KR20230162409A - Method, system and non-transitory computer-readable recording medium for detecting near-duplicate images - Google Patents

Abstract

The present disclosure relates to a method, system, and non-transitory computer-readable recording medium for detecting duplicate images. A method of detecting a duplicate image including the same subject according to an embodiment of the present disclosure includes receiving a reference image including at least one first subject, and determining at least one similar image similar to the reference image. , calculating a similarity between at least one first subject included in the reference image and at least one second subject included in the similar image, if the similarity satisfies a predetermined first standard, the similar image is It includes determining a duplicate image of a reference image and providing information about the duplicate image.

Description

METHOD, SYSTEM AND NON-TRANSITORY COMPUTER-READABLE RECORDING MEDIUM FOR DETECTING NEAR-DUPLICATE IMAGES}

The present disclosure relates to a method, system, and non-transitory computer-readable recording medium for detecting duplicate images, and more specifically, to detecting and recognizing a modified image including a subject identical to a subject included in a reference image as a duplicate image. It relates to a method, system, and non-transitory computer-readable recording medium.

Recently, as the online market's share of the overall retail sales market has rapidly increased, consumers can search for desired products and compare products found on multiple online sales sites to decide whether to purchase them.

In particular, in online open markets, multiple brands and different businesses sell products, and images for the same product may be provided with some different backgrounds or rotated images. In this case, not only consumers but also open market operators may need a filtering function for the same product.

In this regard, the development and research of technology for detecting some modified images as duplicate images has been actively conducted recently.

For example, Patent Registration No. 10-2260631 discloses a method and device for searching duplicate image files. In the prior art, the feature points of each image file are extracted, the Euclidean distance of the feature points extracted from the same image file is calculated, and the number of feature points and/or the Euclidean distance are compared to determine whether or not there is a duplicate image.

According to this prior art, when only some features of each image are compared, quick comparison is possible, but there is a problem of low accuracy, and when the features compared in each image are increased, the processing speed decreases and the server workload increases. .

Accordingly, the present inventor proposes a duplicate image detection service that is efficient and improves accuracy by first setting a candidate group of duplicate images by briefly comparing the similarity of the entire area of the image, and then detecting duplicate images by comparing detailed regions. .

Registered Patent Publication No. 10-2260631 (May 31, 2021)

The purpose of the present disclosure is to solve all the problems of the prior art described above.

Additionally, the purpose of the present disclosure is to provide a method and system for detecting images containing the same subject but with a changed angle or focus or a changed posture of the subject as duplicate images.

Another purpose of the present disclosure is to provide a method and system for quickly and efficiently detecting a modified image containing the same subject as a duplicate image.

A method of detecting a duplicate image including the same subject according to an embodiment of the present disclosure includes receiving a reference image including at least one first subject, and determining at least one similar image similar to the reference image. , calculating a similarity between at least one first subject included in the reference image and at least one second subject included in the similar image, if the similarity satisfies a predetermined first standard, the similar image is It includes determining a duplicate image of a reference image and providing information about the duplicate image.

According to an embodiment of the present disclosure, determining at least one similar image similar to the reference image includes calculating an overall degree of similarity by comparing at least one image included in a predetermined image database with the reference image; The method may include determining the at least one image as a similar image when the overall similarity satisfies a second predetermined criterion.

According to an embodiment of the present disclosure, calculating the similarity between at least one first subject included in the reference image and at least one second subject included in the similar image includes at least one subject included in the reference image. Detecting a first subject, detecting at least one second subject included in the similar image, and calculating similarity by comparing the at least one first subject and the at least one second subject. It can be included.

According to an embodiment of the present disclosure, the step of calculating similarity by comparing the at least one first subject and the at least one second subject includes determining a first category of the at least one first subject, A second category of the at least one second subject is determined, so that among the at least one first subject and the at least one second subject, the first subject and the second subject have the same first and second categories. The similarity can be calculated by comparing .

According to an embodiment of the present disclosure, the step of calculating similarity by comparing the at least one first subject and the at least one second subject may include comparing the at least one first subject to a second subject located in the center of the reference image. 1 A subject may be determined, a second subject located in the center of the similar image among the at least one second subject may be determined, and the degree of similarity may be calculated by comparing the first subject and the second subject.

