KR20230065750A - Method, system, and computer program for classify place review images based on deep learning - Google Patents

Abstract

Disclosed are a method, system, and computer program for classifying place review images based on deep learning. A method for classifying place review images includes the steps of: receiving images submitted as reviews for a place; and classifying the image based on information related to the place from results of analyzing the image based on the deep learning, wherein images submitted for the review can be classified into high-quality images and low-quality images through a deep learning model learned with an object-centered learning data set related to the place and rule-based logic which serves as an informative standard for the place.

Description

Method, system, and computer program for classifying place review images based on deep learning {METHOD, SYSTEM, AND COMPUTER PROGRAM FOR CLASSIFY PLACE REVIEW IMAGES BASED ON DEEP LEARNING}

The description below relates to a technique for classifying images submitted for place review.

In the place review service, after authenticating a place visited or used by a user by using a place reservation function, a receipt authentication function, etc., the authenticated place information can be submitted together with the review.

As an example, Korean Patent Publication No. 10-2020-0020457 (published on February 26, 2020) discloses a technique for proving a user's visit confirmation based on a receipt.

A user review of a place may be exposed in conjunction with a place review service as well as other services providing place information such as a search service or a map service.

Images submitted for place review can be classified into low-quality images and high-quality images in terms of information about the place.

Place review images can be classified through a deep learning model using various training data sets and rule-based logic.

A review image classification method executed on a computer system, wherein the computer system includes at least one processor configured to execute computer readable instructions contained in a memory, and the review image classification method is performed by the at least one processor. , receiving an image submitted as a review for a place; and classifying, by the at least one processor, the image based on information related to the place from a result of analyzing the image based on deep learning.

According to one aspect, the step of classifying is submitted to the review through a deep learning model learned with an object-oriented learning data set related to the place and a rule-based logic that serves as a criterion for information about the place Classifying the processed image into a high-quality image and a low-quality image.

According to another aspect, the classifying may include classifying an image submitted to the review including at least one object related to the place as a high-quality image and classifying the remaining images as low-quality images. .

According to another aspect, the classifying may include classifying the image into a low-quality blacklist when the image is classified as a receipt image through a deep learning-based image classification model. .

According to another aspect, the classifying may include classifying the image into a low-quality blacklist when a ratio occupied by a human object detected in the image in an entire area is greater than or equal to a threshold value through an object detection model. can include

According to another aspect, the classifying may include classifying the image into a low-quality blacklist when an object prediction probability value of a face object detected in the image through an object detection model is greater than or equal to a threshold value.

According to another aspect, the classifying may include classifying blurry images among images submitted to the review into a low-quality blacklist through blurry image detection logic.

According to another aspect, the classifying may include classifying a screenshot image among images submitted to the review into a low-quality blacklist through a screenshot image detection logic.

According to another aspect, the review image classification method may further include, by the at least one processor, limiting exposure of the low-quality image when exposing a review of the place.

According to another aspect, the review image classification method may include, by the at least one processor, at least one of a review including the high-quality image of the place or a reviewer submitting the high-quality image as a review of the place. A recommendation step may be further included.

A computer program stored in a computer readable recording medium is provided to execute the review image classification method on a computer.

A computer system comprising at least one processor configured to execute computer readable instructions contained in a memory, wherein the at least one processor receives an image submitted as a review for a place and, based on deep learning, the image Provided is a computer system characterized in that the image is classified based on the information related to the place in the analysis result.

According to the embodiments of the present invention, review quality is managed by classifying images submitted for place reviews into low-quality images and high-quality images in terms of information about places through a deep learning model using various learning data sets and rule-based logic. Through this, the recommendation performance for reviews or reviewers can be improved.

1 is a diagram illustrating an example of a network environment according to an embodiment of the present invention.
2 is a block diagram illustrating an example of a computer system according to one embodiment of the present invention.
3 is a diagram showing an example of components that may be included in a processor of a computer system according to an embodiment of the present invention.
4 is a flowchart illustrating an example of a method that a computer system may perform according to an embodiment of the present invention.
5 to 9 are exemplary diagrams for explaining a process of classifying a place review image according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to techniques for classifying images submitted for placement reviews.

Embodiments including those specifically disclosed in this specification may classify images submitted for place review into low-quality images and high-quality images in terms of information about the corresponding place.

In the present specification, a place may mean all objects that can review user experiences according to visits or use, such as restaurants, shops, famous places, and hot places. A review of a place may include a review of an online user experience, such as a commercial transaction or a virtual performance/exhibition, as well as an offline user experience.

