JP2010204733A - Device, method, program and system for managing found article information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically or semi-automatically perform "input of found article information" by only a minimum operation of "imaging of found article" in a system capable of intensively managing found article information via a network. <P>SOLUTION: A tag is automatically assigned to a photographic image of found article, whereby a database allowing search and management of images of found articles with the tag as a key can be constructed. Thus, an operation such as manual labeling of each found article can be eliminated. A loser can search a lost article the loser falls from the database using the tag as the key. Further, even when found article information matched to a search condition designated by the loser is searched, the found article information is not output as it is, but presented through authentication such that the loser is caused to select a partial image from the partial image and a dummy image. Thus, illicit browsing of found article information by a person who impersonates the loser can be prevented. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a technique for managing information on lost items (so-called lost items).

  Conventionally, a system for efficiently managing lost item information has been developed.

  In Patent Document 1, when a finder transmits found item information to the found item information management server using the first terminal device, a part of the found item information is set as undisclosed specific information by the found item notification processing means. The found item information is registered in the found item information database. After that, when the lost person sends a found-item information reference request to the found-item information management server using the second terminal device, the found-item information registered in the found-item information database is identified by the information disclosure means. Are transmitted to the second terminal device. Then, when the lost person transmits the lost property information to the found-item information management server using the second terminal device, the consistency determination unit determines the consistency between the lost-item information and the specific information in the found-item information database. Then, the determination result is returned to the second terminal device.

  In Patent Document 2, a lost person terminal that acquires lost article information, a finder terminal that acquires found article information, and a lost article / found article management apparatus are connected via a communication line, and the lost article / found article management apparatus. The verification processing unit for authenticating the user of the lost person terminal, the storage means for storing the found item information, the found item information based on the lost item information, and the found result information list display screen A display screen that displays a question text corresponding to the specific information included in the found-item information corresponding to the information for selecting the found item and the information for selecting the found item, receives the answer information for the question statement, and receives this answer information A lost and found item management system comprising a delivery validity checking unit that compares the information with specific information is disclosed.

  Patent Documents 3 and 4 show an example of a technique for setting an ROI (Region of Interest) in an image. Patent Document 3 discloses a technique for determining an ROI region from image motion information or the like. Patent Document 4 discloses a technique for setting an ROI with a camera operation means.

  Patent Document 5 discloses an example of a technique for generating annotation information from an image analysis result. Here, it is information corresponding to an image file, and is an annotation information generation device for generating annotation information prepared separately from the image file. An image analysis unit that extracts a feature of an image to be represented, a label derivation unit that derives a label corresponding to the feature based on at least the extracted feature, and an annotation information generation unit that generates annotation information including the derived label. An annotation information generation apparatus provided is disclosed.

  Patent Document 6 is an example of an object recognition technique, and the model feature amount extraction unit is a point on the edge of the model image for extracting a model image feature amount that is a feature amount of the model image. A model image feature amount is extracted in each of a plurality of feature amount extraction regions for one feature point in the vicinity of the feature point, and the matching unit inputs the input image at a point on the edge of the input image corresponding to the feature point The input image feature quantity, which is the feature quantity, is matched with each of the plurality of model image feature quantities at the feature points.

  Patent Document 7 shows an example of calculating a feature amount of a block set in an image.

JP 2004-30485 A JP 2006-31338 A JP 2007-329788 A JP 2001-230947 A JP 2008-33954 A JP 2008-243175 A JP 2009-27664 A

  According to Patent Documents 1 and 2, using a computer terminal connected to a network, a pickup owner can register and manage lost article information, while allowing a general lost owner to search for lost article information while preventing "spoofing". become.

  However, in Patent Documents 1 and 2, it is only possible to know how to input lost article information in general, and in particular, how to input information necessary to prevent spoofing and to identify the dropped owner reliably. There is no mention of how to sort out the “lost property information”. In general, the selection and input of such information is very complicated and requires careful consideration. In the present situation where this point has not been solved, it can be said that the burden on the system operator side (that is, the finder side) is still high.