According to an embodiment of the present disclosure, the first subject located in the center of the reference image is determined to be the first subject in which the distance between the center point of the reference image and the center point of the at least one first subject is closest, The second subject located in the center of the similar image may be determined as the second subject whose distance between the center point of the similar image and the center point of the at least one second subject is closest.

According to an embodiment of the present disclosure, the similarity is determined by comparing a first subject and a second subject in which the first category and the second category are the same among the at least one first subject and the at least one second subject. Thus, the subject similarity can be calculated and determined by assigning a weight to the subject similarity.

According to an embodiment of the present disclosure, calculating the overall similarity includes calculating a first similarity of the center of the reference image and the center of the at least one image, and calculating the first similarity of the center of the reference image and the center of the at least one image. The overall similarity can be calculated by calculating the second similarity of the periphery of and assigning weights to the first similarity and the second similarity.

According to an embodiment of the present disclosure, calculating the overall similarity includes comparing each of the at least one first subject included in the reference image and the at least one second subject included in the similar image to determine the similarity. Calculate, select the closest similarity among them as the first similarity, calculate the second similarity by comparing the first subject located in the center of the reference image with the second subject located in the center of the similar image, and calculate the second similarity. The total similarity can be calculated by assigning a predetermined weight to the first similarity and the second similarity.

According to an embodiment of the present disclosure, in the step of providing information about the duplicate image, information about the duplicate image may be provided by sorting based on the similarity.

A system for detecting duplicate images including the same subject according to an embodiment of the present disclosure includes a reference image receiving unit that receives a reference image including at least one first subject, and a similar image receiving unit that determines at least one similar image similar to the reference image. An image determination unit, calculates the similarity between at least one first subject included in the reference image and at least one second subject included in the similar image, and if the similarity satisfies a predetermined first standard, the similar image It includes a duplicate image determination unit that determines a duplicate image of the reference image and a duplicate image information provider that provides information about the duplicate image.

In addition, other methods for implementing the present disclosure, other systems, and non-transitory computer-readable recording media for recording computer programs for executing the methods are further provided.

According to the present disclosure, it is possible to provide a method and system that can detect a modified image including the same subject as a duplicate image.

Additionally, according to the present disclosure, a modified image including the same subject can be quickly and efficiently detected as a duplicate image.

1 is a diagram illustrating a schematic configuration of a system environment for detecting duplicate images according to an embodiment of the present disclosure.
FIG. 2 is a diagram illustrating a reference image and a duplicate image (eg, a duplicate image set) of the reference image according to an embodiment of the present disclosure.
Figure 3 is a functional block diagram schematically showing the functional configuration of a duplicate image detection server according to an embodiment of the present disclosure.
FIG. 4 is a diagram illustrating an entire area comparison between a reference image and a comparison target image according to an embodiment of the present disclosure.
FIG. 5 is a diagram illustrating a method of comparing a first subject of a reference image and a second subject of a similar image according to an embodiment of the present disclosure.
FIG. 6 is a diagram illustrating another example of a method of comparing a first subject of a reference image and a second subject of a similar image according to an embodiment of the present disclosure.
Figure 7 is an operation flowchart showing a process for detecting duplicate images according to an embodiment of the present disclosure.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Below, if it is judged that there is a risk of unnecessarily obscuring the gist of the present disclosure, detailed descriptions of already known functions and configurations will be omitted. In addition, it should be noted that the content described below only relates to one embodiment of the present disclosure and is not limited thereto.

The terms used in this disclosure are only used to describe specific embodiments and are not intended to limit the disclosure. For example, a component expressed in the singular should be understood as a concept that includes plural components unless the context clearly indicates only the singular. It should be understood that the term “and/or” as used in this disclosure encompasses any and all possible combinations of one or more of the listed items. Terms such as 'include' or 'have' used in this disclosure are only intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in this disclosure, and the meaning of these terms is The use is not intended to exclude the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

In the embodiments of the present disclosure, 'module' or 'unit' refers to a functional part that performs at least one function or operation, and may be implemented as hardware or software, or as a combination of hardware and software. Additionally, a plurality of 'modules' or 'units' may be integrated into at least one software module and implemented by at least one processor, except for 'modules' or 'units' that need to be implemented with specific hardware. there is.

In addition, unless otherwise defined, all terms used in this disclosure, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the technical field to which this disclosure pertains. It should be noted that terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with the context meaning of the related technology, and should not be interpreted in an overly restrictive or expanded manner unless clearly defined otherwise in the present disclosure. .