The review image classification system according to embodiments of the present invention may be implemented by at least one computer system, and the review image classification method according to embodiments of the present invention may include at least one computer system included in the review image classification system. can be performed through At this time, a computer program according to an embodiment of the present invention may be installed and driven in the computer system, and the computer system may perform the review image classification method according to the embodiments of the present invention under the control of the driven computer program. there is. The above-described computer program may be combined with a computer system and stored in a computer readable recording medium to execute a review image classification method on a computer.

1 is a diagram illustrating an example of a network environment according to an embodiment of the present invention. The network environment of FIG. 1 shows an example including a plurality of electronic devices 110 , 120 , 130 , and 140 , a plurality of servers 150 and 160 , and a network 170 . 1 is an example for explanation of the invention, and the number of electronic devices or servers is not limited as shown in FIG. 1 . In addition, the network environment of FIG. 1 only describes one example of environments applicable to the present embodiments, and the environment applicable to the present embodiments is not limited to the network environment of FIG. 1 .

The plurality of electronic devices 110, 120, 130, and 140 may be fixed terminals implemented as computer devices or mobile terminals. Examples of the plurality of electronic devices 110, 120, 130, and 140 include a smart phone, a mobile phone, a navigation device, a computer, a laptop computer, a digital broadcast terminal, a personal digital assistant (PDA), and a portable multimedia player (PMP). ), tablet PC, game console, wearable device, internet of things (IoT) device, virtual reality (VR) device, augmented reality (AR) device, and the like. As an example, FIG. 1 shows the shape of a smartphone as an example of the electronic device 110, but in the embodiments of the present invention, the electronic device 110 substantially uses a wireless or wired communication method to transmit other information via the network 170. It may refer to one of various physical computer systems capable of communicating with the electronic devices 120 , 130 , and 140 and/or the servers 150 and 160 .

The communication method is not limited, and short-distance wireless communication between devices as well as a communication method utilizing a communication network (eg, a mobile communication network, a wired Internet, a wireless Internet, and a broadcasting network) that the network 170 may include may also be included. For example, the network 170 may include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , one or more arbitrary networks such as the Internet. In addition, the network 170 may include any one or more of network topologies including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, and the like. Not limited.

Each of the servers 150 and 160 communicates with the plurality of electronic devices 110, 120, 130, and 140 through the network 170 to provide commands, codes, files, contents, services, and the like, or a computer device or a plurality of computers. Can be implemented in devices. For example, the server 150 may be a system that provides a first service to a plurality of electronic devices 110, 120, 130, and 140 accessed through the network 170, and the server 160 may also include a network ( It may be a system that provides a second service to a plurality of electronic devices 110, 120, 130, and 140 accessed through 170). As a more specific example, the server 150 provides a service targeted by the application (eg, a place review service) through an application as a computer program that is installed and driven in the plurality of electronic devices 110, 120, 130, and 140. may be provided to the plurality of electronic devices 110, 120, 130, and 140 as the first service. As another example, the server 160 may provide a service for distributing files for installing and running the above-described application to the plurality of electronic devices 110, 120, 130, and 140 as a second service.

2 is a block diagram illustrating an example of a computer system according to one embodiment of the present invention. Each of the plurality of electronic devices 110 , 120 , 130 , and 140 or each of the servers 150 and 160 described above may be implemented by the computer system 200 shown in FIG. 2 .

As shown in FIG. 2 , the computer system 200 may include a memory 210, a processor 220, a communication interface 230, and an input/output interface 240.

The memory 210 is a computer-readable recording medium and may include a random access memory (RAM), a read only memory (ROM), and a permanent mass storage device such as a disk drive. Here, a non-perishable mass storage device such as a ROM and a disk drive may be included in the computer system 200 as a separate permanent storage device distinct from the memory 210 . Also, an operating system and at least one program code may be stored in the memory 210 . These software components may be loaded into the memory 210 from a computer-readable recording medium separate from the memory 210 . The separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD/CD-ROM drive, and a memory card. In another embodiment, software components may be loaded into the memory 210 through the communication interface 230 rather than a computer-readable recording medium. For example, software components may be loaded into memory 210 of computer system 200 based on a computer program installed by files received over network 170 .

The processor 220 may be configured to process commands of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 220 by memory 210 or communication interface 230 . For example, processor 220 may be configured to execute received instructions according to program codes stored in a recording device such as memory 210 .

