JP2008217117A - Image retrieval method and image retrieval system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image retrieval method and an image retrieval system, capable of reaching a target image in a short time period. <P>SOLUTION: A personal computer 10 is provided with a CPU 20, a memory 21, a monitor 22, an operation unit 23, an image database 25, and a keyword database 26. If an image retrieval program PG stored in the memory 21 is started, the CPU 20 functions as a browser control unit 29, a related keyword retrieval unit 30, and an image retrieval unit 31. If a retrieval keyword is input by operation of the operation unit 23, the related keyword retrieval unit 30 retrieves a related keyword from the keyword database 26. The image retrieval unit 31 retrieves image data corresponding to the retrieval keyword and the related keyword from the image database. The browser control unit 29 makes the monitor 22 display the retrieval keyword, the related keyword, and a thumbnail image of the corresponding image data. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image retrieval method and an image retrieval system for retrieving an image from an accumulated image database.

  It is common to accumulate a large number of images such as photographs and illustrations and store them in an image database, browse the images, and read out and process the necessary images. It is desired to quickly search for an image that approximates the user's image from a large number of images stored in such an image database. For example, in the image search method described in Patent Document 1, a keyword-based method is used. Image search and content-based similar image search are combined. From these search results, images determined to have similar keywords, colors, textures, shapes, etc. are weighted with their assigned coefficients. Then, feedback is performed so that a highly relevant image and a less relevant image can be identified to the user according to these weightings.

Patent Document 2 describes an image search method for searching for a similar image using a color and a feature amount of an image as a search key. By specifying an image and a keyword as a search key by an image searcher in advance, It is possible to search for an image having a similar keyword or image feature amount in contrast to a keyword assigned to the image or a category in which the image is classified.
JP 2002-197117 A JP 2002-63172 A

  However, if the image of the target image has already been determined by the user when searching for the image, it is effective to narrow down the search using a keyword or shape selection indicating the target image, as in the above-mentioned patent document. However, if the image of the image desired by the user is still vague, the keyword or shape that accurately indicates the target image is not determined, and a search method such as the above-mentioned patent document does not reach the target image. It may take time, and even if you browse a large number of images, you may not be able to associate the images and reach the images.

  The present invention has been made in view of the above circumstances, and provides an image search method and an image search system that can arouse an image of a user who performs an image search and can reach a target image in a short time. For the purpose.

    In the image search method of the present invention, in an image search method for searching for an image from an image database storing a plurality of image data, a first related keyword search step of searching for a related keyword related to an input search keyword from the keyword database. A first image search step of searching the image database for images associated with the search keyword and the related keyword, and displaying thumbnail images of a plurality of images searched in the image search step on the display means. And a first image display step.

  In the first image display step, it is preferable to place an image corresponding to the related keyword around the image corresponding to the search keyword. After the first image display step, when one image is selected from the images displayed on the display means, a related keyword related to the keyword of the selected image is searched from the keyword database. A second related keyword search step; a second image search step of searching the image database for an image associated with the related keyword; and a plurality of images associated with one selected image and the related keyword. And a second image display step for displaying on the display means. Alternatively, after the first image display step, when one image is selected from among the images displayed on the display means, an image associated with the same keyword as the selected image is stored in the database. It is preferable to include a third image search step for searching from the third image display step and a third image display step for displaying a plurality of images associated with these same keywords on the display means.

  In addition, information data related to a plurality of items is added to the image stored in the image database, and a plurality of evaluation axes indicating the items of the information data are set on the display screen in the image display step. The images are arranged in a matrix in the order of the information data along these evaluation axes, and one image is selected from the displayed images after the first or second image display step. A display scroll step of scrolling to a display screen centered on the image. Furthermore, information data relating to a plurality of items is added to the images stored in the image database, and only image data that meets the narrowing conditions when the narrowing conditions for these information data are input. And a refinement search step for searching for.