Hereinafter, a method for automatically extracting image data features according to an embodiment of the present disclosure will be described in detail with reference to the drawings.

1 is a diagram illustrating a schematic configuration of a system environment for detecting duplicate images according to an embodiment of the present disclosure.

As shown in FIG. 1, the duplicate image detection system 100 according to an embodiment of the present disclosure may include a plurality of user terminals 110, a communication network 120, and a duplicate image detection server 130.

The user terminal 110 according to an embodiment of the present disclosure is a digital device that includes a function to connect to and communicate with the duplicate image detection server 130 through the communication network 120. It may be a portable digital device, such as a smartphone, tablet PC, etc., equipped with a memory means and equipped with a microprocessor to have computing power, and is not limited to a specific form. Although three user terminals are shown in this drawing, the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, various types of user input received on the user terminal 110 may be transmitted to the duplicate image detection server 130 through the communication network 120. According to an embodiment of the present disclosure, the user terminal 110 may receive various signals transmitted from the outside (eg, the duplicate image detection server 130) through the communication network 120.

According to an embodiment of the present disclosure, the user terminal 110 may include an application to support the function according to the present disclosure. Such an application may be downloaded from the duplicate image detection server 130 or an external application distribution server (not shown).

The communication network 120 according to an embodiment of the present disclosure may be configured regardless of communication mode, such as wired communication or wireless communication, and may include a local area network (LAN), a metropolitan area network (MAN), It can be composed of various communication networks, such as a wide area network (WAN). Preferably, the communication network 120 referred to in this specification may be the known Internet or World Wide Web (WWW). However, the communication network 120 is not necessarily limited thereto and may include at least a portion of a known wired or wireless data communication network, a known telephone network, or a known wired or wireless television communication network.

For example, the communication network 120 is a wireless data communication network, including Wi-Fi communication, WiFi-Direct communication, Long Term Evolution (LTE) communication, and Bluetooth communication (e.g., low-power Bluetooth (e.g., low-power Bluetooth). It may implement at least part of a conventional communication method such as BLE (Bluetooth Low Energy) communication, infrared communication, ultrasonic communication, etc.

The duplicate image detection server 130 according to an embodiment of the present disclosure may detect duplicate images for a predetermined reference image and provide information. The duplicate image detection server 130 may detect an image containing the same subject as a subject included in a predetermined reference image as a duplicate image and provide information.

FIG. 2 is a diagram illustrating a reference image and a duplicate image (eg, a duplicate image set) of the reference image according to an embodiment of the present disclosure.

Here, the duplicate image is an image containing the same subject as the subject included in the reference image, and includes an image identical to the reference image or an image in which the reference image has been modified. For example, Figure 2(a) shows a reference image, and Figure 2(b) shows a duplicate image in which the background color of the reference image has been modified. As another example, Figure 2(c) is a reference image, and Figure 2(d) shows a duplicate image transformed by rotating or inverting the subject of the reference image. As another example, Figure 2(e) is a reference image, and Figure 2(f) shows a duplicate image taken of the same location and subject by changing the shooting angle or focus, or a duplicate image in which the subject's posture has been changed. As shown in FIG. 2, the duplicate image detection server 130 according to an embodiment of the present disclosure includes an image that includes the same subject as the reference image, but has a different background (FIG. 2(a), (b)), and a rotated image. Information is collected by detecting duplicate images for images with modified reference images, such as images (Figures 2(c) and (d)) and images of different cuts for the same location and subject (Figures 2(e) and (f)). can be provided. In this drawing, an image in which the modification of the reference image can be distinguished by the naked eye is shown as an example of a duplicate image, but this is not limited. In addition, a duplicate image may also include cases where the image quality of the reference image is changed or the resolution of the image is changed. there is.

As shown, the duplicate image detection server 130 may communicate with the user terminal 110 through the communication network 120.

The duplicate image detection server 130 according to an embodiment of the present disclosure may be a digital device equipped with a memory means and equipped with a microprocessor to have computing capabilities. This duplicate image detection server 130 may be a server system.

Figure 3 is a functional block diagram schematically showing the functional configuration of a duplicate image detection server according to an embodiment of the present disclosure.

