique effect that both high-speed image retrieval and reliable image retrieval can be achieved.

[Brief description of drawings]

FIG. 1 is a flowchart illustrating a process of registering an image in an image database in one embodiment of the image search method of the present invention.

FIG. 2 is a flowchart illustrating a process of searching an image based on an image database obtained by the process of the flowchart of FIG.

FIG. 3 is a diagram illustrating an observation window for smoothing processing.

FIG. 4 is a diagram illustrating an operation window for thinning processing.

FIG. 5 is a diagram showing a relationship between a central pixel to be processed and an adjacent pixel.

FIG. 6 shows another embodiment of the image search method of the present invention.
7 is a flowchart illustrating a process of registering an image in an image database.

FIG. 7 is a flowchart illustrating a process of searching an image based on an image database obtained by the process of the flowchart of FIG.

FIG. 8 is a diagram showing an unidentifiable area.

FIG. 9 is a block diagram showing a portion for creating an image database in the embodiment of the image retrieval apparatus of the present invention.

FIG. 10 shows an embodiment of the image search device of the present invention.
It is a block diagram showing a portion which performs an image search based on an image database.

FIG. 11 is a block diagram showing a portion for creating an image database in another embodiment of the image retrieval apparatus of the present invention.

FIG. 12 is a block diagram showing a portion that performs image retrieval based on an image database in another embodiment of the image retrieval apparatus of the present invention.

[Explanation of symbols]

7,27 Image database 12,32 Image retrieval unit 15,35 Image display time calculation unit 16,36
Image display section

Claims (4)

[Claims] 1. A plurality of images are stored in an image storage means (7) (2).
Image retrieval in which the corresponding images are sequentially stored in the image storage means (7) and (27) based on predetermined retrieval conditions and visually displayed by the display means (16) and (36). In the method, each image is stored in the image storage means (7) (2
When storing in 7), the information indicating the complexity of the image is also stored in the image storing means (7) (27), and the image is stored in the image storing means (7) (27) together with the corresponding image. An image retrieval method characterized by extracting information indicating the image quality and controlling the visible display time of the corresponding image based on the information indicating the complexity of the image. 2. The image search method according to claim 1, wherein the information indicating the complexity is an occupancy rate of a boundary line in the image. 3. Information indicating complexity is included in an image,
The image search method according to claim 1, wherein the number of areas is defined by a plurality of representative colors. 4. An image storage means (7) (2) for storing a plurality of images.
Image retrieval in which the corresponding images are sequentially stored in the image storage means (7) and (27) based on predetermined retrieval conditions and visually displayed by the display means (16) and (36). In the device, the image storage means (7) (27)
Information that indicates the complexity of the image is stored together with the stored image, and extraction means (12) (32) that extracts the information indicating the complexity of the image together with the corresponding image from the image storage means (7) (27). ) And display time control means (15) (35) for controlling the visible display time of the corresponding image based on the information indicating the complexity of the image.