    The image search system according to claim 7, wherein an image database storing a plurality of images, a keyword database storing keywords associated with the images, an input means for inputting a search keyword, and an input search keyword In the image search system comprising display means for displaying an image searched from the image database using, the keyword is associated with another keyword in advance, and when the search keyword is input, Related keyword search means for searching for a related keyword related to the search keyword from the keyword database, image search means for searching the image database for the search keyword and an image associated with the related keyword, and the image search means Place the image searched in A first image display screen is characterized in that a display control means for displaying on the display means.

  The first image display screen preferably has an image corresponding to the search keyword and an image corresponding to the related keyword arranged around the image. A selection input unit that selects one image from the first image display screen; and the related keyword search unit stores a related keyword related to a keyword of the image selected by the selection input unit from a keyword database. The image search unit searches the image database for an image associated with the related keyword, and the display control unit arranges one selected image and a plurality of images associated with the related keyword. The displayed second image display screen is preferably displayed on the display means. Alternatively, the image search device includes selection input means for selecting one image from the first image display screen, and the image search means associates the same keyword with the image selected by the selection input means. It is preferable that the display control unit displays a third image display screen in which a plurality of images associated with the same keyword are arranged on the display unit, retrieved from the database.

  Further, the image data stored in the image database is respectively added with information data regarding a plurality of items, and on the image display screen, a plurality of evaluation axes indicating the items of the information data are set. The display control means arranges the images in a matrix in the order of the information data along the evaluation axis, and scrolls around the image when one image is selected from the image display screen. It is preferable to display. Still further, information data related to a plurality of items is added to the images stored in the image database, and a narrowing condition input means for inputting a narrowing condition related to the information data, and the image search means includes the narrowing down It is preferable to perform a narrowing search that uses only image data that meets the conditions as a search target.

  The image search method and the search system according to the present invention search for a related keyword related to an input search keyword from the keyword database, and search for images associated with the search keyword and the related keyword from the image database. Since the image is displayed on the display means, the image of the user who searches for the image can be aroused and the target image can be reached in a short time.

  Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows an example of a personal computer (PC) 10 as an image search system to which the present invention is applied. The PC 10 is totally controlled by the CPU 20 as a whole. The CPU 20 is connected to a memory 21, a monitor 22, an operation unit 23, an image database 25, a keyword database 26, and the like.

  The memory 21 is composed of a semiconductor memory, a magnetic disk, or the like, and stores various programs such as an operating system and data necessary for starting up the PC 10 and an image search program PG described later. The memory 21 temporarily stores image data and keyword data read from the image database 25 and the keyword database 26.

  The monitor 22 uses an LCD, a CRT, or the like, and displays each execution screen when the image search program PG is activated. The operation unit 23 includes a keyboard 24a and a mouse 24b. The CPU 20 controls the operation of each unit of the PC 10 in response to the operation of the operation unit 23.

  When the image processing program PG is activated by the operation of the operation unit 13, the CPU 20 generates processing functions such as a related keyword search unit 30, an image search unit 31, and a synonym dictionary 32 in addition to the browser control unit 29. The browser control unit 29 displays an image display browser 30 as a GUI (Graphical User Interface) as shown in FIG. 2 on the screen of the monitor 22.

  The image display browser 30 is provided with an input space 32a for inputting a keyword for search, a thumbnail window 32b for displaying a thumbnail image 33 of the searched image, and a main window 32c for displaying a selected image 34. ing. The selected image 34 is displayed on the main window 32c by placing the cursor 36 on the thumbnail image 33 and clicking the click button of the mouse 24b. The selected image 34 is displayed larger than the thumbnail image so that the contents of the image can be easily recognized. The cursor 36 is a pointer that moves in conjunction with the movement of the mouse 24b.

  In the present embodiment, the keyword data stored in the keyword database 26 has a structure shown in FIG. 3, and a keyword to be matched with a search keyword and a related keyword related to the keyword are tagged and stored. ing. Further, usage frequency data that is weighted according to the usage frequency is added to the keyword data, and the weighting value of the usage frequency data is updated every time each keyword matches in the search.