Referring to FIG. 3, the duplicate image detection server 130 includes a reference image receiving unit 302, a similar image determining unit 304, a duplicate image determining unit 306, a duplicate image information providing unit 308, and a communication unit 310. may include. The components shown in FIG. 3 do not reflect all functions of the duplicate image detection server 130 and are not essential, so the duplicate image detection server 130 may include more or fewer components than the components shown. May contain components.

According to an embodiment of the present disclosure, the duplicate image detection server 130 includes a reference image receiving unit 302, a similar image determining unit 304, a duplicate image determining unit 306, a duplicate image information providing unit 308, and a communication unit. 310 may be a program module, at least some of which communicate with an external system. These program modules may be included in the duplicate image detection server 130 in the form of an operating system, application program module, or other program module, and may be physically stored in various known storage devices. Additionally, these program modules may be stored in a remote storage device capable of communicating with the duplicate image detection server 130. Meanwhile, such program modules include, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc. that perform specific tasks or execute specific abstract data types according to the present disclosure.

According to an embodiment of the present disclosure, the reference image receiver 302 performs a function of receiving a predetermined reference image. According to an embodiment of the present disclosure, the reference image receiver 302 may receive an image including at least one subject as a reference image. According to an embodiment of the present disclosure, the reference image receiver 302 may receive an image including a fashion item as a subject as a reference image. In one embodiment, the reference image receiver 302 may receive an image including at least one subject and a background image as a reference image.

According to an embodiment of the present disclosure, the similar image determination unit 304 performs a function of determining images similar to a reference image as a duplicate image candidate group. Here, an image similar to the reference image is an image whose similarity to the entire area (hereinafter referred to as global area) of the reference image satisfies a predetermined value or more, and is a group of duplicate image candidates that can be duplicate images of the reference image.

According to an embodiment of the present disclosure, the similar image determination unit 304 may calculate the overall similarity and compare the reference image and the comparison target image without comparing detailed differences. According to one embodiment, the similar image determination unit 304 calculates the similarity between the global area of the comparison target image and the global area of the reference image included in a predetermined image database, and selects the comparison target image if it satisfies a predetermined value or more. You can decide based on similar images.

According to an embodiment of the present disclosure, similarity can be expressed as the distance of each feature vector. Here, the feature vector is a vector containing attribute information of the image in the style space. In one embodiment, the feature vector may be a vector in a style space that includes attribute information such as item type, category, color, pattern, style, and length included in the image. According to an embodiment of the present disclosure, the closer the distance between each feature vector, the higher the similarity can be determined.

In one embodiment of the present disclosure, the similar image determination unit 304 extracts a feature vector for the global area of the reference image and extracts a feature vector for the global area of the comparison target image included in a predetermined image database to determine the similarity. can be calculated. In one embodiment, the feature vector for the global region may be a vector in a style space that includes attribute information for the global region of the image, that is, the entire region. In one embodiment of the present disclosure, the feature vector for the global area may include attribute information about at least one subject included in the image and attribute information about the background image.

According to an embodiment of the present disclosure, the similar image determination unit 304 may calculate the overall similarity by assigning different weights to the degree of similarity between the center and the periphery of each image included in a predetermined image database and the reference image. .

According to an embodiment of the present disclosure, the similar image determination unit 304 calculates the first similarity by comparing the center of the reference image with the center of the comparison image, and compares the peripheral portion of the reference image with the peripheral portion of the comparison image. By calculating the second similarity, the total similarity can be calculated by adding the first similarity and the second similarity.

According to an embodiment of the present disclosure, the similar image determination unit 304 may calculate the degree of similarity by assigning a predetermined weight to each component value of the feature vector of the reference image and the feature vector of the image to be compared. In one embodiment, the similar image determination unit 304 may calculate the similarity by assigning a greater weight to the vector component value representing the properties of the subject included in the image than the component value representing the properties of the background.

According to an embodiment of the present disclosure, when the similarity between the reference image and the image to be compared satisfies a predetermined second standard, the similar image determination unit 304 selects the image to be compared as a duplicate image candidate group for the reference image. You can decide based on similar images.

FIG. 4 is a diagram illustrating an entire area comparison between a reference image and a comparison target image according to an embodiment of the present disclosure.

Referring to FIG. 4, the reference image 401 and the comparison target image 403 have different visible areas for the same image. Although the reference image 401 and the comparison target image 403 do not correspond to the same image, the similar image determination unit 304 compares the global area of the reference image 401 and the comparison target image 403 to compare the comparison target image ( 403) can be determined as a similar image.

