Image extraction apparatus, image extraction method, and recording medium

Abstract

Image extraction apparatus, image extraction method, and recording medium. Provided are: a similarity judgment condition reception mechanism that receives input of a similarity judgment condition made by a user to judge whether any two images among a plurality of images are similar to each other; a similar image group determination mechanism that determines one or more similar image groups each composed of two or more images similar to each other, among images having a matching degree of a threshold value or more, by determining images satisfying the similarity judgment condition received by the similarity judgment condition reception mechanism, or determines one or more similar image groups each composed of two or more images similar to each other, among images satisfying the similarity judgment condition received by the similarity judgment condition reception mechanism, by determining images having a matching degree of a threshold value or more; and an image extraction mechanism that extracts at least one image in each of the one or more similar image groups determined by the similar image group determination mechanism.

Description

[0001] This application is a divisional application of application number "201811144335.5", filed on September 28, 2018, entitled "Image Extraction Apparatus, Image Extraction Method and Image Extraction Program and Recording Medium storing the Program". Technical Field

[0002] This invention relates to an image extraction apparatus, an image extraction method, an image extraction program, and a recording medium storing the program. Background Technology

[0003] Many images are taken at events such as weddings. From these numerous images, high-quality images are selected for purposes such as creating a photo album. Therefore, devices that automatically select suitable images for the album (Patent Document 1), devices that specify the image state for the image insertion area in the album and insert images matching the specified state into the image insertion area (Patent Document 2), etc., could also be considered. Furthermore, devices that notify the user of failed images before printing (Patent Document 3), devices that perform focus confirmation by magnifying a specified area of ​​the displayed image (Patent Document 4), etc., could also be considered.

[0004] Furthermore, devices that allow selecting images while confirming the layout of a photo album (Patent Document 5), devices that effectively reproduce a large number of images (Patent Document 6), devices that allow easy searching of photos (Patent Document 7), and devices that allow easy switching between similar image screens and screens containing dissimilar images (Patent Document 8) are also possible.

[0005] Patent Document 1: Japanese Patent No. 5655112

[0006] Patent Document 2: Japanese Patent No. 4455302

[0007] Patent Document 3: Japanese Patent Publication No. 2002-187329

[0008] Patent Document 4: Japanese Patent No. 4346859

[0009] Patent Document 5: Japanese Patent Publication No. 2017-010247

[0010] Patent Document 6: Japanese Patent Publication No. 2013-246555

[0011] Patent Document 7: Japanese Patent Publication No. 2017-054552

[0012] Patent Document 8: Japanese Patent Publication No. 2011-257980

[0013] The apparatus described in Patent Document 1 selects images with high evaluation values, thus resulting in the selection of multiple similar images simply because an image has a high evaluation value. The apparatus described in Patent Document 2 inserts images that match a specified state into an image insertion area, which also results in multiple similar images being inserted into the album. The apparatus described in Patent Document 3 notifies the user of failed images before printing, and the apparatus described in Patent Document 4 performs focus confirmation by magnifying a specified area of ​​the displayed image; therefore, neither of these methods takes into account multiple similar images.

[0014] Furthermore, the device described in Patent Document 5 selects an image, sets similarity conditions from the selected images, and displays candidate images that meet the similarity conditions in the display area. The device described in Patent Document 5 cannot set similarity conditions if no image is selected, therefore each image must be checked when there are multiple images. The device described in Patent Document 6 displays similarity information indicating whether two consecutively displayed images are similar, therefore it does not consider what kind of images are considered similar. The device described in Patent Document 7, for example, displays images associated with "rose" by inputting the keyword "rose," but it does not consider conditions for judging image similarity. Moreover, the device described in Patent Document 8 is a device that displays images of dogs taken in real-time horizontally on the screen, similar to images in a live view, but it does not consider conditions for judging image similarity. Therefore, it is sometimes impossible to find similar images that meet the user's expectations. Summary of the Invention

[0015] The purpose of this invention is to be able to find similar images that match the user's preferences.

[0016] The image extraction apparatus based on the present invention is characterized by comprising: a similarity judgment condition receiving mechanism, which receives input of a similarity judgment condition for determining whether any two images are similar to each other when comparing them among a plurality of images; a similar image group determining mechanism, which determines one or more similar image groups consisting of two or more similar images according to the similarity judgment condition received by the similarity judgment condition receiving mechanism; and an image extraction mechanism, which extracts at least one image from each of the similar image groups determined by the similar image group determining mechanism.

[0017] The image extraction apparatus based on the present invention may include: a similarity judgment condition receiving device, which receives input of a similarity judgment condition for judging whether any two images are similar to each other when comparing them among a plurality of images; and a processor, which, according to the similarity judgment condition received by the similarity judgment condition receiving device, determines one or more similar image groups consisting of two or more similar images, and extracts at least one image from each of the determined one or more similar image groups.

[0018] The present invention also provides an image extraction method suitable for an image extraction apparatus. Specifically, the method is characterized in that a similarity judgment condition receiving mechanism receives input of a similarity judgment condition for determining whether any two images are similar to each other when comparing them among multiple images; a similar image group determining mechanism determines one or more similar image groups consisting of two or more similar images according to the similarity judgment condition received by the similarity judgment condition receiving mechanism; and an image extraction mechanism extracts at least one image from each of the similar image groups determined by the similar image group determining mechanism.

[0019] The image extraction device may also include: a first image display control mechanism that displays multiple images in a first display window formed on the display screen of the display device, and distinguishes between images included in one or more similar image groups determined by the similar image group determination mechanism and images not included in the similar image groups.

[0020] The first image display control mechanism preferably distinguishes one similar image group from other similar image groups among multiple similar image groups.

[0021] The image extraction apparatus may further include: a similar image group designation mechanism, which designates one of one or more similar image groups determined by the similar image group determination mechanism; and a similar image display control mechanism, which displays adjacent images contained in the similar image group designated by the similar image group designation mechanism in a second display window different from the first display window. Adjacent display of similar images means that similar images (photographic images) are displayed continuously without inserting dissimilar images between them. If no dissimilar images are inserted between these images, decorative elements such as text or markings can be inserted between the similar images. The images can be arranged horizontally or vertically. Furthermore, the first and second display windows can be displayed without overlapping, or with a portion of the second display window overlapping a portion of the first display window from above, or with the second display window contained within the first display window and overlapping the first display window.

[0022] The image extraction apparatus may also include a similarity judgment condition specification window display control mechanism, which displays a similarity judgment condition specification window containing multiple similarity judgment conditions on the display screen of the display device. In this case, the similarity judgment condition receiving mechanism receives the input of similar image judgment conditions, for example, by specifying one of the multiple similarity judgment conditions displayed in the similarity judgment condition specification window.