  Accordingly, an object of the present invention is to automatically or semi-automatically “enter lost item information” with only a minimum effort of “photographing a lost item image” in a system capable of centrally managing lost item information via a network. And Further, preferably, the present invention aims to drastically improve the safety and convenience of such a system by automatically performing an operation of “identification of a drop owner for preventing impersonation”.

  The lost article information management device according to the present invention relates to an image from a lost object image input by the input unit, an input unit for inputting a lost object image obtained by imaging the lost object as a subject, information on a pickup situation of the lost object, and information on an owner of the lost object. A feature quantity extraction unit for extracting feature quantities, an annotation section for adding a tag corresponding to a lost article image based on the feature quantities extracted by the feature quantity extraction section, and an image feature from the lost article image input by the image input section Register the partial image creation unit that creates a partial image by extracting the part, the information about the acquisition status of the lost article, the information about the owner of the lost article, the lost article image, the tag, and the lost article information that associates the partial image with the database. A registration unit.

  Preferably, a search condition input unit for inputting a search condition for lost article information, a search unit for searching for lost article information including a lost article image provided with a tag that matches the search condition input from the search condition input unit, An output unit that outputs search result information corresponding to the lost article information searched by the search unit.

  Preferably, the search result information includes information obtained by excluding information related to the lost item image and / or the owner of the lost item from the lost item information.

  Preferably, a search result information selection unit that accepts selection of desired search result information from the search result information output to the output unit is provided, and the output unit displays the search result information selected from the search result information selection unit. A partial image included in the corresponding lost article information and a predetermined dummy image are mixed and output.

  Preferably, an image selection unit that accepts selection of a desired image from the partial images and dummy images output by the output unit, and image selection according to whether or not the image selected from the image selection unit is a partial image And an authentication unit that authenticates whether or not the person is the owner of the genuine lost article.

  Preferably, the output unit outputs information on the pick-up person included in the lost article information corresponding to the selected search result information in response to the authentication unit authenticating that the image selector is a true lost owner. To do.

  Preferably, the difference between the feature amount of the dummy image and the feature amount exceeds the predetermined threshold value.

  Preferably, the feature amount extracted by the feature amount extraction unit represents information on the shape or color of the subject.

  Preferably, the information related to the lost item acquisition status of the lost item information registered by the registration unit includes the shooting date and time of the lost item image, the shooting location, and other incidental information.

  Preferably, a calculation unit that calculates an elapsed time starting from the date and time of picking up a lost item included in the information related to the detection of the lost item or at the time of registration of the lost item management information, and the elapsed time calculated by the calculation unit exceeds a predetermined time A notification unit that notifies predetermined information to a predetermined partner.

  The lost article information management method according to the present invention includes a step in which a computer inputs a lost article image obtained by taking a lost article as a subject, information on a pickup situation of the lost article, and information on an owner of the lost article, and an image characteristic from the input lost article image. A step of extracting a quantity, a step of assigning a tag corresponding to the lost article image based on the feature quantity extracted by the feature quantity extraction unit, and extracting a partial image by extracting an image feature from the inputted lost article image A step of creating, and a step of registering in the database the lost item information in which the information regarding the state of finding the lost item, the information regarding the owner of the lost item, the lost image, the tag, and the partial image are associated with each other are executed.

  A program for causing a computer to execute the lost article information management method is also included in the present invention.

  A lost article management system according to the present invention includes a first terminal for inputting a lost article image, information relating to the pickup status of a lost article, and information relating to the owner of the lost article, to an input unit of the lost article information management apparatus, and a lost article information management apparatus. And a second terminal that reproduces the information output from the output unit of the lost article information management device.

  According to the present invention, it is possible to construct a database capable of automatically tagging an image of a lost article and searching and managing the image of the lost article using the tag as a key. Therefore, the trouble of manually labeling each dropped item manually can be saved. Also, the tag can be used as a key, and the owner can search the database for lost items. Furthermore, even if lost item information that matches the search condition specified by the owner of the lost item is searched, the lost item information is not output as it is, but the lost item information is presented through authentication that allows a partial image to be selected from the partial image and the dummy image. Therefore, it is possible to prevent a person pretending to be a dropper from browsing the lost item information illegally.