  Further, the image data stored in the image database is associated with all the keywords stored in the keyword database. For example, as shown in FIG. 4, the keywords are written in tag information attached to the image data. It is. Further, in the present embodiment, in the tag information of the image data, usage frequency data weighted according to the usage frequency, information data indicating shooting information, and the like are written. Furthermore, thumbnail image data obtained by reducing the image data is added.

  The related keyword search unit 30 accesses the keyword database 26, collates keywords that match the search keyword input by the operation unit 23, reads the matched keyword data from the keyword database 26, and related keywords related to the search keyword. Search for. Furthermore, in this embodiment, when one image is selected from a plurality of images displayed in the thumbnail window 32b as described later, a related keyword associated with a keyword corresponding to this image is searched.

  The image search unit 31 accesses the image database 25 and searches for an image associated with the search keyword and the related keyword searched in the keyword database 26.

  Next, a processing procedure when the image search program PG is executed by the PC 10 having the above configuration will be described with reference to the flowchart of FIG. First, when the user operates the operation unit 23 to start the image search program PG (st1), a browser control unit 29, a related keyword search unit 30, an image search unit 31, a synonym dictionary 32, and the like are generated in the CPU 20, and browser control is performed. The image search browser 30 shown in FIG. 2 is displayed on the screen of the monitor 22 by the unit 29. Next, a search keyword is input to the input space displayed in the search keyword window 32a by an input operation using the operation unit 23 (st2).

  When a search keyword is input, the related keyword search unit 30 functions to search for a keyword related to the input search keyword (st3). At this time, the search keyword and the related keyword are temporarily stored in the memory 23.

  Next, it is determined whether or not the number of related keywords extracted by the search is 8 or less in the present embodiment (st4). Note that the number of related keywords determined at this time may be one less than the number of thumbnail images displayed in the thumbnail window 32c, and is not limited thereto.

  If the number of related keywords is less than 8, the synonym dictionary 33 functions to search for similar keywords similar to the search keyword (st5). Usage frequency data indicating weighting according to usage frequency is added to this similar keyword in advance, and a similar keyword with higher usage frequency is selected based on this usage frequency data, and the number of related keywords described above is insufficient. These similar keywords are supplemented as related keywords (st6).

  On the other hand, when the number of related keywords is eight or more, the CPU 20 determines the use frequency based on the use frequency data of the related keyword, and determines eight related keywords having the highest use frequency (st7).

  When eight related keywords are determined or supplementation with similar keywords is completed, the image search unit 31 functions to search for an image corresponding to the search keyword and the related keyword. At this time, since there are a plurality of images corresponding to the keyword, the CPU 20 determines the use frequency based on the use frequency data of each image data, and determines the image data with the highest use frequency (st8).

  When the most frequently used image corresponding to the search keyword and the related keyword is determined, the browser control unit 29 functions to read the thumbnail image data added to the image data, and display it in the thumbnail display window 32b together with each keyword. A first image display screen 41 (see FIG. 6A) on which these thumbnail images 33 are arranged is displayed (st9). In FIG. 6, the image corresponding to the search keyword “bright” is centered, and the related keywords with the highest frequency of use are “fresh”, “fun”, “sea”, “sky”, “smile” The figure shows an example of a first image display screen 41 in which images corresponding to “light”, “summer”, and “orange” are arranged around it.

  When one of the nine thumbnail images in the first image display screen 41 is selected by operating the operation unit 23, the selected image 34 is displayed on the main window 32c. When the selected image is determined, the CPU 20 displays the question screen 46 as to whether or not to continue the search shown in FIG. 7 in another window, or switches the main window 32c to the question screen shown in FIG. 7 (st10). ). If it is selected not to continue the search, the image search process ends, and another program that can save the selected image or process it using the image operates.