[0023] The image extraction apparatus may further include: a similar image group designation mechanism, which designates one of one or more similar image groups determined by the similar image group determination mechanism; a compliance condition receiving mechanism, which receives input of compliance conditions for the image; a compliance degree calculation mechanism, which calculates the compliance degree of the designated image among the similar images contained in the similar image group designated by the similar image group designation mechanism, under the compliance conditions received by the compliance condition receiving mechanism; and a compliance degree notification mechanism, which notifies the compliance degree calculated by the compliance degree calculation mechanism.

[0024] The image extraction apparatus may also include a condition-specifying window display control mechanism, which displays a condition-specifying window containing multiple matching conditions on the display screen of the display device. In this case, the condition-receiving mechanism receives, for example, input of a matching condition specified from among the multiple matching conditions displayed in the condition-specifying window.

[0025] In a group of similar images, an image contained in one group of similar images is not similar to an image contained in another group of similar images, for example, based on the similarity judgment condition received by the receiving agency.

[0026] The image extraction apparatus may also include a similarity judgment condition input window display control mechanism, which displays the similarity judgment condition input window on the display screen of the display device. In this case, the similarity judgment condition receiving mechanism receives, for example, the similarity judgment conditions entered in the similarity judgment condition input window.

[0027] The image extraction apparatus may also include a condition-compliant input window display control mechanism, which displays the condition-compliant input window on the display screen of the display device. In this case, a condition-compliant receiving mechanism, for example, receives the condition-compliant input entered in the condition-compliant input window.

[0028] The image extraction device may further include: a similar image group designation mechanism, which designates one of one or more similar image groups determined by the similar image group determination mechanism; a conformity calculation mechanism, which calculates the conformity degree of the images contained in the similar image group designated by the similar image group designation mechanism, which is received by the conformity condition receiving mechanism; and a recommended image notification mechanism, which notifies the image with the highest conformity degree calculated by the conformity calculation mechanism as the recommended image extracted by the image extraction mechanism.

[0029] Invention Effects

[0030] This system receives input from similarity criteria used to determine whether two images are similar to each other when comparing them. Based on these criteria, it determines one or more similar image groups consisting of two or more similar images. From each determined similar image group, at least one image is extracted. The user can determine the similarity criteria and identify similar image groups accordingly. The user can determine the following similarity criteria: when performing similarity based solely on image comparison, images judged as dissimilar are judged as similar; conversely, when performing similarity based solely on image comparison, images judged as similar are judged as dissimilar. It can find similar images that meet the user's desired criteria. Because similar image groups are determined, even if multiple images exist, it is relatively easy to identify which image is similar. It can extract at least one image from each similar image group. Attached Figure Description

[0031] Figure 1 This is a block diagram representing the electrical structure of an image extraction device.

[0032] Figure 2 This is a flowchart illustrating the processing sequence of the image extraction device.

[0033] Figure 3 This is a flowchart illustrating the processing sequence of the image extraction device.

[0034] Figure 4 This is a flowchart illustrating the processing sequence of the image extraction device.

[0035] Figure 5 This is a flowchart illustrating the processing sequence of the image extraction device.

[0036] Figure 6 This is a flowchart illustrating the processing sequence of the image extraction device.

[0037] Figure 7 This is an example of an image overview window.

[0038] Figure 8 This is an example of a window specified by similarity judgment conditions.

[0039] Figure 9 This is an example of adding a window to the similarity judgment condition.

[0040] Figure 10 This is an example of a similarity judgment condition input window.

[0041] Figure 11 This is an example of an editing window.

[0042] Figure 12 This is an example of an image confirmation window.

[0043] Figure 13 This is an example of a window that meets the specified conditions.

[0044] Figure 14 This is an example of a conditional input window.

[0045] Figure 15 This is an example of an image confirmation window.

[0046] Figure 16 This is an example of an editing window.

[0047] Figure 17 This is a flowchart illustrating an example of the processing sequence of an image extraction device.

[0048] Figure 18 This is an example of an editing window.

[0049] Figure 19 This is an example of an image confirmation window.

[0050] Symbol Explanation

[0051] 1-Image extraction device, 2-CPU, 3-Display device, 4-Communication device, 5-Hard disk, 6-Hard disk drive, 7-Memory, 8-Keyboard, 9-Mouse, 10-Optical disc drive, 11-Optical disc, 12-Memory card reader, 13-Memory card, 50-Image overview window, 50A-Editing window, 51-Image overview display area, 52-Slider, 53-Similarity judgment condition button, 54-Page image overview display area, 55-End button, 60-Similarity judgment condition specification window, 61-65-Check boxes, 66-Similarity judgment condition append button, 67-OK button, 70-Similarity judgment condition append window, 71-72-Check boxes, 76-Similarity judgment condition creation button, 77-Back button, 80-Similarity judgment condition input window, 81-Similarity judgment condition input area, 82-OK button, 83-Back button, 90-Image confirmation window, 91-Similar image display area, 92-Image Confirmation area, 93 - Meets criteria button, 94 - OK button, 95 - Compliance display area, 100 - Meets criteria specification window, 101-104 - Checkboxes, 105-107 - Slider, 108-110 - Checkboxes, 112 - Meets criteria input button, 113 - OK button, 120 - Meets criteria input window, 121 - Meets criteria input area, 122 - Meets criteria input area, 123 - Meets criteria input area, 124 - Back button Buttons, A1~A18 - Image pasting area, A11~A16 - Image pasting area, A21~A25 - Image pasting area, A31~A36 - Image pasting area, A41~A46 - Image pasting area, G1 - 1st similar image group, G2 - 2nd similar image group, G3 - 3rd similar image group, G4 - 4th similar image group, G5 - 5th similar image group, G6 - 6th similar image group, I1~I28 - Images, P1~P5 - Page images. Detailed Implementation

[0052] (First Embodiment)

[0053] Figure 1 This is a diagram illustrating an embodiment of the present invention, and a block diagram illustrating the electrical structure of the image extraction device 1.

[0054] The overall operation of the image extraction device 1 is centrally controlled by the CPU (Central Processing Unit) 2.

[0055] The image extraction device 1 includes a display device 3 for displaying images and other information on a display screen, and a communication device 4 for communicating with devices other than the image extraction device 1 by connecting to the Internet and other networks. Furthermore, the image extraction device 1 also includes a hard disk 5, a hard disk drive 6 for accessing the hard disk 5, a memory 7 for storing data, a keyboard 8 for inputting commands, and a mouse 9. Additionally, the image extraction device 1 includes an optical disk drive 10 for accessing an optical disk 11, and a memory card reader 12 for writing data to and reading data stored in a memory card 13.