Schematic configuration diagram of the lost article information management system Management server block diagram Flow chart of lost article information registration process Flow chart of lost article information search processing

  FIG. 1 is a schematic configuration diagram of a lost article information management system 100 according to a preferred embodiment of the present invention. This system includes a pick-up main terminal 1, a management server 2, and a dropping main terminal 3. Pick-up main terminal 1, management server 2 and dropping main terminal 3 are each a single computer (CPU, RAM, ROM, etc. equipped with circuits necessary for arithmetic processing, data input / output circuit, display circuit, operation device, communication circuit, etc.) However, there is no essential difference in equipment configuration between the picking-up main terminal 1 and the dropping main terminal 3, and it is distinguished whether the user has picked up the lost item (the picking-up owner) or the person who picked up the item (the dropping-off owner). Is done. The picking-up main terminal 1 and the dropping main terminal 3 can be commercially available information terminals such as personal computers, mobile phones, and PDAs as they are. Each of these devices need not be installed together in the same location.

  Although only one pick-up main terminal 1 and drop main terminal 3 are shown in this figure, there may be a plurality.

  The pick-up main terminal 1 is a terminal for the owner of the lost item to connect to the management server 2 via the network such as the Internet and register information about the lost item. The camera 4 is for a pickup person to take an image of a lost item, and is typically a digital camera that records digital image data on an electronic recording medium such as a memory card. However, instead of the camera 4, a film scanner that scans an analog film on which a subject image is recorded and converts it into digital data can also be used.

  The pickup main terminal 1 includes an interface, for example, a USB port, for inputting an image (still image or moving image) from the camera 4 or its substitute. The subject of the image input to the pick-up main terminal 1 is a lost article. This lost object image is called a lost object image. The number of lost object images input corresponding to the same lost object from the same pickup main terminal 1 may be one or more.

  The pick-up main terminal 1 is connected to the management server 2 via a LAN, the Internet, or other network, and transmits lost item information including a lost item image and image incidental information other than the image to the management server 2. The information attached to the image is an ID uniquely assigned to the picker (can be replaced with an e-mail address), the name of the picker, and information about the picker such as contact information such as the address, phone number, and e-mail address. Related information), information on the acquisition status of the lost property (date related information) indicating the date and location of picking up the lost property, the item of the lost property, the name of the owner listed on the lost property, its address and telephone number, etc. Information about lost items such as contact information, lost item storage location (the location may be under the owner himself / herself, or the facility administrator or police station of a public institution (station, etc.) where lost items are reported) Lost article related information), and various information that can be used as a key for the owner to match the search condition of the lost item.

  The item of the lost item and the place where it was dropped may be hierarchized like a large classification and a small classification. For example, mobile phones are classified into large categories, mobile phones manufactured by A company, mobile phones manufactured by B company, etc., are classified into small categories. Alternatively, the major classification of the dropped place is “street”, “store”, “public office”, etc., and the minor classification is its address.

  Of course, the handling of the lost item by the owner should be stipulated by the law, but in this system it is sufficient if the storage location of the lost item itself is included in the information attached to the image, and the system itself physically stores the lost item. I do not know.

  The management server 2 is connected to the pick-up main terminal 1 and the dropping main terminal 3 via a network. The drop owner terminal 3 is an apparatus for the drop owner to input a search condition for lost article information related to his / her lost article to the management server 2 to request a search, receive the search result, and output it to a display device or a printer. is there. The drop owner terminal 3 may be a personal computer owned by the drop owner itself, or may be a terminal connected to the management server 2 via a LAN.

  As shown in FIG. 2, the management server 2 includes a lost article DB 21, an image input device 22, an input device 23, an image analysis device 24, a registration device 25, a search device 26, a display device 27, and an operation device 28. In addition, a program module executed by a computer of each device can be substituted, and a specific hardware configuration is not necessarily required.

  These devices may be integrally configured in one computer, but each device may be configured as a single computer, and the management server 2 may be configured by connecting them through a network.