  On the other hand, when the search continuation is selected, the related keyword search unit 30 functions to start searching for related keywords related to the keyword of the selected image (st11). Then, when the related keyword is searched, the image corresponding to the related keyword with the highest use frequency is searched in the same manner as in the above processes (st3 to st7), and the thumbnail images 33 added to the plurality of images are arranged. 2 is displayed on the thumbnail window 32b (see FIG. 6B). In FIG. 7, a thumbnail image 33 of “smile” is selected from the thumbnail window 32b in the state of FIG. 6A which is the first image display screen 41, and the search keyword “smile” is selected. , Images of the top related keywords “happy”, “birthday”, “love”, “children”, “happiness”, “holiday”, “facial expression” “good friend” are searched, and thumbnail images of each 6 shows that the state is switched to the state shown in FIG. 6B, which is the second image display screen 42 arranged in the thumbnail window 32b. When an image is selected in this manner, a related keyword related to the keyword and an image corresponding to the related keyword are searched to display the image, and this is repeated until the selected image is confirmed.

  As described above, a screen on which a plurality of images corresponding to a related keyword associated with one keyword is arranged is displayed, and further, a corresponding keyword associated with the keyword of the selected image selected from the plurality of images is supported. Since the display of the screen with multiple images is repeated, the user's image is aroused while the related keywords and images are connected, and the image and keywords are gradually decided. It is possible to reach the image to be performed in a short time.

  In the first embodiment, after displaying the first image display screen in which a plurality of thumbnail images corresponding to the search keyword and the related keyword related to the search keyword are displayed, the search keyword is associated with the keyword of one selected image. However, the present invention is not limited to this, and in the second embodiment of the present invention described below, the first image display screen is arranged in which a plurality of images corresponding to the related keywords are arranged. After the image display screen is displayed, a second image display screen in which another plurality of images corresponding to the keyword of the selected image is arranged is displayed. The personal computer as an image search apparatus to which the second embodiment is applied has the same component configuration as that of the first embodiment, and the image search method to which this embodiment is applied will be described with reference to the flowchart of FIG. To do. When the image search program is started (st1), the image search browser 30 is displayed. When a search keyword is input (st2), the related keyword search unit 30 functions to search the keyword database 26 for related keywords. Further, when the image search unit 31 functions to search the image corresponding to the search keyword and the related keyword from the image database 25 and determine the image (st3 to st8), the thumbnail image data added to the image data is read, A first image display screen 51 (see FIG. 10A) in which these thumbnail images 33 are arranged is displayed in the thumbnail display window 32b together with the keywords (st9).

  When one of the nine thumbnail images in the first image display screen 51 is operated to select one, the selected image 34 is displayed on the main window 32c. When it is selected that the search is not continued after the selected image is determined, the image search process is ended, and another program that can store or process the image is operated.

  On the other hand, when the search continuation is selected, in this embodiment, the image search unit 31 functions to search for image data associated with the same keyword as the selected image 34 (st10), and from the images with the matching keyword. Nine image data with the highest usage frequency are determined (st11). In this way, the top nine usage frequencies of the same keyword, that is, the previously selected image 34 is the highest usage frequency. When the second to ninth usage images are determined in addition to this selection image 34, browser control is performed. The unit 29 functions to read out the thumbnail image data added to the image data, and a third image display screen 53 (FIG. 10 (FIG. 10) in which these thumbnail images 33 are arranged in the thumbnail window 32b together with the keywords and the order of use frequency. B)) is displayed (st12). In FIG. 10, the thumbnail image 33 of “smile” is selected from the thumbnail window 32 b in the state of FIG. 10A that is the first image display screen 51, and the usage associated with “smile” is used. It shows that the nine images with the highest frequency are retrieved and switched to the state of FIG. 10B, which is the third image display screen 53 in which the respective thumbnail images 33 are displayed in the thumbnail window 32b. A keyword (smile) displayed on the third image display screen 53 is accompanied by a number indicating the order of use frequency. When one of the nine thumbnail images in the third image display screen 53 is selected by operating the operation unit 23, the selected image is displayed on the main window 32c. When the selection is confirmed, the image search ends.