[0056] The operation program of the image extraction device 1, described later, is received by the communication device 4 via the Internet. The received operation program is installed in the image extraction device 1. Alternatively, the operation program may not be received and installed in the image extraction device 1 via a network such as the Internet, but rather recorded on a portable recording medium such as an optical disc 11 and read from that portable recording medium. In this case, the operation program read from the portable recording medium is installed in the image extraction device 1. It goes without saying that the computer (CPU 2) of the image extraction device 1 is capable of reading the operation program.

[0057] Figures 2 to 6 This is a flowchart illustrating the processing sequence of the image extraction device 1.

[0058] In this embodiment, similar images are identified from a plurality of images, and the user is informed which of the images is considered similar. Specifically, the system receives input of similarity criteria used to determine whether two images are similar to each other when comparing them among a plurality of images, and identifies similar images according to the received similarity criteria. The user can modify or create the similarity criteria, thus allowing the system to determine which images are considered similar according to the user's preferences.

[0059] Image data representing multiple images is stored on memory card 13, and the image data stored on memory card 13 is read by memory card reader 12. Figure 2 Step 21). A list of most (or most) images represented by the read image data is displayed on the display screen of display device 3. Figure 2 Step 22).

[0060] Figure 7 This is an example of an image overview window 50 displayed on the display screen of the display device 3.

[0061] An image overview display area 51 is formed in the image overview window 50. Multiple images represented by image data read from the memory card 13 are displayed in the image overview display area 51. A slider 52 is formed on the right side of the image overview display area 51, which can be freely slid up and down according to the user's command (drag action). A similarity judgment condition button 53, labeled with the text "Similarity Judgment Condition", is formed by clicking in the lower right corner of the image overview display area 51.

[0062] Images I1 to I28 (with regard to images I22 to I28 being a portion of the image) are displayed in the image overview display area 51. By moving the slider 52 downwards, images represented by image data read from the memory card 13 that are not displayed in the image overview display area 51 are displayed in the image overview display area 51.

[0063] If the user clicks the similarity judgment condition button 53 using the mouse (9), Figure 2 If "Yes" is selected in step 23, then the display will be shown on the screen of the display device 3 under the control of CPU2 (an example of a window display control mechanism specified by a similar judgment condition). Figure 8 The similarity judgment condition specification window 60 is shown. The similarity judgment condition specification window 60 can be displayed overlappingly with the image overview window 50, overlapping a portion of the image overview window 50, or not overlapping the image overview window 50.

[0064] refer to Figure 8 Five checkboxes 61 to 65 are formed in the similarity judgment condition specification window 60. These checkboxes 61 to 65 can be selected by using the mouse 9. By selecting any one or all of the checkboxes 61 to 65, a similarity judgment condition is specified from multiple similarity judgment conditions, thereby the image extraction device 1 receives the input of the similarity judgment condition.

[0065] Pre-defined similarity criteria are displayed corresponding to checkboxes 61 to 65. To the right of the first checkbox 61, the first similarity criterion is displayed: "Only images of the same person are considered similar." Regarding this first similarity criterion, when comparing two images, they are considered similar if the matching degree between the two images containing only images of the same person is above a threshold (therefore, if one image contains images of someone other than the same person, it will not be considered similar). Normally, when matching two images, the similarity criterion is based on a matching degree above a threshold. However, in this embodiment, different conditions can be added to this criterion to make it either more stringent or more lenient. The first similarity criterion is that even with a high matching degree, if the two images do not contain the same person, they will not be considered similar (equivalent to a stricter similarity criterion. The following four similarity criterions are considered similar).

[0066] To the right of the second checkbox 62, there is a second similarity judgment condition: "Only set images with the same expression (smile) as similar." Regarding the second similarity judgment condition, two images are judged to be similar when the matching degree of the two images is above the threshold and the expressions of the people contained in the two images are determined to be the same (the consistency of the expressions is above the threshold). (Therefore, if the expressions of the people contained in the two images are determined to be different, even if the matching degree of the two images is above the threshold, they will not be judged to be similar.)

[0067] To the right of the third checkbox 63, there is a third similarity judgment condition: "Only images with the same size of human subjects are considered similar." Regarding the third similarity judgment condition, only images with a matching degree of more than the threshold and whose human subjects (the size ratio of the human subject to the overall image size) are the same (within a certain range) are judged as similar. (Therefore, if the human subjects in two images are not the same size, even if the matching degree of the two images is more than the threshold, they will not be judged as similar.)

[0068] To the right of checkbox 64, there is a fourth similarity criterion: "Only images captured in consecutive time periods are considered similar." This fourth similarity criterion defines two images as similar if their matching degree is above a threshold and if they were captured in consecutive time periods. These two images are considered similar only if there are no other images captured between them. (Therefore, even if the matching degree is above the threshold, images compared to each other will not be considered similar if there are other images captured between them.)

[0069] To the right of checkbox 65, the fifth similarity criterion is displayed: "Set the vertical image and the horizontal image as similar." In this embodiment, when the fifth similarity criterion is not applied, the matching degree between the vertical image (image whose vertical length is longer than its horizontal length) and the horizontal image (image whose horizontal length is longer than its vertical length) is not calculated (therefore, they are judged as dissimilar). However, when the fifth similarity criterion is applied, the matching degree of the two images is calculated. If the matching degree is above a threshold, they are judged as similar images (the matching degree of the vertical image and the horizontal image can be calculated by comparing the square portion after cropping the center, aligning the aspect ratio to one side, etc., after performing various operations on one or two images. Furthermore, when there are square images mixed in, it is better to check the fifth checkbox). When the fifth similarity criterion is applied, compared to when the fifth similarity criterion is not applied, more images may be judged as similar (the similarity criterion becomes more lenient).

[0070] When two or more checkboxes are selected, images that all meet their respective similarity criteria will be extracted as similar images (judged as AND conditions). Alternatively, when two or more checkboxes are selected, images that meet either condition will be extracted as similar images (judged as OR conditions). Furthermore, a GUI (Graphical User Interface) can be set up that allows input of logical expressions, where the user can input whether the conditions shown in the two or more checkboxes are judged as AND or OR conditions.

[0071] At the bottom of the similarity judgment condition specification window 60, there are two clickable buttons: a "Add Similarity Judgment Condition" button 66, labeled "Add Similarity Judgment Condition," and an "OK" button 67. The "Add Similarity Judgment Condition" button 66 is clicked by the user when adding conditions other than those displayed in the similarity judgment condition specification window 60 as similarity judgment conditions. The "OK" button 67 is clicked by the user when the similarity judgment condition is confirmed.