Furthermore, a part of these devices may be picked up and included in the main terminal 1 or the drop main terminal 3. For example, the image input device 22, the input device 23, and the image analysis device 24 are picked up and provided in the main terminal 1, the search device 26, the display device 27, and the operation device 28 are provided in the drop main terminal 3, and the management server 2 is connected to the search device 26. You may have lost article DB21. In this way, from the analysis of the image to the registration of the lost item information can be controlled by the main terminal 3, and the management server 2 only stores the lost item information in the lost item DB 21 according to the control from the lost main terminal 3 via the network. Or
The contents of the lost property DB 21 of the management server 2 can be browsed and the contents of the lost property DB 21 can be added, deleted, changed, or updated via the operation device 28 only by an authorized administrator who is authenticated by a user ID, a password, or the like. To do.

  The image input device 22 inputs a lost article image from the pick-up main terminal 1 and transfers it to the image analysis device 24. The input device 23 inputs image supplementary information from the main terminal 1 and transfers it to the registration device 25. Both the image input device 22 and the input device 23 can be constituted by a network adapter connected to the main terminal 1 via the network. Alternatively, it may be a camera or a film scanner itself.

  The image analysis device 24 analyzes and specifies the characteristic part of the lost article from the lost article image input from the pickup main terminal 1, extracts the characteristic region indicating the specified characteristic part from the lost article image, and registers the partial image as a registration device To 25. Further, the feature amount data of the partial image is acquired by image analysis and is output to the registration device 25.

  The method for analyzing the feature region and extracting the partial image is arbitrary. For example, as in Patent Document 3, the ROI is automatically determined from the moving image, and the image (one-point embroidery, pattern, accessory, strap, etc.) included in the ROI is determined. Mascots, standby screens, bag scratches, etc.) are partial images. Alternatively, as in Patent Document 4, an image included in the ROI set by a user operation from the camera 4, the pickup main terminal 1, or the operation device 28 including a keyboard and a mouse is set as a partial image.

  The feature amount data of the partial image is a value indicating the feature of the image information. For example, as shown in paragraph 0013 of Patent Document 7, RGB three-color frequency component, pixel value histogram, dynamic range (partial image) The range from the maximum pixel value to the minimum pixel value), the number of colors (number of gradations), entropy, variance, and the like. Furthermore, the feature value data of the partial image is not related to pixel values such as the average value, variance, maximum value, minimum value, and luminance histogram of the pixel value, but also the circularity, moment, radius, volume, and area of the contour. A feature amount related to a simple shape may be included.

  The image analysis device 24 presents the automatically identified feature region on the display device 27, and one or more features arbitrarily selected by the authorized administrator's selection operation via the operation device 28 from the feature region. Only the regions may be extracted as final partial images and secured in the RAM, and the feature regions that are not selected may be deleted without being extracted as partial images. By doing so, it is possible to check through a human eye beforehand whether or not the partial image has an appropriate content for allowing the true droper to recognize the lost item.

  The partial image has a role of recognizing whether or not the lost item is the original one without presenting the lost item image. Therefore, a plurality of partial images may exist. For example, if the grip portion of the handbag is red and a yellow mascot is attached thereto, “the red grip portion of the bag” and “yellow mascot” will be extracted as partial images.

  However, if the lost article image itself can be reproduced by combining all of the partial images, an unauthenticated person who is not the original owner can grasp the entire lost article. Therefore, one or a plurality of partial images extracted from the same lost article image does not have information for reproducing the entire lost article image, in other words, information is always lost when the partial image is extracted from the lost article image. To do. The size of the partial image to be extracted may be arbitrarily set from the operation device 28. Alternatively, the condition and the number of partial images to be extracted (for example, 16 × 16 pixels or more) may be arbitrarily set from the operation device 28, and partial images that meet this condition may be extracted.

  Further, the image analysis device 24 performs annotation (tagging) based on the feature analysis result of the lost article image. For example, tags corresponding to all patterns that can be taken by the characteristic analysis result of the lost article image are prepared in advance in the lost article DB 21, and tags corresponding to the actual analysis result of the lost article image by the image analysis device 24 are prepared from the lost article DB 21. In particular, the identified tag is used as annotation information corresponding to the analyzed lost article image. There may be a plurality of annotation information.