  As described above, the first image display screen in which a plurality of images corresponding to the related keyword associated with the search keyword is arranged is displayed, and the same keyword as the selected image selected from the plurality of images is displayed. Since the third image display screen on which a plurality of images are arranged is displayed, a keyword and an image to be imaged by the user are determined from one keyword, and further images can be obtained by browsing a plurality of images with the same keyword. Since it is confirmed, it is possible to reach the target image in a short time.

  The first and second embodiments can be combined with each other. That is, the same process is performed until the first image display screen is displayed. When performing a selection operation for determining a selection image from the first image display screen, for example, one is simply a normal click with the mouse, and the other is the mouse. Hold the clicked state (while keeping the button pressed for a long time) or double-click, etc., and set different operations. Depending on the difference in the operation, the second image is displayed in one operation as in the first embodiment. The process proceeds to the display screen, and the other operation can be a process to proceed to the third image display screen as in the second embodiment. Alternatively, the first image display screen may be switched to the second image display screen, and the selection may be switched to the third image display screen when the selected image is determined from the second image display screen. As a result, the same effects as those of the first and second embodiments can be obtained, and options for the image search method can be increased. In the second embodiment, the first image display screen is completely switched to the third image display screen. However, the present invention is not limited to this. For example, the selected image is determined from the first image display screen. Sometimes, the third image display screen (that is, the thumbnail image having the same keyword as the selected image) may be reduced and displayed by dividing only the portion where the thumbnail image of the selected image exists.

    In the first and second embodiments, the search keyword or the related keyword is searched for and an image having the highest use frequency is arranged. However, the third embodiment of the present invention described below is used. In the embodiment, a configuration in which an image is searched by setting an evaluation axis based on information data in addition to keyword matching is illustrated. In the third embodiment, shooting information data when an image is shot as data added to the image data, for example, shooting date, lens type, exposure value, aperture value, and the like are stored in association with each other. Also, information extracted from the image data, for example, density data relating to color shading, is digitized in advance and stored together with the image data. The image search apparatus to which the third embodiment is applied has the same component configuration as the first and second embodiments, and an image search method to which this embodiment is applied will be described with reference to the flowchart of FIG. . When the image search program PG is started (st1), the image search browser 30 is displayed. When a search keyword is input, the related keyword search unit functions to search for a related keyword (st2). Further, when the image corresponding to the search keyword and the related keyword is searched from the image database 25 and the image is determined, the thumbnail image data added to the image data is read and these thumbnail images are arranged in the thumbnail display window 32b together with each keyword. The first image display screen is displayed (st3 to st9).

  When one of the nine thumbnail images in the first image display screen is operated to select one, the selected image is displayed on the main window 32c. When the selected image is determined, the CPU 20 displays a question screen as to whether or not to continue the search. Another program that selects not to continue the image search, saves the image, or processes using the image operates.

  On the other hand, when search continuation is selected, in this embodiment, the image search unit 31 functions to search for image data associated with the same keyword as the selected image in accordance with the setting of the evaluation axis (st12). At this time, surrounding images are determined along the above-described evaluation axis with the thumbnail image 33e of the selected image 34 as the center from the images with the matching keyword, and the image display screen 61 shown in FIG. 11 is displayed in the thumbnail window 32b. To display. In the present embodiment, since the X axis is an evaluation axis indicating the shooting date, the images 33b and 33h at the same position as the thumbnail image 33e of the selected image 34 on the X axis are those having the same shooting date or close date as the selected image 34. Yes, so that the date gradually changes to a newer one when proceeding in the positive direction of the X axis (positions of images 33c, 33f, and 33i), and the older one when proceeding further in the negative direction (positions of images 33a, 33d, and 33g). Since the Y axis is an evaluation axis indicating the image density, the images 33d and 33f at the same position as the thumbnail image 33e of the selected image 34 on the Y axis have the same or similar brightness (image density as the selected image). ) In the positive direction of the Y-axis (positions of the images 33a, 33b, 33c) gradually becomes brighter (low density), and further in the negative direction (positions of the images 33g, 33h, 33i) is dark ( Degree is high) to place things.