[0072] When adding similarity judgment conditions ( Figure 2 In step 25 ("Yes"), use mouse 9 to click the similar judgment condition append button 66. Then, the display on the display device 3 will show... Figure 9 The similarity judgment condition is added to window 70 (shown). Figure 3 Step 26).

[0073] refer to Figure 9The similarity judgment condition addition window 70 also contains five checkboxes 71 to 75. To the right of checkbox 71, an additional similarity judgment condition is displayed: "Only set images with the same hue as similar." Regarding this additional similarity judgment condition, images with a matching degree above a threshold and the same hue intensity (same in terms of above-threshold aspects) are judged as similar. Even if the matching degree of two images is above the threshold, images with different hues, such as black and white images and color images, will not be judged as similar. Additional similarity judgment conditions are also displayed to the right of the other checkboxes 72 to 75.

[0074] By checking checkboxes 71 to 75, you can add additional specified options. Figure 8 The similarity judgment criteria shown in window 60 are used to specify similarity judgment criteria other than the standard similarity judgment criteria.

[0075] At the bottom of the similarity judgment condition addition window 70, there is a similar judgment condition creation button 76 with the text "Similar Judgment Condition Creation" and a return button 77 with the text "Return".

[0076] Figure 8 The similarity judgment conditions shown in window 60 are the similarity judgment conditions and Figure 9 The similarity judgment condition addition window 70 shows the added similarity judgment conditions that are preset in the image extraction device 1. Similarity judgment conditions are specified from the preset similarity judgment conditions and the added similarity judgment conditions. However, the user can create similarity judgment conditions other than these. When the user creates similarity judgment conditions, the similarity judgment condition creation button 76 is clicked. The return button 77 is used when... Figure 9 The similarity judgment condition append window 70 shown returns to Figure 8 The similarity judgment condition shown specifies that window 6() was clicked by the user.

[0077] When creating similarity judgment conditions ( Figure 3 In step 27 ("Yes"), as described above, click the similarity judgment condition creation button 76. Then, the CPU2 (an example of a similarity judgment condition input window display control mechanism) displays the result on the display screen of the display device 3. Figure 10 The similarity judgment condition input window shown is 80 ( Figure 3 Step 28).

[0078] refer to Figure 10 A similarity judgment condition input area 81 is formed in the similarity judgment condition input window 80. The user creates similarity judgment conditions and inputs them into the similarity judgment condition input window 80 using the keyboard 8. The image extraction device 1 then receives the input similarity judgment conditions.

[0079] The image extraction device 1 includes a similarity judgment condition analysis mechanism (not shown). This mechanism performs lexical analysis on the natural language input by the user and generates similarity judgment conditions that can be processed by the image extraction device 1. Alternatively, the similarity judgment condition analysis mechanism can also apply a pre-learned learning model to the natural language input by the user to generate similarity judgment conditions that can be processed by the image extraction device 1.

[0080] Below the similarity judgment condition input area 81, there are a "OK" button 82 and a "Back" button 83. If the user clicks the "OK" button 82, the condition entered in the similarity judgment condition input area 81 will be used as the similarity judgment condition. If the user clicks the "Back" button 83... Figure 3 If "Yes" is selected in step 29, the similarity judgment condition input window 80 disappears from the display screen. The display screen then shows... Figure 8 The similarity judgment criteria specification window 60 is shown. If you click the OK button 67 included in the similarity judgment criteria specification window 60... Figure 4 If "yes" is selected in step 30, then the CPU2 determines the similar image group according to the similarity judgment conditions. Figure 4 Step 31). (The display shows...) Figure 9 In the state of the similarity judgment condition append window 70 shown, when the return button 77 contained in the similarity judgment condition append window 70 is clicked ( Figure 3 In step 27, "No" is selected. Figure 3 In step 29, if "Yes" is selected, the similar judgment condition append window 70 also disappears from the display screen, and the display screen shows... Figure 8 The similarity judgment criteria specification window 60 is shown. If you click the OK button 67 included in the similarity judgment criteria specification window 60... Figure 4 If "yes" is selected in step 30, then the CPU2 determines the similar image group according to the similarity judgment conditions. Figure 4 Step 31). (The display shows...) Figure 8 In the state of the window 60 showing the similarity judgment conditions, when the OK button 67 is clicked instead of the similarity judgment condition append button 66 ( Figure 2 In step 25, "No" is selected. Figure 4 In step 30, "Yes" is also used by CPU2 to determine similar image groups according to similarity judgment conditions. Figure 4(In step 30, "Yes"). Among the images represented by the image data read from memory card 13, a specific image is identified, and the CPU2 calculates the matching degree between all images represented by the image data read from memory card 13 and the specific image. For images with a calculated matching degree of more than a threshold, the CPU2 determines whether they meet the similarity judgment criteria based on their relationship with the identified specific image, and as a result, whether they are images similar to the specific image is determined.

[0081] The similarity judgment condition specification window 60, similarity judgment condition addition window 70, and similarity judgment condition input window 80 displayed on the screen of the keyboard 8, mouse 9, and display device 3 serve as a similarity judgment condition receiving mechanism that receives inputs determining whether any two images among multiple images are similar to each other when comparing them. Furthermore, the CPU 2 serves as a similarity image group determining mechanism that determines one or more similar image groups consisting of two or more similar images according to the received similarity judgment conditions.

[0082] As a similarity judgment condition, let it be specified. Figure 8 The similarity judgment criteria shown in window 60 specify the first similarity judgment criterion: "Only images of the same person are considered similar." Therefore, images containing the same person are identified as a similar image group among images whose matching degree is above a certain threshold. For example, in... Figure 7 In the image overview display area 51 of the image overview window 50 shown, the CPU2 determines that images I1, I2, I3, and I11 are similar images and identifies them as the first similar image group G1 containing these images I1, I2, I3, and I11. Similarly, in Figure 7 In the image overview display area 51 of the image overview window 50 shown, image I6 and image I13 are judged to be similar and are identified as the second similar image group G2. Similarly, images I8, I9, I10 and I12 are judged to be similar and are identified as the third similar image group G3, images I16, I17 and I18 are judged to be similar and are identified as the fourth similar image group G4, images I19 and I20 are judged to be similar and are identified as the fifth similar image group G5, and images I22 and I23 are judged to be similar and are identified as the sixth similar image group G6.

[0083] If a similar image group is identified, the image overview window 50 displayed on the display screen of display device 3 disappears from the display screen. Figure 11 The editing window 50A shown is displayed on the display screen of the display device 3.

[0084] Figure 11 This is an example of an editing window, 50A.

[0085] The editing window 50A includes an image overview display area 51, which is identical to the image overview window 50. As described above, this image overview display area 51 displays a plurality of images represented by image data read from the memory card 13.