The communication interface 230 may provide a function for the computer system 200 to communicate with other devices (eg, storage devices described above) through the network 170 . For example, a request, command, data, file, etc. generated according to a program code stored in a recording device such as the memory 210 by the processor 220 of the computer system 200 is controlled by the communication interface 230 to the network ( 170) to other devices. Conversely, signals, commands, data, files, etc. from other devices may be received into computer system 200 via communication interface 230 of computer system 200 via network 170 . Signals, commands, or data received through the communication interface 230 may be transferred to the processor 220 or the memory 210, and files may be stored as storage media that the computer system 200 may further include (described above). permanent storage).

The input/output interface 240 may be a means for interface with the input/output device 250 . For example, the input device may include a device such as a microphone, keyboard, or mouse, and the output device may include a device such as a display or speaker. As another example, the input/output interface 240 may be a means for interface with a device in which functions for input and output are integrated into one, such as a touch screen. The input/output device 250 may be configured as one device with the computer system 200 .

Also, in other embodiments, computer system 200 may include fewer or more elements than those of FIG. 2 . However, there is no need to clearly show most of the prior art components. For example, the computer system 200 may be implemented to include at least some of the aforementioned input/output devices 250 or may further include other elements such as a transceiver and a database.

Hereinafter, specific embodiments of a method and system for classifying place review images based on deep learning will be described.

3 is a block diagram showing an example of components that may be included in a processor of a computer system according to an embodiment of the present invention, and FIG. 4 is a review that a computer system according to an embodiment of the present invention may perform. It is a flowchart illustrating an example of an image classification method.

The computer system 200 according to the present embodiment may provide a place review service to clients through a dedicated application installed on the client or access to a web/mobile site related to the computer system 200 .

The processor 220 of the computer system 200 may include an image classification unit 310 and a review providing unit 320 as shown in FIG. 3 as components for performing a review image classification method to be described later. can Depending on embodiments, components of the processor 220 may be selectively included in or excluded from the processor 220 . Also, components of the processor 220 may be separated or merged to express functions of the processor 220 according to embodiments.

The processor 220 and components of the processor 220 may control the computer system 200 to perform steps included in a review image classification method to be described later. For example, the processor 220 and components of the processor 220 may be implemented to execute instructions according to an operating system code and at least one program code included in the memory 210 .

Here, elements of the processor 220 may be representations of different functions performed by the processor 220 according to instructions provided by program codes stored in the computer system 200 . For example, image classification unit 310 may be used as a functional representation of processor 220 that controls computer system 200 according to the instructions described above so that computer system 200 classifies images submitted for place reviews. there is.

The processor 220 may read necessary commands from the memory 210 loaded with commands related to the control of the computer system 200 . In this case, the read command may include a command for controlling the processor 220 to execute a review image classification method to be described later.

Steps included in the review image classification method to be described later may be performed in an order different from the order shown, and some of the steps may be omitted or additional processes may be further included.

Referring to FIG. 4 , in step S410, the image classification unit 310 receives the image submitted as a place review (hereinafter, referred to as a 'place review image') and converts the place review image into a high-quality product in terms of information about the place. It can be classified into images and low-quality images. A high-quality image is an image rich in information and may mean an image including an effective object related to a place. On the other hand, a low-quality image may mean an image lacking information about a place. For example, in the case of a restaurant, a high-quality image can be an image that clearly reveals the food and information of the restaurant. If food is essential and a table, chair, plate, etc. appear together with the food, the information about the restaurant is good. This rich image can be defined. On the other hand, a blurry image, an image captured from content on the Internet, an image with only one cup, and an image in which a person occupies a large portion may be defined as low-quality images in terms of information. The image classification unit 310 may classify the place review image into a high-quality image and a low-quality image through a deep learning model using various training data sets and a rule-based logic serving as a criterion for information about a place. The image classification unit 310 may classify an image including an effective object as a high-quality image and classify the remaining images as low-quality images based on a result of image analysis using a deep learning model for the place review image. To this end, the image classification unit 310 may include a function of detecting whether the place review image is a receipt image, a function of detecting an object in the place review image, a function of inspecting the image quality of the place review image, and the like. . The image classification unit 310 may blacklist a reviewer who submitted a low-quality image, a review including a low-quality image, or a reviewer who submitted a low-quality image as a place review, and may classify a high-quality image, a review including a high-quality image, or a high-quality image. A reviewer who submitted an image as a place review can manage it by assigning a separate tag (eg, good, etc.).

In addition to classifying place review images based on informativeness, image classification by food menu, image classification by atmosphere, and image classification using an index representing image sharpness can be applied.