  When nine images associated with the same keyword are determined in this way, the browser control unit 29 functions to read out the thumbnail image data added to the image data and to display the thumbnail window 32c together with each keyword and shooting information data. The third image display screen 61 in which these thumbnail images 33a to 33i are arranged is displayed (st13).

  At this time, on the third image display screen 61, the X and Y axes are set as described above, and are arranged in a matrix in the order of the shooting date and the brightness of the image. When one of the nine thumbnail images 33a to 33i is selected, this selected image is displayed on the main window 32c. Then, whether or not to continue the search is selected. When the search is continued, the image search unit 31 and the browser control unit 29 function, and the selected image selected earlier is located at the center. Since other images are arranged along the X and Y axes, the screen scrolls as a result (st13 to st15). That is, if an image at a position that advances in the positive / negative direction of the X axis with respect to the central image is selected, the image is gradually switched to a new or old image on the shooting date, and exposure is performed by selecting an image at a position that advances in the positive / negative direction of the Y axis. Scroll so that the value gradually switches to a larger or smaller image. Thereafter, each time an image is selected, the display screen is switched along the X and Y axes to scroll the image.

  In this way, it is possible to search for an image with a matching keyword, and further to search while scrolling the image along the evaluation axis based on the shooting information data. In addition to the keyword, other information can be taken into account. Therefore, it is possible to make the user arouse the image from the image, the keyword, and various information, and it is possible to reach the target image in a short time.

  Further, in the third embodiment, as an example of the image display screen for setting the evaluation axis, a case where it is applied to the third image display screen in which a plurality of images corresponding to the same keyword are arranged is exemplified. The present invention is not limited to this. Even when the second image display screen described in the first embodiment, that is, the selected image and the image corresponding to the keyword associated with the keyword of the selected image are arranged, As in the third embodiment, when an evaluation axis is set on the display screen and an image is arranged along the evaluation axis and an image on the image display screen is selected, the image is scrolled around the selected image. You may arrange so that.

  Furthermore, in the third embodiment, only two axes of the vertical and horizontal (X, Y axes) are set as the evaluation axes. However, the present invention is not limited to this, and 3 in the oblique direction in addition to the vertical and horizontal directions. An evaluation axis is set in the direction of the four or four axes, an image is arranged along the evaluation axis, and when an image is selected in the direction of the evaluation axis, the image is scrolled vertically and horizontally and diagonally. You can do it.

  In addition, when a search is performed in the above embodiment, a plurality of images that approximate the target may be found. Accordingly, a fourth embodiment of the present invention that can further narrow down an image in addition to the configurations of the first to third embodiments will be described below. Note that the image search apparatus to which the fourth embodiment is applied has the same component configuration as that of the first and second embodiments, and a description thereof will be omitted.

  Shooting information data is added to the image data. Note that data conforming to the EXIF standard is used as an example of photographing information data. Furthermore, in this embodiment, when image data is stored in the image database 25 so that the narrowing search described later can be performed in a short time, the shooting information data is applied to the search condition data applied to the search conditions defined in advance. Convert and save with image data.

  As an example of the conversion from the shooting information data to the search condition, the shooting information data whose shooting date is “1990 to 1999”, “April to June”, and “December 24” are “ It is converted into search condition data such as “1990s”, “Spring”, “Christmas Eve”, and written in the tag information area of image data. Similarly, in the photographing time of the photographing information data, “5-7 am” and “5-7 pm” are converted to “morning” and “evening”, and “20-30 mm” for the photographing lens type. And “200 to 300 mm” (focal length) are converted into “wide angle” and “telephoto”, respectively, and “1 million pixels” and “10 million pixels” (the number of effective pixels at the time of shooting) are “ “Low resolution” and “High resolution”, and “f1.4 to 2.8” and “f11 to 16” (aperture value) are “shallow” and “deep” (depth of field) at the photographing aperture. The shutter speeds “1/500 to 1/2000” and “1/4 to 1” (seconds) are converted into search conditions “fast” and “slow”.