[0086] Furthermore, a thumbnail page image overview display area 54 for editing is formed below the image overview display area 51. The thumbnail page image overview display area 54 displays thumbnail page images P1 to P5 representing the page layouts that constitute the electronic photo album. In thumbnail page image P1, image pasting areas A1 to A8 formed on the first page of a double-page spread of the electronic photo album are defined. Similarly, in thumbnail page image P2, image pasting areas A11 to A16 formed on the second page of a double-page spread of the electronic photo album are defined. In thumbnail page images P3, P4, and P5, image pasting areas A21 to A25, A31 to A35, and A41 to A46 formed on the third, fourth, and fifth pages of a double-page spread of the electronic photo album, respectively, are formed. An "End" button 55 is formed on the right side of the thumbnail page image overview display area 54.

[0087] As described above, in the image overview display area 51 of the editing window 50A (the first display window), the first similar image group G1, which includes similar images I1, I2, I3, and I11, is surrounded by a dotted line. This distinguishes the images I1, I2, I3, and I11 included in the first similar image group G1 from other images not included in the first similar image group G1. Similarly, the second similar image group G2, which includes similar images I6 and I13, is surrounded by a dotted line; the third similar image group G3, which includes similar images I8, I9, I10, and I12, is surrounded by a dotted line; the fourth similar image group G4, which includes similar images I16, I17, and I18, is surrounded by a dotted line; the fifth similar image group G5, which includes similar images I19 and I20, is surrounded by a dotted line; and the sixth similar image group G6, which includes similar images I22 and I23, is surrounded by a dotted line. In this way, images that are similar to each other are displayed separately from other images, allowing users to determine whether images among multiple images are similar. Moreover, the similarity criteria are specified by the user, thus enabling images that meet the user's desired criteria to be identified as similar.

[0088] In similar image groups G1 to G5, images contained in one similar image group are dissimilar to images contained in other similar image groups under the same similarity judgment conditions. For example, images I1, I2, I3, and I11 contained in similar image group G1 are dissimilar to images I8, I9, I10, and I12 contained in similar image group G3.

[0089] exist Figure 11 In the example, similar image groups G1 to G5 are all surrounded by dashed lines. However, within similar image groups G1 to G5, it's also possible to differentiate one similar image group from others. For example, within similar image groups G1 to G5, the way the box surrounding each similar image group is changed can be configured to make it easier to distinguish between similar image groups. Furthermore, as long as it's possible to distinguish between images contained within a similar image group and images not contained within it, it's also possible to configure it so that instead of using a box to surround each similar image group, each similar image group is highlighted by changing its color, etc.

[0090] exist Figure 11 In the example, if the similarity judgment condition is specified as "only images with consecutive shooting times are set as similar" (see reference), Figure 8 If images I3 and I11, I6 and I13, and I10 and I12 are not consecutively captured, they will not be considered similar images. As a result, image I11 is excluded from the first similar image group G1, and the similar images contained in G1 become images I1, I2, and I3. Furthermore, image I6 and image I13 contained in the second similar image group G2 are considered dissimilar. Moreover, image I12 is excluded from the third similar image group G3, and the similar images contained in G3 become images I8, I9, and I10.

[0091] Furthermore, for example, if no similarity judgment condition is specified (see reference), such as "only images of the same person are considered similar" (see reference). Figure 8 If the image I21 is similar to the image I19 or the image I20, then the fifth similar image group G5 contains the image I21 in addition to the images I19 and I20.

[0092] In the similar image groups displayed in the image overview display area 51, if you use mouse 9 (an example of specifying the organization of similar image groups) to drag and drop any similar image group onto any image paste area in images P1 to P5 on the thumbnail page ( Figure 4 If "Yes" is selected in step 33, then the display screen on display device 3 will show... Figure 12 The image confirmation window 90 shown is a second display window, different from the first display window. Figure 5 Step 35). If a similar image group is not dragged and dropped into the image pasting area formed by any of the thumbnail page images P1 to P5, but an image not included in the similar image group (referred to as a separate image) is dragged and dropped into the image pasting area formed by any of the thumbnail page images P1 to P5 (…), then… Figure 4 In step 33, "No" is used. Figure 4If "Yes" is selected in step 34, further processing is performed. In this embodiment, the third similar image group G3 is dragged and dropped onto the image pasting area A1 of the thumbnail page image P1. The mouse 9 corresponds to the designated similar image group of one or more similar image groups.

[0093] refer to Figure 12 The image confirmation window 90 includes a similar image display area 91. This similar image display area 91 displays similar images I8, I9, I10, and I12 contained in the dragged and dropped similar image group G3. The similar images I8, I9, I10, and I12 are displayed adjacently via CPU2 (an example of a similar image display control mechanism). No dissimilar images are displayed between the similar images I8, I9, I10, and I12, and the images are displayed closely together. Furthermore, no text or other materials besides images are displayed between the images. However, the images do not necessarily have to be closely connected; text or other materials can be displayed between images. Even if dissimilar images are included between images, it is sufficient that the dissimilar images and similar images can be distinguished. Similar images can be displayed brightly, while dissimilar images can be displayed darkly. For example, images I8, I9, I10, I11, and I12 can be displayed in the similar image display area 91, with the similar images I8, I9, I10, and I12 displayed brightly, and the dissimilar image I11 displayed darkly.

[0094] Below the similar image display area 91, there is an image confirmation area 92. The image confirmation area 92 displays the image clicked by the user among the similar images displayed in the similar image display area 91 (initially a predetermined image, such as the image initially captured among the images displayed in the similar image display area 91).

[0095] A "Meets Criteria" button 93 is formed to the right of the similar image display area 91. A "OK" button 94 is formed to the right of the image confirmation area 92.

[0096] Can Figure 12 The size of the image confirmation window shown is set to the same as the image size shown. Figure 11 The editing window 50A shown is the same size as, and is in line with... Figure 11 The image confirmation window 90 can be displayed on the editing window 50A in the same manner as shown, or it can be displayed in a manner that overlaps with a part of the editing window 50A, or it can be displayed in a manner that does not overlap with the editing window 50A.

[0097] If the user clicks the button that meets the conditions using the mouse (9) Figure 5If "Yes" is selected in step 36, then the CPU2 (an example of a window display control mechanism that specifies the condition) will display the information on the display screen of the display device 3. Figure 13 The specified window 100 shows the conditions. Figure 5 Step 37).

[0098] Figure 13 This is an example of a window that meets the specified conditions (window 100).

[0099] The criteria for meeting the requirements are those used to determine the completeness of an image. These criteria can be included within similarity criteria. The criteria specification window 100 (an example of a criteria receiving agency) contains four checkboxes 101 to 104. The criteria corresponding to these four checkboxes 101 to 104 are displayed. Multiple criteria are displayed in the criteria specification window 100.