  The contents of the lost object image feature analysis are arbitrary. For example, as in Patent Document 6, object recognition by matching a model feature amount for each subject prepared in advance with an actual feature amount of the image, and color recognition of the object. And the manufacturer's name logo printed on the surface of the object and the character recognition of the model name. The tags corresponding to the analysis result pattern prepared in the lost article DB 21 are related to the type or shape of the subject specified by the object recognition such as “umbrella”, “mobile phone”, “bag”, “purse”, “ Examples include those related to the color of the subject specified by color recognition such as “red” and “light blue”, and those related to the manufacturer (manufacturer) of lost items such as “manufactured by company A” and “made by company B”.

  If the analysis of the lost article image is normally completed, the annotation information includes a tag based on the analysis result of the lost article image, but the annotation information may include other information not based on the analysis result of the lost article image. For example, as is well known, the date and time of shooting and the shooting location (GPS) described by the camera 4 in the meta information of the image file may be included in the annotation information. Further, not the meta information itself but the result of analyzing the meta information may be included in the annotation information. For example, “AM” is included in the annotation information if the shooting date / time belongs to AM, and “PM” is included in the annotation information if the shooting date / time belongs to PM. Alternatively, if the shooting date / time belongs to a summer vacation period (for example, 7/21 to 8/31) set in the lost article DB 21 in advance, “summer vacation” is included in the annotation information. Further, text information describing a comment input from the operation device 28 and other supplementary information may be included in the annotation information.

  The lost article DB 21 is composed of a large-capacity storage medium such as a hard disk. The registration device 25 is a database engine that manages storage, deletion, update, change, addition, and other management of lost property information in the lost property DB 21. The registration device 25 associates the image supplementary information output from the input device 23 with the partial image output from the image analysis device 24 and annotation information, and stores them in the lost article DB 21. A set of these information in which the lost article image, the partial image, the feature data of the partial image, the image incidental information (including the picker related information and the lost article related information) and the annotation information are associated is referred to as lost article information. The registered lost article information may be displayed on the display device 27 in accordance with an instruction from the operation device 28 and the contents thereof may be confirmed.

  The tag, meta information, and text information that are part of the annotation information stored in the lost article DB 21 may be arbitrarily added, modified, deleted, and changed from the operation device 28 or the pickup main terminal 1.

  The search device 26 matches the search conditions specified from the operating means of the drop main terminal 3 or the operation device 28 with each piece of information included in the lost item information in the lost item DB 21, and specifies lost item information corresponding to the specified search conditions, The specified lost article information is output to a display device 27 constituted by a liquid crystal display or the like. Alternatively, when the search condition is designated from the operating means of the dropping main terminal 3, the specified lost article information may be transmitted to the dropping main terminal 3 via the network. The database engines of the search device 26 and the registration device 25 may be common or different.

  Hereinafter, processing executed in the present system will be described with reference to FIGS. 3 and 4.

  FIG. 3 is a flowchart of lost article information registration processing in which the CPU of the management server 2 (or the CPU of the picking-up main terminal 1; the same applies to the description of FIG. 3) is controlled. A program for causing the management server 2 to execute this processing is stored in the ROM of the management server 2 or the pick-up main terminal 1, and the CPU of the management server 2 or the pick-up main terminal 1 interprets the program and performs the following processing. Execute.

  In S <b> 1, the lost article image obtained by the camera 4 or the like is input to the image analysis device 24 via the image input device 22. Also, the image supplementary information is input to the registration device 25 via the input device 23.

  In S2, the image analysis device 24 creates annotation information based on the input lost article image.

  In S3, the image analysis device 24 specifies a feature region from the input lost article image.

  In S4, the image analysis device 24 creates a partial image by extracting the specified feature region from the lost article image. Also, feature data of the feature region is extracted.

  In S <b> 5, the registration device 25 creates lost article information by associating the lost article image, the partial image, the feature data of the partial image, the image incidental information (including the picker related information and the lost article related information), and the annotation information.

  In S6, the registration device 25 stores the lost article information in the lost article DB 21.

  FIG. 4 is a flowchart of lost article information search processing in which the CPU of the management server 2 (or the CPU of the dropped main terminal 3; the same applies to the explanation of FIG. 4) controls the execution. A program for causing the management server 2 to execute this processing is stored in the ROM of the management server 2 or the dropping main terminal 3, and the CPU of the management server 2 or the dropping main terminal 3 interprets the program and executes the following processing. .