  The search condition data is not limited to the above-described shooting information data, and may be formed from color and shape feature amounts extracted from image data. For example, the color feature amount of the image, that is, the ratio of each color included in the image is digitized in advance and stored together with the image data. Alternatively, the feature amount of the shape, that is, a representative shape included in the image is extracted in advance, and the shape data is stored together with the image data. As the shape extraction, for example, when a plurality of templates and shapes in an image are matched and a matching or similar shape is included, data indicating the template is stored as shape data.

  Furthermore, in this embodiment, the image display browser when the image program PG is started is provided with a search condition input window 70 as shown in FIG. 12, and the search condition input window includes a shooting information data selection field. 71, a color input field 72, a shape input field 73, and a refinement search button 74 are provided. The shooting information data selection field 71 can be input by clicking each option. The color input field 72 can be input by clicking a desired color from the color palette 76, and a plurality of colors can be selected at the same time. Furthermore, the shape input field 73 is a well-known handwriting input palette 77, and a mouse can be used to draw and input a shape as if it were handwritten in the palette.

  The operation of the above configuration will be described. For example, when the first keyword is input in the first to third embodiments, the search refinement of this embodiment can be combined. That is, in the first to third embodiments, after starting the image search program PG, the search keyword is input, the search condition for the search is input, and the search search button is clicked, the image search unit 31 functions. Start a refined search. In the narrowing search, first, an image whose search condition data matches the search condition input earlier is searched from images stored in the image database. As a result, the search target is narrowed down to a partial image in the image database.

  Next, the CPU operates related keyword search means to search for related keywords. Then, the image search unit 31 functions to search the image database 25 for images corresponding to the search keyword and the related keyword. At this time, when an image with a matching keyword is determined from the images whose total number has been narrowed down by the above-described narrowing search, the thumbnail image data added to the image data is read, and these thumbnail images are displayed in the thumbnail display window together with each keyword. A first image display screen in which is placed is displayed. Thereafter, a process of switching to the second and third screen display or a process of determining one image is executed, and an image search is performed.

As described above, since the search using the search keyword is performed after narrowing down the images in the image database using the search condition data added to the image data, the search time is shortened, and the target image is reduced to a short time. Can be reached.

  In the first to fourth embodiments, the image database and the keyword database are built in the PC 10 and the keyword and the image are searched by accessing the database in the PC 10. The present invention is not limited to this, and the present invention can be applied to an image search system including a PC having a communication I / F and a server connected via the Internet. That is, when the image search program is executed on the PC 101 and the related keyword search unit functions, the communication I / F connects to the Internet to search for a related keyword from the keyword database, and when the image search unit functions, the communication I / F F connects to the Internet and retrieves images from the image database. The keyword data and image data extracted by the search are stored in the memory. In this way, keyword search and image search can be performed, and the same effect as in the above embodiment can be obtained.

It is a block diagram which shows the schematic structure of the image search device to which this invention is applied. It is explanatory drawing which shows the display browser when performing an image search. It is explanatory drawing which shows an example of the structure of keyword data. It is explanatory drawing which shows an example of the structure of image data. It is a flowchart which shows a sequence when performing the image search to which 1st Embodiment is applied. It is explanatory drawing which shows the image display screen when performing the image search to which 1st Embodiment is applied. It is explanatory drawing which shows the question screen which asks whether it continues search. It is a flowchart which shows a sequence when performing the image search to which 2nd Embodiment is applied. It is explanatory drawing which shows the image display screen when performing the image search to which 2nd Embodiment is applied. It is a flowchart which shows a sequence when performing the image search to which 3rd Embodiment is applied. It is explanatory drawing which shows the image display screen when performing the image search to which 3rd Embodiment is applied. It is explanatory drawing which shows a display browser when performing the image search to which 4th Embodiment is applied.