In step S420, the review providing unit 320 may limit the exposure of low-quality images in providing place reviews and provide recommendation information on places based on high-quality images. For example, the review provider 320 may recommend at least one of a review including a high-quality image as a place review or a reviewer who submitted a high-quality image as a place review. In the state where place review images are classified into high-quality images and low-quality images in terms of information about places, the exposure of reviews or reviewers using low-quality images can be limited, and recommendations can be made focusing on reviews or reviewers using high-quality images. Restricting the exposure of low-quality images may mean reducing exposure opportunities compared to high-quality images or excluding them from exposure.

5 to 9 are exemplary diagrams for explaining a process of classifying a place review image according to an embodiment of the present invention.

Referring to FIG. 5 , in step S501, the image classification unit 310 may classify a receipt image for a place review image by using an image classification model. The image classification unit 310 determines three classes of images, for example, paper receipts, e-receipts, and not receipts based on deep learning. Through a multi-class classification model that can be used, it is possible to classify a place review image to which class it belongs to. Depending on the embodiment, it is also possible to apply an image classification model capable of discriminating images into two classes (receipts and non-receipts).

In step S502, the image classification unit 310 may determine whether the place review image is an image of a receipt or an image unrelated to the receipt based on the classification result using the image classification model.

In step S503, if the place review image is determined to be a receipt image as a result of the determination in step S502, the image classification unit 310 may filter the place review image into a receipt class and classify the place review image into a low quality image blacklist.

Referring to FIG. 6, in step S604, the image classification unit 310 uses an object detection model when the place review image is determined to be an image unrelated to the receipt as a result of the determination in step S502. Objects in the review image may be detected. The image classification unit 310 may recognize objects included in the place review image through an object detection model capable of detecting various objects related to a place to be reviewed based on deep learning. The object detection model uses a variety of learning datasets, such as COCO dataset, WIDER FACE dataset, UEC Food 256 dataset, etc. It corresponds to a deep learning model trained with a training data set.

In step S605, the image classification unit 310 may determine whether at least one valid object related to a place is detected in the place review image. For example, when a place review service is provided to restaurants, effective objects may include food, cups, plates, tables, chairs, menu boards, and the like.

In step S606, the image classification unit 310 filters the place review image into an unknown target class and classifies it into a low-quality image blacklist when valid objects are not detected in the place review image as a result of the determination in step S605. can

In step S607, the image classification unit 310 may classify the place review image into a high-quality image list when a valid object is detected in the place review image as a result of the determination in step S605.

For example, in the case of a restaurant, referring to FIG. 7 , in step S71, the image classification unit 310 determines whether a human object having an object area equal to or greater than a threshold value is detected in the place review image, for example, 30% of the entire image. It may be determined whether an occupied human object is detected.

In step S72, the image classification unit 310 filters the place review image into a person class and classifies it into a low-quality image blacklist when a human object having an object area equal to or greater than a threshold value is detected in the place review image as a result of the determination in step S71. can

Many images featuring visitors are submitted for place reviews, but if human objects occupy most of the images, they cannot be viewed as good images in terms of information. In consideration of this point, the image classification unit 310 may classify a place review image having a human object region equal to or greater than a threshold value as a low quality image.

In step S73, the image classification unit 310 determines that, as a result of the determination in step S71, if a human object is not detected in the place review image or a human object whose object area is less than the threshold value is detected, the object prediction probability value in the place review image is It may be determined whether a face object having a threshold value or higher is detected, for example, whether a face object corresponding to a probability value of 0.75 or higher is detected. In addition to object prediction probability, object size and object ratio can be used for image classification.

The image classification unit 310 may filter the place review image into a person class and classify it into a low-quality image blacklist when a facial object having an object prediction probability value equal to or greater than a threshold value is detected in the place review image as a result of the determination in step S73.

If an image of the face of a person other than a visitor is submitted as a place review and the review is exposed on the service, there may be a problem with portrait rights. In consideration of this point, the image classification unit 310 may classify a place review image having a probability value of a face object equal to or greater than a threshold value as a low quality image.

Referring to FIG. 8 , in step S84, the image classification unit 310, as a result of the determination in step S73, determines whether a face object is detected in the place review image or a face object whose object prediction probability value is less than a threshold value is detected. In the review image, it may be determined whether a cup object having an object area equal to or greater than a threshold value is detected, for example, whether a cup object occupying 30% of the entire image is detected.

In step S85, the image classification unit 310 filters the place review image into a cup class and classifies it into a blacklist of low quality images when a cup object having an object area equal to or greater than a threshold value is detected in the place review image as a result of the determination in step S84. can

In the case of cafe reviews, images containing cups are mainly submitted, and cups can be a criterion for a good image given the nature of a place called a cafe. However, if the cup object occupies most of the image, it cannot be regarded as a good image in terms of information. Therefore, in consideration of this point, the image classification unit 310 may classify a place review image in which the cup object area is greater than a threshold value as a low quality image. can

In step S86, the image classification unit 310 determines at least one food object in the place review image when no cup object is detected in the place review image or a cup object whose object area is less than a threshold is detected as a result of the determination in step S84. It can be determined whether or not is detected.