  In S11, the dropping main terminal 3 accepts the designation of the search condition via an operation means such as a keyboard, connects to the management server 2, and transmits the accepted search condition. The management server 2 drops the search condition via the input device 23 and receives it from the main terminal 3 and inputs it to the search device 26. Alternatively, the search condition accepted by the operation device 28 may be input to the search device 26.

  Search conditions are as follows.

  What is the type (name) of the lost item? For example, “umbrella”, “mobile phone”, “bag”, etc.

  What is the main color of the lost item? For example, “blue” or “orange”. The color of the search condition may be specified by dropping a color palette (color sample) and displaying it on the display of the main terminal 3, and selecting a desired color from the color palette.

  The place where I dropped it. Transportation, station, store name, address, etc. These may be specified by text or may be specified on a map.

  The date and time when it was dropped. What year, month, day, etc. A date and time range may be specified.

  The amount that was dropped. However, it can be specified only when the type of the lost item is valuable (cash, securities, driver's license, mobile phone, etc.).

  The input method of the search condition is arbitrary, and may be a direct character input, or a format in which any item is selected from options such as a drop-down list.

  In S12, the search device 26 searches for lost article information in the lost article DB 21 using the input search condition as a key. When the input source of the search condition is the dropping main terminal 3, the lost article information extracted as a result of the search is transmitted to the dropping main terminal 3 and displayed on the display. When the input source of the search condition is the operation device 28, the lost article information extracted as a result of the search is displayed on the display device 27.

  However, of the lost item information, information that allows the person pretending to be the lost owner (spoofer) to uniquely identify the lost item is not displayed and transmitted to the lost owner terminal 3. This is to prevent the misrepresentation dropper from illegally acquiring someone else's lost property.

  Specifically, of the lost item information, information (critical information) that allows the spoofed dropper such as the lost item image, the partial image, and the picker-related information to identify the lost item itself is not transmitted / displayed. Instead, information that makes it easy to narrow down lost items with information that excludes critical information from lost item information (search result display information), for example, “Product Name” + “Date and Time of Dropped” + “Place of Dropped (only major category)” + Send / display the “color of the lost item”.

  In addition, when the lost article information corresponding to the designated search condition cannot be extracted, it is preferable that a search condition can be added or the search condition can be specified again.

  Hereinafter, the display destination of information is assumed to be the display of the dropping main terminal 3 or the display device 27 corresponding to the input source of the search condition, and is not described one by one. Further, instead of video display on the display, audio output by an audio playback device or printout by a printer can be used, and the information playback output form is not limited to video display.

  In S <b> 13, the search device 26 causes the drop main terminal 3 or the operation device 28 to select from the displayed search result display information what the user who searches for the lost item thinks about his / her own lost item. Here, as will be described below, before the details of the lost property information corresponding to the selected search result display information are displayed in detail, it is authenticated whether or not the owner who has designated the search condition is authentic. .

  In S14, the search device 26 displays all or part of the partial image group included in the lost article information including the selected search result display information. In addition, the search device 26 randomly selects a predetermined number of dummy images from the dummy image group stored in advance in the lost item DB 21 regardless of the lost item information, and selects the selected dummy images as lost item information. Are displayed together with the partial image group included in the. The arrangement of the dummy image and the partial image is random (no regularity). For example, partial images are arranged in cells randomly selected from 4 × 4 cells, and dummy images are arranged in the remaining cells. It should be noted that a random number serving as a criterion for randomness is generated using a pseudo-random number generated by a CPU or a chip set according to a known method.

  In the dummy image group, dummy images corresponding to each analysis pattern are prepared in the lost article DB 21 so as to cover the analysis patterns of the lost article image. For example, “umbrella”, “mobile phone”, “bag”, “wallet”, etc., dummy corresponding to the type of subject specified by object recognition, for example “umbrella dummy”, “mobile phone dummy”, “bag” `` Dummy '', `` Wallet dummy '', a dummy corresponding to the subject type and corresponding to the subject color specified by color recognition such as `` red '', `` light blue '', such as `` red umbrella dummy '', There is a “light blue bag dummy”.