Explanation of symbols

10 Personal computer (image search system)
20 CPU
22 Monitor (display means)
23 Operation Unit 25 Image Database 26 Keyword Database 30 Related Keyword Search Unit 31 Image Search Unit 32 Synonym Dictionary

Claims (12)

In an image search method for searching for an image from an image database storing a plurality of image data,
A first related keyword search step of searching a keyword database for a related keyword related to the input search keyword, and a first image search for searching the image database for the search keyword and an image associated with the related keyword An image search method comprising: a step; and a first image display step for displaying thumbnail images of a plurality of images searched in the image search step on a display means.   2. The image search method according to claim 1, wherein in the first image display step, an image corresponding to the related keyword is arranged around the image corresponding to the search keyword.   After the first image display step, when one image is selected from the images displayed on the display means, a related keyword related to the keyword of the selected image is searched from the keyword database. A second related keyword search step; a second image search step of searching the image database for an image associated with the related keyword; and a plurality of images associated with one selected image and the related keyword. The image search method according to claim 1, further comprising a second image display step of displaying on the display means.   After the first image display step, when one image is selected from the images displayed on the display means, the database is searched for an image associated with the same keyword as the selected image. 4. A third image search step, and a third image display step for displaying a plurality of images associated with the same keyword on the display means. Image search method.   Information stored in the image database is appended with information data related to a plurality of items, and in the image display step, a plurality of evaluation axes indicating the items of the information data are set on a display screen, When the images are arranged in a matrix in the order of information data along these evaluation axes, and one image is selected from the displayed images after the first or second image display step. 5. The image search method according to claim 1, further comprising a display scroll step of scrolling to a display screen centered on the image.   Information stored in the image database includes information data related to a plurality of items, and when the narrowing-down conditions regarding these information data are input, only image data that meets the narrowing-down conditions is searched. The image search method according to claim 1, further comprising a target narrowing search step. An image database storing a plurality of images, a keyword database storing keywords associated with the images, an input means for inputting a search keyword, and searching from the image database using the input search keyword In a search system comprising display means for displaying an image,
The keyword is associated with another keyword in advance, and when the search keyword is input, a related keyword search means for searching for a related keyword related to the search keyword from the keyword database, the search keyword, Image search means for searching for an image associated with the related keyword from the image database; and display control means for displaying on the display means a first image display screen on which the image searched by the image search means is arranged. An image search system comprising:   8. The image search system according to claim 7, wherein the first image display screen arranges an image corresponding to the related keyword around the image corresponding to the search keyword.   A selection input unit that selects one image from the first image display screen; and the related keyword search unit stores a related keyword related to a keyword of the image selected by the selection input unit from a keyword database. The image search unit searches the image database for an image associated with the related keyword, and the display control unit arranges one selected image and a plurality of images associated with the related keyword. 9. The image search system according to claim 7, wherein the second image display screen is displayed on the display means.   Selection input means for selecting one image from the first image display screen; and the image search means from the image database associated with the same keyword as the image selected by the selection input means. 9. The image search system according to claim 7 or 8, wherein the display control means causes the display means to display a third image display screen in which a plurality of images associated with the same keyword are arranged. .   Information stored in the image database is appended with information data related to a plurality of items, and a plurality of evaluation axes indicating the items of the information data are set on the image display screen. The means arranges the images in a matrix in the order of the information data along the evaluation axis, and displays a scroll display centered on the image when one image is selected from the image display screen. The image search system according to claim 7, wherein the image search system is performed. Information stored in the image database is appended with information data related to a plurality of items, and a filtering condition input means for inputting filtering conditions related to the information data and the image search means conform to the filtering conditions. The image search system according to claim 7, wherein the search is performed by limiting only the image data to be searched.

Images