[0100] The word "Closed Eyes" is displayed to the right of the first checkbox 101. By checking the first checkbox 101, the degree to which the eyes of the people in the image are closed can be specified as a compliance condition. To the right of the word "Closed Eyes" is a ruler with the words "Complete" on the left and "Slightly" on the right. On this ruler, the slider 105 can be moved left and right using the mouse 9. Moving the slider 105 further to the left increases the compliance of images with no closed eyes, while moving the slider 105 further to the right increases the compliance of images with closed eyes, even if some closed eyes are present. By checking the checkbox 101 and moving the slider 105, the effect of closed eyes on the compliance can be adjusted.

[0101] The word "Blurred" is displayed to the right of the second checkbox 102. By checking the second checkbox 102, the degree of image blur can be specified as a compliance condition. To the right of the word "Blurred" is a ruler with the words "Severe" on the left and "Slight" on the right. On this ruler, the slider 106 can be moved left and right using the mouse 9. Moving the slider 106 further to the left increases the compliance with less blurry images, while moving the slider 106 further to the right increases the compliance with blur, even if some blur is present. By checking the checkbox 102 and moving the slider 106, the effect of blur relative to the compliance can be adjusted.

[0102] The word "Exposure" is displayed to the right of the third checkbox 103. By checking the third checkbox 103, the brightness level of the image can be specified as a compliance condition. To the right of the word "Exposure" is a ruler with the words "Bright" on the left and "Dark" on the right. On this ruler, the slider 107 can be moved left and right using the mouse 9. Moving the slider 107 further to the left results in a brighter image and a higher compliance; moving the slider 107 further to the right results in a darker image and a higher compliance. By checking the checkbox 103 and moving the slider 107, the effect of exposure relative to compliance can be adjusted.

[0103] The word "Expression" is displayed to the right of checkbox 4 104. By checking checkbox 4 104, the facial expression of the person in the image can be specified as a matching condition. To the right of the word "Expression" are the words "Laugh," "Smile," "Serious," and "Crying." Below each of the words "Laugh," "Smile," "Serious," and "Crying" are checkboxes 108, 109, 110, and 111. The closer the expression is to the expression corresponding to the checked checkbox in checkboxes 108, 109, 110, and 111, the higher the degree of facial expression matching. By checking checkbox 104 and any of the checkboxes 108, 109, 110, and 111, the influence of facial expression on the degree of matching can be adjusted.

[0104] At the bottom of the specified window 100, there is a "Compliant Input" button 112 labeled "Compliant Input" and an "OK" button 113 labeled "OK".

[0105] If you click the input button 112 (which meets the conditions) Figure 5 If "Yes" is selected in step 38, then the CPU2 (an example of a conditional input window display control mechanism) will display the result on the display screen of the display device 3. Figure 14 The matching input window 120 shown is an example of a matching receiving agency. Figure 5 Step 39).

[0106] Figure 14 This is an example of a window that meets the conditions for inputting window 120.

[0107] The conditional input window 120 contains three conditional input areas 121, 122, and 123. A "Back" button 124 labeled "Back" is located in the lower right corner of the conditional input window 120.

[0108] If a matching condition is entered using keyboard 8 in at least one of the matching condition input areas 121, 122, and 123, then the matching condition becomes user-specified. For example, if "resolution" is entered in any of the matching condition input areas 121, 122, or 123, then resolution becomes a matching condition; the higher the resolution, the stronger the matching. Figure 13 As shown, it can also be configured to display a slider, and the slider can be used to adjust the influence of the degree of compliance of the input compliance conditions. The image extraction device 1 receives the input of compliance conditions.

[0109] If you click the back button 124 Figure 5 If "Yes" is selected in step 40, then... Figure 14 The matching input window 120 shown disappears from the display screen of the display device 3. Figure 13 The specified window 100, which meets the conditions, is displayed on the display screen of the display device 3. Figure 5 Step 37). If you click the OK button 113 contained in the specified window 100 that meets the conditions ( Figure 5 In step 38, "No" is used. Figure 5 If "Yes" is selected in step 41, then CPU2 (an example of a compliance calculation mechanism) calculates the compliance according to the specified compliance conditions. Figure 12 The conformity of the image displayed in the image confirmation area 92 of the image confirmation window 90 shown ( Figure 6 Step 42). If the conformity is calculated, the image confirmation window 90 updates the display screen on the display device 3 via CPU2. Figure 6 Step 47).

[0110] Figure 15 This is an example of an updated image confirmation window 90.

[0111] A compliance display area 95 is displayed in the image confirmation window 90. In this embodiment, "closed eyes," "blurred," and "expression (smile)" are specified as compliance conditions. The compliance display area 95 displays the specified compliance conditions "closed eyes," "blurred," and "expression (smile)," and evaluations A, B, and C are displayed to the right of each compliance condition. Regarding the compliance condition of "closed eyes," A is surrounded by a circle, indicating that the compliance condition of image 18 displayed in the image confirmation area 92 is rated A for "closed eyes." Similarly, regarding the compliance condition of "blurred," B is surrounded by a circle, indicating that the compliance condition of image 18 is rated B for "blurred." Regarding the compliance condition of "expression (smile)," A is surrounded by a circle, indicating that the compliance condition of image 18 is rated A for "expression (smile)."

[0112] If, in the similar image display area 91, among the similar images I8, I9, I10, and I11, you use mouse 9 to click on an image other than image I8 displayed in the image confirmation area 92 (…). Figure 6 If "Yes" is selected in step 44, then CPU2 (an example of a compliance calculation mechanism) calculates the compliance score for the clicked image based on the specified compliance criteria. Figure 6 Step 42). The image confirmation window 90 is updated, and the calculated compliance score is displayed in the compliance score display area 95 (an example of a compliance score notification agency). Figure 6 Step 47).

[0113] If you click the OK button 94 ( Figure 6 If "Yes" is selected in step 45, the image confirmation window 90 will disappear from the display screen of the display device 3. Figure 16 As shown, the editing window 50A is displayed on the display screen of the display device 3. Figure 6 Step 46).

[0114] refer to Figure 16 In the editing window 50A, the image displayed in the image confirmation area 92 is pasted into the image pasting area A1 where the similar image group G3 is dragged and dropped by the CPU2. This allows the CPU2 (an example of a similar image extraction mechanism) to extract at least one image from the similar image group G3.

[0115] Repeat until the end button 55 in the editing window 50A is clicked. Figure 4 Step 33 to Figure 6 Step 46 processing ( Figure 6 Step 47).