  The feature amount data of the dummy image selected by the search device 26 is different from the feature amount data of the partial image of the lost article information corresponding to the selected search result display information. This is because if there is no difference between them, the difference between the dummy image and the partial image of the lost item is not visually recognized even by the owner. On the contrary, the difference between the two needs to be a level that can be recognized only when the dropping person actually sees the dropping person, that is, a level that cannot be recognized by a spoofing dropping person who has never seen the real thing. For example, for the partial image of “red umbrella”, the search device 26 uses a dummy image “blue umbrella dummy image” or “yellow umbrella dummy” having frequency components different from the three RGB color components of the partial image. Select “Image”. Alternatively, for a partial image of “a red umbrella with a grip pattern J”, “a dummy image of a red umbrella with a grip pattern L” is selected. Alternatively, a type of dummy that is completely different from the type of object in the partial image, such as a wallet or a bag, may be selected for the partial image of “umbrella”. An administrator having authority may arbitrarily set what kind of feature amount is to be different between the partial image and the dummy image to be selected via the operation device 28.

  In general, the search device 26 uses a dummy image in which the difference (difference) between the feature amount data of the partial image and the feature amount data of the dummy image, such as color difference, luminance difference, and shape difference, exceeds a predetermined threshold. Select. A dummy image may be selected such that only the degree of difference relating to the pixel value, the degree of difference relating to the shape, or both of which exceed the corresponding predetermined threshold. It may be determined at random which pixel attribute / shape attribute difference is compared with a predetermined threshold.

  Further, even if a dummy image that is different from the feature amount of the partial image is clearly displayed, it is highly likely that it will be overlooked by the spoofer if it is not a partial image. Therefore, the degree of difference based on the partial image has a predetermined first value. A dummy image that exceeds the threshold but does not satisfy the predetermined second threshold may be selected.

  Note that the dummy image has the same size as the partial image, and the difference between the dummy image and the partial image cannot be recognized by visually observing the difference in size.

  The number of dummy images to be selected may be determined according to the number of partial image groups. For example, the maximum integer that does not exceed three times the number of partial images is set as the number of dummy images to be selected. This is because displaying an extremely larger number of dummy images than the partial image causes mischief. With each partial image as a reference, the same number of dummy images that have a degree of difference that exceeds a predetermined first threshold but does not satisfy a predetermined second threshold may be selected.

  In S <b> 15, the search device 26 causes the dropped main terminal 3 or the operation device 28 to select an image that the user himself / herself determines as a partial image related to the lost item of the user from the displayed partial image and dummy image. Instead of directly selecting the partial image itself, the dummy image may be selected as an exclusion candidate, and the partial image may be indirectly selected by regarding the remaining image not selected as the exclusion candidate as the partial image.

  In S16, the search device 26 determines whether a dummy image is included in the image selected in S15. In the case of No (when no dummy image is included), it is determined that the access from the authentic user is authenticated (authentication OK), and the process proceeds to S17. In the case of Yes (at least one dummy image is included) If YES, the access from the unauthenticated user is authenticated (authentication NG), and the process returns to S13. In the case of authentication NG, the process may return to S12 to prompt the user to reset the search condition. It should be noted that the number of times of authentication NG is counted, and when this number reaches a predetermined threshold value (for example, 3 times), it is determined that this user is a non-authentic user, and the process does not return to S13. After finishing this process, a message such as “authentication failed” may be displayed, and thereafter, access from the same drop main terminal 3 that selected the image in S15 may be restricted. This is for preventing a non-authentic user from seeing through a partial image while selecting an image many times.

  In S17, all or a part of each piece of information included in the lost article information, for example, a lost article image is displayed. It is sufficient that the information related to the picker is displayed as long as the dropper can contact the picker and take the lost item to an appropriate storage location. For example, only the mail address of the picker is displayed, and the name and address of the picker may not be displayed. It may be possible to preliminarily select from the main terminal 1 which item of the lost article information is to be displayed to the authenticated user.

  The lost owner can check the displayed lost item information, contact the lost owner according to the information, and get the lost item stored by the picker or other items.