According to an embodiment of the present disclosure, the similar image determination unit 304 extracts a feature vector containing attribute information for the global area of the reference image 401 and attributes for the global area of the comparison target image 403. After extracting a feature vector containing information, similarity can be calculated by comparing the feature vectors.

According to an embodiment of the present disclosure, the feature vector for the global area may include attribute information about the person as the subject included in the reference image 401 and the comparison target image 403 and information about the shape of the globe as the background image. You can. In one embodiment, the similar image determination unit 304 calculates the difference in component values including attribute information about the subject of the reference image 401 and the comparison target image 403 into component values including attribute information about the background image. Similarity can be calculated by assigning a weight greater than the difference. According to an embodiment of the present disclosure, the similar image determination unit 304 calculates the degree of similarity between the reference image 401 and the comparison target image 403, and, if a predetermined second standard is satisfied, the comparison target image 403 can be determined as a similar image to the reference image 403.

Referring again to FIG. 3, according to an embodiment of the present disclosure, the duplicate image determination unit 306 compares the local area of the reference image and the local area of the similar image to determine whether the similar image corresponds to a duplicate image of the reference image. It performs the function of deciding. Here, the local area of the image refers to the subject area included in each image.

According to one embodiment, the duplicate image determination unit 306 calculates the degree of similarity by comparing the first subject included in the reference image and the second subject included in the similar image, and when the similarity satisfies a predetermined value or more. The similar image may be determined as a duplicate image for the reference image.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 may extract a subject included in the reference image into a local area and determine it as the first subject. Additionally, the duplicate image determination unit 306 may extract a subject included in a similar image to a local area and determine it as the second subject. In one embodiment, the reference image and the similar image may include a plurality of subjects. In this case, the duplicate image determination unit 306 may determine each subject included in the reference image as the first subject and each subject included in the similar image as the second subject.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 may calculate the degree of similarity by comparing at least one first subject included in the reference image and at least one second subject included in the similar image.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 may extract feature vectors for each first subject and second subject, and calculate similarity by comparing the respective feature vectors.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 calculates the degree of similarity by comparing each of the plurality of first subjects included in the reference image and the plurality of second subjects included in the similar image. You can select the similarity corresponding to the closest one.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 first determines the category of the first subject included in the reference image and the second subject included in the similar image, and determines the category of the first subject and the second subject. Only when the categories are the same, similarity can be calculated by comparing the first subject and the second subject. In one embodiment, the duplicate image determination unit 306 may extract the category of the first subject from the feature vector of the first subject and extract the category of the second subject from the feature vector of the second subject.

According to an embodiment of the present disclosure, when the reference image and the similar image each include a plurality of subjects, the duplicate image determination unit 306 may preferentially determine the category of each first subject and the category of the second subject. there is. In one embodiment, the duplicate image determination unit 306 selects the first and second subjects corresponding to the same category among the plurality of first subjects included in the reference image and the plurality of second subjects included in the similar image. Similarity can be calculated by comparison.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 determines a first subject located in the center of the reference image and a second subject located in the center of a similar image, and compares the first subject and the second subject. Similarity can be calculated. In one embodiment, the duplicate image determination unit 306 selects the first subject in which the distance between the center point of the reference image and the center point of the at least one first subject is closest, and the distance between the center point of the reference image and the center point of the at least one first subject is closest. 2 The similarity can be calculated by selecting the second subject whose central point distance is closest to the subject and comparing the first subject and the second subject. In one embodiment, even when the reference image and the similar image each include a plurality of subjects, the similarity calculation process can be simplified by calculating similarity through comparison between subjects located in the center of the image.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 may calculate the similarity by calculating the subject similarity of each of the plurality of first and second subjects and adding them together. According to an embodiment of the present disclosure, the duplicate image determination unit 306 may calculate the similarity by assigning a weight to the similarity of each subject and adding them together.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 may calculate the similarity by assigning a large weight to the similarity between the first subject located in the center of the reference image and the second subject located in the center of the similar image.

According to an embodiment of the present disclosure, the duplicate image determination unit 306 calculates the degree of similarity by comparing each of the at least one first subject included in the reference image and the at least one second subject included in the similar image, Among them, the closest one is selected as the first similarity, the second similarity is calculated by comparing the first subject located in the center of the reference image and the second subject located in the center of the similar image, and then the first similarity and the second similarity are calculated. 2 The overall similarity can be calculated by assigning weights to the similarity.