[0116] According to the first embodiment, the user can specify similarity judgment conditions, and thus find similar images that correspond to the user's wishes.

[0117] (Second Embodiment)

[0118] Figures 17 to 19 This refers to the second embodiment.

[0119] In the second embodiment, among the images contained in the specified similar image group, the image with the highest degree of similarity is notified to the user as the recommended image.

[0120] Figure 17 Indicates and Figure 4 The processing order corresponding to steps 33 and 34 is a flowchart representing part of the processing order of the image extraction device 1.

[0121] For reference Figures 2 to 6 The flowchart illustrates that, if Figure 16The editing window 50A shown is updated as long as the end button 55 is not pressed. Figure 6 Step 47), then proceed to Figure 17 Step 131 involves CPU2 calculating the similarity score of the images contained in each similar image group (this similarity score is based on the similarity criteria specified or created by the user). Figure 17 Step 131). Therefore, for each group of similar images, the image with the highest similarity within that group is displayed relatively brighter compared to the other images in the group. Figure 17 Step 132).

[0122] Figure 18 This is an example of an editing window, 50A.

[0123] The first similar image group G1 contains similar images I1, I2, I3, and I11, and it is assumed that among these images I1, I2, I3, and I11, image I2 has a relatively higher similarity than the other images I1, I3, and I11. Therefore, image I2 is displayed relatively brighter than the other images I1, I3, and I11. Figure 18 In the example shown, images I1, I3, and I11 are shaded, while image I2 is not shaded, indicating that image I2 is relatively brighter than images I1, I3, and I11. Thus, by distinguishing the image I2, which has a relatively high similarity, from the image with a relatively low similarity, it can be seen that the image with the high similarity is the one contained in the similar image group.

[0124] Similarly, it can be seen that in the second similar image group G2, the similarity of image I13 is relatively higher than that of image I6; in the third similar image group G3, the similarity of image I9 is ​​relatively higher than that of images I8, I10 and I12; in the fourth similar image group G4, the similarity of image I18 is relatively higher than that of images I16 and I17; in the fifth similar image group G5, the similarity of image I19 is relatively higher than that of images I20 and I21; and in the sixth similar image group G6, the similarity of image I122 is relatively higher than that of image I123.

[0125] Similarly, for example, if the third similar image group G3 is dragged and dropped onto the image paste area A1 of page image P1 ( Figure 17 If "Yes" is selected in step 33, then... Figure 19 An image confirmation window 90, as shown, is displayed on the display screen of the display device 3.

[0126] refer to Figure 19In the similar image display area 91 of the image confirmation window 90 (an example of a recommended image notification mechanism), images I8, I9, I10, and I12 contained in the dragged and dropped third similar image group G3 are displayed. However, similar to the display in the editing window 50A, the image I9 with relatively high conformity is displayed brightly (no shadow is applied to show brightness), while the images I8, I10, and I12 with relatively low conformity are displayed darkly (shadow is applied to show darkness).

[0127] Furthermore, an image I9 with a relatively high degree of conformity is displayed in the image confirmation area 92 (an example of a recommended image notification agency), and the conformity of the image I9 displayed in the image confirmation area 92 is displayed in the conformity display area 95.

[0128] According to the second embodiment, among the images contained in a similar image group, the image with a relatively high degree of similarity can be easily identified.

[0129] Furthermore, in the first and second embodiments described above, similarity judgment conditions and compliance conditions are preset, and the display is performed as needed. Figure 10 The similarity judgment condition input window 80 shows the user inputting similarity judgment conditions, or displays... Figure 14 The input window 120 shows the criteria for matching, and the user inputs the criteria. However, it can also be set so that similarity judgment conditions and matching conditions are not preset, and the user uses the input window. Figure 10 The similarity judgment condition input window 80 shows the input of similarity judgment conditions, or you can use... Figure 14 The input window 120 shown shows the similarity judgment criteria.

[0130] Furthermore, in the first and second embodiments described above, in the display Figure 7 After the image overview window 50 shown, the following is displayed. Figure 8 The similarity judgment conditions shown specify window 60. Figure 9 The similarity judgment condition shown in window 70 is an additional window. Figure 10 The similarity judgment condition input window 80 is shown. Figure 13 The specified window 100 and the conditions shown are met. Figure 14 The conditional input window shown is 120, etc., but it can also be displayed... Figure 7 Before the image overview window 50 shown, windows 60, 70, 80, 100, 120, etc., are displayed, and the similarity judgment conditions and compliance conditions are initially determined. In short, as long as the user can specify or input (create) the similarity judgment conditions and compliance conditions, it is possible.

[0131] In addition to the CPU2, which executes software and functions as various processing units, the processing unit that performs the above-mentioned processing also includes programmable logic devices such as FPGA (field-programmable gate array) whose circuit structure can be changed after manufacturing, processors with circuit structures specifically designed for performing specific processes such as ASIC (application specific integrated circuit), and special electrical circuits.

[0132] A processing unit can be composed of a single processor from these various processors, or it can be composed of a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs, CPUs, and FPGAs). Examples of constructing multiple processing units using a single processor include: First, as exemplified by a client computer or server, a single processor is constructed from a combination of one or more CPUs and software, functioning as multiple processing units. Second, as exemplified by a system-on-a-chip (SoC), a processor uses a single integrated circuit (IC) chip to implement the overall functionality of a system containing multiple processing units. Thus, various processing units are constructed using one or more processors as hardware structures.

[0133] Moreover, more specifically, the hardware architecture of these various processors consists of electrical circuits that combine circuit elements such as semiconductor components.

[0134] In the above embodiments, the image extraction device 1 is described as a dedicated device, but it may not be a dedicated device but a personal computer, or a so-called smart device such as a smartphone or tablet, or a mobile phone such as a feature phone.

Claims

1. An image extraction apparatus characterized by comprising: Possessing: a similarity judgment condition receiving mechanism that receives input of a similarity judgment condition made by a user of the image extraction device, the similarity judgment condition being whether, when comparing any two images out of a plurality of images, the two images are similar to each other and made by the user; a similar image group determination mechanism that, out of images having a degree of matching above a threshold value, determines images that satisfy the similarity judgment condition received by the similarity judgment condition receiving mechanism and made by the user, thereby determining one or more similar image groups composed of two or more images that are similar to each other, or, out of images that satisfy the similarity judgment condition received by the similarity judgment condition receiving mechanism and made by the user, determines images having a degree of matching above a threshold value, thereby determining one or more similar image groups composed of two or more images that are similar to each other; and an image extraction mechanism that extracts at least one image in each of the one or more similar image groups determined by the similar image group determination mechanism.