  Note that the RTC (Real Time Clock) (not shown) provided in the management server 2 measures the registration date of the lost item information and the elapsed time from the date of picking up the lost item, and if this elapsed time exceeds a predetermined period, the lost item The information is considered to be related to the lost property that the lost person did not enter, and the sale of the lost property is started by notifying the fact or the lost property information to an auction server (not shown), or it is picked up based on the information related to the owner corresponding to the lost property information. You may connect to the main terminal 4 and transmit the notification to the effect that the picker became the owner. This will make it easier for lost items to be left behind.

1: Picking main terminal, 2: Management server, 3: Dropping main terminal, 4: Camera

Claims (13)

An input unit for inputting a lost object image obtained by imaging a lost object as a subject, information on the state of picking up the lost object, and information on the owner of the lost object;
A feature amount extraction unit that extracts a feature amount related to an image from a lost article image input by the input unit;
Based on the feature amount extracted by the feature amount extraction unit, an annotation unit that gives a tag corresponding to the lost article image;
A partial image creation unit for creating a partial image by extracting an image feature from the lost article image input by the image input unit;
A registration unit for registering in the database information relating to the state of picking up the lost property, information relating to the owner of the lost property, and lost property information that associates the lost property image, the tag, and the partial image;
Lost and found information management device. A search condition input section for inputting search conditions for the lost property information;
A search unit that searches the database for lost article information including a lost article image that has been assigned a tag that matches the search condition input from the search condition input unit;
An output unit for outputting search result information corresponding to lost article information searched by the search unit;
The lost article information management apparatus according to claim 1.   The lost article information management apparatus according to claim 2, wherein the search result information includes information obtained by excluding information regarding the lost article image and / or the owner of the lost article from the lost article information. A search result information selection unit that accepts selection of desired search result information from the search result information output to the output unit,
4. The lost article according to claim 2, wherein the output unit outputs a partial image included in lost article information corresponding to the search result information selected from the search result information selection unit and a predetermined dummy image in a mixed manner. Information management device. An image selection unit that accepts selection of a desired image from the partial image and the dummy image output by the output unit;
An authenticating unit for authenticating whether or not the image selector is a true dropper, depending on whether the image selected from the image selecting unit is a partial image;
The lost article information management device according to claim 4.   In response to the fact that the authentication unit has authenticated that the image selector is a genuine lost property owner, the output unit obtains information regarding the picker included in the lost property information corresponding to the selected search result information. 6. The lost article information management apparatus according to claim 5, wherein the lost article information management apparatus outputs the lost article information.   The lost article information management device according to any one of claims 4 to 6, wherein a difference between the feature amount of the dummy image and the feature amount of the partial image exceeds a predetermined threshold.   The lost article information management apparatus according to claim 1, wherein the feature quantity extracted by the feature quantity extraction unit represents information related to a shape or color of a subject.   The lost article information management apparatus according to any one of claims 1 to 8, wherein the information related to the lost object acquisition status of the lost article information registered by the registration unit includes shooting date and time, a shooting location, and other incidental information of the lost article image. A calculating unit that calculates an elapsed time starting from the date and time of picking up the lost item included in the information related to the state of picking up the lost item or when the lost item management information is registered;
A notification unit that notifies predetermined information to a predetermined partner when the elapsed time calculated by the calculation unit exceeds a predetermined time;
The lost article information management device according to any one of claims 1 to 9. Computer
A step of inputting a lost article image obtained by imaging a lost article as a subject, information on the state of acquisition of the lost article, and information on the owner of the lost article;
Extracting a feature amount related to the image from the inputted lost article image;
Assigning a tag corresponding to the lost article image based on the feature quantity extracted by the feature quantity extraction unit;
Creating a partial image by extracting an image feature from the input lost article image;
Registering in the database information relating to the state of picking up the lost property, information relating to the owner of the lost property, lost property information that associates the lost property image, the tag, and the partial image;
Lost and found information management method.   A program for causing a computer to execute the lost article information management method according to claim 11. A first terminal for inputting the lost article image, information on the pickup situation of the lost article, and information on the owner of the lost article, to the input unit of the lost article information management device according to any one of claims 1 to 10.
A second terminal that inputs the search condition to the search condition input unit of the lost article information management device and reproduces the information output from the output unit of the lost item information management device;
Lost and found information management system.

Images