According to an embodiment of the present disclosure, when the similarity between the first subject and the second subject satisfies a predetermined first standard, the duplicate image determination unit 306 may determine the similar image as a duplicate image for the reference image. . In one embodiment of the present disclosure, a separate verification data set may be configured in advance, and a reference value that can maximize the accuracy of detecting duplicate images may be determined using the verification data set as a predetermined first standard.

In one embodiment, the predetermined first criterion for determining a similar image as a duplicate image in the duplicate image determination unit 306 is a predetermined first criterion for determining the image to be compared as a similar image to the reference image in the similar image determination unit 304. It can be determined more strictly than the second standard.

FIG. 5 is a diagram illustrating a method of comparing a first subject of a reference image and a second subject of a similar image according to an embodiment of the present disclosure.

Referring to FIG. 5, the duplicate image determination unit 306 determines the first subjects 503 and 505 included in the reference image 501 and the second subjects 509 and 511 included in the similar image 507. can be decided. In one embodiment of the present disclosure, the duplicate image determination unit 305 may extract feature vectors for each of the first subjects 503 and 505 and feature vectors for each of the second subjects 509 and 511.

In one embodiment, the duplicate image determination unit 306 determines the category (503: top, 505: bottom) of the first subjects (503, 505) and the category (509: top) of the second subject (509, 511). , 511: bottoms) can be determined.

In one embodiment of the present disclosure, the duplicate image determination unit 306 compares the feature vectors of the first subject 503 and the second subject 509 corresponding to the top among the plurality of first and second subjects. The similarity of the first subject may be calculated and the similarity of the second subject may be calculated by comparing the feature vectors of the first subject 505 and the second subject 511 corresponding to the bottoms.

In one embodiment, the duplicate image determination unit 306 may calculate the similarity of the local area of the reference image 501 and the similar image 507 by adding the first and second subject similarities.

FIG. 6 is a diagram illustrating another example of a method of comparing a first subject of a reference image and a second subject of a similar image according to an embodiment of the present disclosure.

In FIG. 6 , the duplicate image determination unit 306 may determine the first subject 603 located in the center of the reference image 601 and the second subject 605 located in the center of the similar image 605. The duplicate image determination unit 306 may calculate similarity by comparing only the first subject 603 and the second subject 605 among the plurality of subjects.

According to an embodiment of the present disclosure, when the similarity between the first and second subjects satisfies a predetermined value, the duplicate image determination unit 306 selects the similar images 403 and 507 as the reference images 401 and 501. ) can be determined by the duplicate image.

Referring again to FIG. 3, according to an embodiment of the present disclosure, the duplicate image information provider 308 performs a function of providing information on a duplicate image for a reference image. In one embodiment, the duplicate image information provider 308 may provide the user terminal 110 with information about an image determined to be a duplicate image with respect to the reference image among images to be compared in a predetermined image database.

According to an embodiment of the present disclosure, the duplicate image information provider 308 may provide information by sorting duplicate images in order of high similarity to the reference image.

The communication unit 310 of the duplicate image detection server 130 according to an embodiment of the present disclosure includes a reference image receiver 302, a similar image determination unit 304, a duplicate image determination unit 306, and a duplicate image information provider ( 308) performs a function that enables data transmission and reception from/to.

Figure 7 is an operation flowchart showing a process for detecting a duplicate image according to an embodiment of the present disclosure.

First, in step S701, the duplicate image detection server 130 receives a predetermined reference image. The reference image is an image including a subject, and the duplicate image detection server 130 may receive an image including at least one subject and a background image as the reference image.

Next, in step S703, the duplicate image detection server 130 calculates the similarity of the entire area (hereinafter referred to as global area) between the image to be compared and the reference image in a predetermined image database. In one embodiment, the duplicate image detection server 130 extracts a global region feature vector of a reference image and a global region feature vector of a comparison target image, and compares them to calculate similarity. In one embodiment of the present disclosure, the duplicate image detection server 130 may calculate the overall similarity by applying a greater weight to the center of the reference image and the image to be compared than to the periphery.

In step S705, the duplicate image detection server 130 may determine the comparison target image as a similar image to the reference image when the overall similarity with the reference image among the comparison target images satisfies a predetermined second standard.