2. The image extraction device according to claim 1, wherein the image extraction device further possesses: a first image display control mechanism that displays the plurality of images in a first display window formed in a display screen of a display device, and distinguishes and displays images included in the one or more similar image groups determined by the similar image group determination mechanism from images not included in the similar image groups.

3. The image extraction device according to claim 2, wherein the first image display control mechanism distinguishes and displays one similar image group from other similar image groups out of the plurality of similar image groups.

4. The image extraction device according to claim 2 or 3, wherein the image extraction device further possesses: a similar image group designation mechanism that designates one similar image group out of the one or more similar image groups determined by the similar image group determination mechanism; and a similar image display control mechanism that adjacently displays images similar to each other included in the similar image group designated by the similar image group designation mechanism in a second display window different from the first display window.

5. The image extraction device according to claim 1 or 2, wherein the image extraction device further possesses: a similarity judgment condition designation window display control mechanism that displays a similarity judgment condition designation window in which a plurality of similarity judgment conditions are displayed in a display screen of a display device, the similarity judgment condition receiving mechanism receives input of a similarity image judgment condition by designating from the plurality of similarity judgment conditions displayed in the similarity judgment condition designation window.

6. The image extraction device according to claim 1 or 2, wherein the image extraction device further possesses: a similar image group designation mechanism that designates one similar image group out of the one or more similar image groups determined by the similar image group determination mechanism; a condition meeting receiving mechanism that receives input of a condition meeting of an image; a degree of meeting calculation mechanism that calculates a degree of meeting of the condition meeting received by the condition meeting receiving mechanism for images similar to each other included in the similar image group designated by the similar image group designation mechanism; and a similar image display control mechanism that adjacently displays the images similar to each other in the similar image group designated by the similar image group designation mechanism in a display window different from the first display window. ​ a conformity notification mechanism that notifies the conformity calculated by the conformity calculation mechanism.

7. The image extraction apparatus according to claim 6, wherein the image extraction apparatus further comprises: a conformity condition designation window display control mechanism that displays a conformity condition designation window in which a plurality of conformity conditions are displayed, on a display screen of a display device, the conformity condition receiving mechanism receives input of a conformity condition designated from among the plurality of conformity conditions displayed in the conformity condition designation window.

8. The image extraction apparatus according to claim 1 or 2, wherein in the plurality of similar image groups, images included in one similar image group are dissimilar to images included in other similar image groups in a judgment based on a similarity judgment condition received by the similarity judgment condition receiving mechanism.

9. The image extraction apparatus according to claim 1, wherein the input of the similarity judgment condition made by the user is input of natural language.

10. The image extraction apparatus according to claim 1 or 2, wherein the image extraction apparatus further comprises: a similarity judgment condition input window display control mechanism that displays a similarity judgment condition input window in which a similarity judgment condition is input, on a display screen of a display device, the similarity judgment condition receiving mechanism receives input of a similarity judgment condition input in the similarity judgment condition input window.

11. The image extraction apparatus according to claim 10, wherein the input in the similarity judgment condition input window is input of natural language.

12. The image extraction apparatus according to claim 6, wherein the image extraction apparatus further comprises: a conformity condition input window display control mechanism that displays a conformity condition input window in which a conformity condition is input, on a display screen of a display device, the conformity condition receiving mechanism receives input of a conformity condition input in the conformity condition input window.

13. The image extraction apparatus according to claim 1, wherein the image extraction apparatus further comprises: a similar image group designation mechanism that designates one similar image group from among one or more similar image groups determined by the similar image group determination mechanism; a conformity calculation mechanism that calculates a conformity of a conformity condition received by the conformity condition receiving mechanism, with respect to images included in the similar image group designated by the similar image group designation mechanism; and a recommended image notification mechanism that notifies an image having the highest conformity among the conformities calculated by the conformity calculation mechanism, as a recommended image extracted by the image extraction mechanism.

14. The image extraction apparatus according to claim 1 or 2, wherein the image extraction apparatus further comprises: a similarity judgment condition addition mechanism that adds the similarity judgment condition, the similar image group determination mechanism determines one or more similar image groups each of which is composed of two or more images similar to each other, in accordance with the similarity judgment condition received by the similarity judgment condition receiving mechanism and the similarity judgment condition added by the similarity judgment condition addition mechanism.

15. The image extraction apparatus according to claim 5, wherein ​ The image extraction apparatus further includes a mechanism that, when a plurality of similarity determination conditions are designated, designates whether to determine as similar images images that satisfy all of the designated plurality of similarity determination conditions or to determine as similar images images that satisfy any one of the designated plurality of similarity determination conditions.

16. The image extraction apparatus according to claim 5, wherein A check box is formed in the similarity determination condition designation window in correspondence with the similarity determination condition, The similarity determination condition receiving mechanism receives input of a similarity image determination condition by designating the check box formed in the similarity determination condition designation window.

17. The image extraction apparatus according to claim 1 or 2, wherein The image extraction apparatus further includes: a similarity determination condition making mechanism that receives a similarity determination condition made by a user in a natural language, and makes the similarity determination condition by morpheme analysis, The similarity determination condition receiving mechanism receives input of a similarity determination condition made by the similarity determination condition making mechanism.

18. An image extraction method, characterized by a similarity determination condition receiving mechanism that receives input of a similarity determination condition made by a user in comparing any two images out of a plurality of images and judging whether the images are similar to each other, a similarity image group determination mechanism that, in images in which the degree of matching is above a threshold value, determines images that satisfy the similarity determination condition received by the similarity determination condition receiving mechanism and made by the user, and thereby determines one or a plurality of similarity image groups composed of two or more images similar to each other, or, in images that satisfy the similarity determination condition received by the similarity determination condition receiving mechanism and made by the user, determines images in which the degree of matching is above a threshold value, and thereby determines one or a plurality of similarity image groups composed of two or more images similar to each other, an image extraction mechanism that extracts at least one image in each of the one or a plurality of similarity image groups determined by the similarity image group determination mechanism.

19. A recording medium having recorded thereon a computer-readable program that controls a computer of an image extraction apparatus, The recording medium is characterized in that the computer-readable program controls a computer of an image extraction apparatus in the following manner: receives input of a similarity determination condition made by a user in comparing any two images out of a plurality of images and judging whether the images are similar to each other, in images in which the degree of matching is above a threshold value, determines images that satisfy the received similarity determination condition made by the user, and thereby determines one or a plurality of similarity image groups composed of two or more images similar to each other, or, in images that satisfy the received similarity determination condition made by the user, determines images in which the degree of matching is above a threshold value, and thereby determines one or a plurality of similarity image groups composed of two or more images similar to each other, extracts at least one image in each of the determined one or a plurality of similarity image groups.

Images