In step S707, the duplicate image detection server 130 detects each subject included in the reference image and determines it as the first subject, and detects each subject included in the similar image and determines it as the second subject. In one embodiment, at least one first subject and at least one second subject may be determined from the reference image and the similar image, respectively.

Next, in step S709, the duplicate image detection server 130 calculates the degree of similarity between the first subject and the second subject. In one embodiment of the present disclosure, the duplicate image detection server 130 may first determine the feature vector of each first subject and the feature vector of the second subject. In one embodiment, the duplicate image detection server 130 may extract the category of the first subject and the category of the second subject, and calculate the degree of similarity by comparing the first subject and the second subject that correspond to the same category. In one embodiment of the present disclosure, when a plurality of first subjects and a plurality of second subjects are determined, the similarity of each subject may be calculated, and a weight may be applied to calculate the similarity between the reference image and the local area of the similar image. .

In step S711, the duplicate image detection server 130 may determine the similar image as a duplicate image for the reference image when the similarity between the first subject and the second subject satisfies a predetermined first standard.

Finally, in step S713, the duplicate image detection server 130 may provide the user terminal 110 with information on the duplicate image with respect to the reference image. In one embodiment, the duplicate image detection server 130 may provide information by sorting duplicate images in order of high similarity to the reference image.

In the embodiments of the present disclosure described above with reference to the drawings (and herein generally), the user terminal 110 and the duplicate image detection server 130 use a client-server model, in particular, the client mainly provides only user input and output functions. Most other functions (in particular, many functions related to duplicate image detection) are described as if implemented based on a server-delegated form, but the present disclosure is not limited thereto. According to another embodiment of the present disclosure, it should be noted that the duplicate image detection system environment may be implemented with its functions evenly distributed between the user terminal and the server, or may rather be implemented depending more on the application environment installed on the user terminal. do. In addition, when implementing the functions of the duplicate image detection system by distributing them between the user terminal and the server according to an embodiment of the present disclosure, the distribution of each function of the duplicate image detection system between the client and the server may be implemented differently for each embodiment. You need to know that you can. According to an embodiment of the present disclosure, it should be noted that the main functions of the duplicate image detection server may be implemented and provided in each user terminal 110 rather than the duplicate image detection server 130.

In addition, in the above-described embodiments of the present disclosure, for convenience, a specific module is described as performing certain operations, but the present disclosure is not limited thereto. According to another embodiment of the present disclosure, it should be noted that operations described above as being performed by a specific module may be performed by separate modules.

Programs executed by the terminals and servers described in this disclosure may be implemented with hardware components, software components, and/or a combination of hardware components and software components. A program can be executed by any system that can execute computer-readable instructions.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device. Software may be implemented as a computer program including instructions stored on computer-readable storage media. Computer-readable recording media include, for example, magnetic storage media (e.g., ROM (Read-Only Memory), RAM (Random-Access Memory), floppy disk, hard disk, etc.) and optical read media (e.g., CD-ROM). (CD-ROM), DVD (Digital Versatile Disc), etc. The computer-readable recording medium is distributed among computer systems connected to a network, so that computer-readable code can be stored and executed in a distributed manner. The media may be readable by a computer, stored in memory, and executed by a processor.

Computer-readable storage media may be provided in the form of non-transitory storage media. Here, 'non-transitory' only means that the storage medium does not contain signals and is tangible, and does not distinguish whether the data is stored semi-permanently or temporarily in the storage medium.

Additionally, programs according to embodiments of the present disclosure may be included and provided as a computer program product. Computer programs are commodities and can be traded between sellers and buyers. A computer program may include a software program and a computer-readable storage medium on which the software program is stored. For example, a computer program may be a product in the form of a software program (e.g., a downloadable application) that is distributed electronically by the device manufacturer or through an electronic marketplace (e.g., Google Play Store, App Store). It can be included. For electronic distribution, at least a portion of the software program may be stored on a storage medium or created temporarily. In this case, the storage medium may be a storage medium of a manufacturer's server, an electronic market server, or a relay server that temporarily stores a software program.

A computer program may include a storage medium of a server or a storage medium of a device in a system comprised of a server and a device. Alternatively, if there is a third device (eg, a smartphone) that is connected to the server or device for communication, the computer program may include the storage medium of the third device. Alternatively, the computer program may include a software program itself that is transmitted from a server to the device or a third device, or from a third device to the device. In this case, one of the server, the device, and the third device may execute the computer program to perform the method according to the disclosed embodiments. Alternatively, two or more of a server, a device, and a third device may execute the computer program and perform the methods according to the disclosed embodiments in a distributed manner. For example, the server may execute a computer program stored on the server and control a device connected to the server to perform the method according to the disclosed embodiments. As another example, a third device may execute a computer program to control a device connected to the third device to perform the method according to the disclosed embodiment. When the third device executes a computer program, the third device may download the computer program from the server and execute the downloaded computer program. Alternatively, the third device may perform the method according to the disclosed embodiments by executing a computer program provided in a pre-loaded state.

As described above, although the embodiments have been described with limited examples and drawings, various modifications and variations can be made by those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and/or components, such as a described computer system or module, may be combined or combined in a form different from the described method, or other components or equivalents may be used. Appropriate results can be achieved even if replaced or replaced by .

100: Duplicate image detection system
110: user terminal
120: communication network
130: Duplicate image detection server

Claims (12)

A method for detecting duplicate images containing the same subject, comprising:
Receiving a reference image including at least one first subject,
determining at least one similar image similar to the reference image;
Calculating a degree of similarity between at least one first subject included in the reference image and at least one second subject included in the similar image;
If the similarity satisfies a predetermined first standard, determining the similar image as a duplicate image of the reference image;
Including providing information about the duplicate image,
method. According to paragraph 1,
Determining at least one similar image similar to the reference image includes:
Comparing at least one image included in a predetermined image database with the reference image to calculate overall similarity;
Comprising the step of determining the at least one image as a similar image when the overall similarity satisfies a predetermined second criterion,
method. According to paragraph 1,
Calculating the degree of similarity between at least one first subject included in the reference image and at least one second subject included in the similar image,
Detecting at least one first subject included in the reference image,
detecting at least one second subject included in the similar image; and
Comprising the step of calculating similarity by comparing the at least one first subject and the at least one second subject,
method. According to paragraph 3,
The step of calculating similarity by comparing the at least one first subject and the at least one second subject,
Determine a first category of the at least one first subject, and determine a second category of the at least one second subject, so that each of the at least one first subject and the at least one second subject is Calculating the similarity by comparing a first subject and a second subject whose first category and the second category are the same,
method. According to paragraph 3,
The step of calculating similarity by comparing the at least one first subject and the at least one second subject,
Determining a first subject located in the center of the reference image among the at least one first subject, and determining a second subject located in the center of the similar image among the at least one second subject, wherein the first subject and the Calculating the similarity by comparing a second subject,
method. According to clause 5,
The first subject located in the center of the reference image is determined to be the first subject whose distance between the center point of the reference image and the center point of the at least one first subject is closest,
The second subject located in the center of the similar image is determined as the second subject in which the distance between the center point of the similar image and the center point of the at least one second subject is closest,
method. According to clause 4,
The similarity is calculated by comparing a first subject and a second subject among the at least one first subject and the at least one second subject whose first and second categories are the same, and wherein the subject Determined by weighting the similarity,
method. According to paragraph 2,
The step of calculating the overall similarity is,
Calculate a first similarity between the center of the reference image and the center of the at least one image, calculate a second similarity between the periphery of the reference image and the periphery of the at least one image, and calculate the first similarity and the second similarity. Calculating the overall similarity by assigning weights to each similarity,
method. According to paragraph 2,
The step of calculating the overall similarity is,
Calculating similarity by comparing each of the at least one first subject included in the reference image and the at least one second subject included in the similar image, and selecting the closest similarity among them as the first similarity,
Calculating a second similarity by comparing a first subject located in the center of the reference image and a second subject located in the center of the similar image,
Calculating the total similarity by assigning a predetermined weight to the first similarity and the second similarity,
method. According to paragraph 1,
In the step of providing information about the duplicate image,
Sorting based on the similarity to provide information about the duplicate images,
method. A non-transitory computer-readable recording medium recording a computer program for executing the method according to claims 1 to 10. A duplicate image detection system including the same subject, comprising:
a reference image receiving unit that receives a reference image including at least one first subject;
a similar image determination unit that determines at least one similar image similar to the reference image;
Calculate the similarity between at least one first subject included in the reference image and at least one second subject included in the similar image, and when the similarity satisfies a predetermined first standard, the similar image is converted to the reference image. A duplicate image determination unit that determines a duplicate image of
Comprising a duplicate image information provider that provides information about the duplicate image,
system.

Images