In step S87, the image classification unit 310 may filter the place review image into an unknown target class and classify it into a low-quality image blacklist when no food object is detected in the place review image as a result of the determination in step S86.

In step S88, the image classification unit 310 may classify the place review image into a high-quality image list through separate tagging when a food object is detected in the place review image as a result of the determination in step S86.

In the process of classifying the place review image, image information can be determined through a method of sequentially detecting objects included in the image, and depending on the embodiment, a method of simultaneously detecting a plurality of objects included in the image It is also possible to judge informativeness.

The present embodiments utilize an image classification model and an object detection model to determine image information, as well as a blurry image detection logic and a screenshot image as an example of logic for inspecting the quality of the image itself ( screenshot image) At least one of the detection logics can be additionally utilized.

Referring to FIG. 9 , in step S901 , the image classification unit 310 may perform a blur check on the place review image through blur image detection logic. The image classification unit 310 may detect a blurry image based on a Laplacian variance.

In step S902, the image classification unit 310 may determine whether the place review image corresponds to a blurred image according to the Laplacian variance value as a result of the blur check in step S901.

In step S903, the image classification unit 310 may filter the place review image into a blur image class and classify it into a low quality image blacklist if the place review image corresponds to a blurry image as a result of the determination in step S902.

In step S904, if the place review image does not correspond to a blurry image as a result of the determination in step S902, the image classification unit 310 may perform a screen shot check on the place review image.

Due to the nature of the screenshot, the upper region (eg, 20% of the entire image) and the lower region (eg, 20% of the entire image) of the entire image are black or white. After resizing the place review image to a certain size (eg, 128×128) using this point, if the same pixel value in the upper and lower areas exceeds a certain percentage (eg, 30%), the screen is displayed. It can be judged by the shot image.

In step S905, the image classification unit 310 may determine whether the place review image corresponds to the screenshot image as a result of the screen shot inspection in step S904.

In step S906, the image classification unit 310 may filter the place review image into a screenshot image class and classify it into a low-quality image blacklist if the place review image corresponds to the screenshot image as a result of the determination in step S905. .

In step S907, the image classification unit 310 may classify the place review image into a high-quality image list through separate tagging when the place review image does not correspond to the screenshot image as a result of the determination in step S905.

The blur image detection process and the screenshot image detection process may be performed at the last step of the image classification process, and depending on the embodiment, an initial step (eg, before step S501) or an intermediate step (eg, before step S501). , before step S604, before step S86, etc.).

In this embodiment, various deep learning models and rule-based logics may be used to classify place review images based on place-related information.

As such, according to the embodiments of the present invention, review quality is improved by classifying images submitted for place reviews into low-quality images and high-quality images in terms of information about places through a deep learning model using various learning data sets and rule-based logic. Management can improve the performance of reviews or recommendations to reviewers.

The devices described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in the embodiments include a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), and a programmable PLU (programmable logic unit). logic unit), microprocessor, or any other device capable of executing and responding to instructions. The processing device may run an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. You can command the device. The software and/or data may be embodied in any tangible machine, component, physical device, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. there is. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. In this case, the medium may continuously store a program executable by a computer or temporarily store the program for execution or download. In addition, the medium may be various recording means or storage means in the form of a single or combined hardware, but is not limited to a medium directly connected to a certain computer system, and may be distributed on a network. Examples of the medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, and ROM, RAM, flash memory, etc. configured to store program instructions. In addition, examples of other media include recording media or storage media managed by an app store that distributes applications, a site that supplies or distributes various other software, and a server.

As described above, although the embodiments have been described with limited examples and drawings, those skilled in the art can make various modifications and variations from the above description. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims (20)

In the review image classification method executed in a computer system,
The computer system includes at least one processor configured to execute computer readable instructions contained in a memory;
The review image classification method,
receiving, by the at least one processor, an image submitted as a review for a place; and
Classifying, by the at least one processor, the image based on information related to the place from a result of analyzing the image based on deep learning;
A review image classification method comprising a. According to claim 1,
The classification step is
Categorizing images submitted to the review into high-quality images and low-quality images through a deep learning model learned with an object-oriented learning data set related to the place and a rule-based logic that serves as an informative criterion for the place step
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying an image submitted to the review, including at least one object related to the place, as a high-quality image and classifying the remaining images as low-quality images.
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying the image into a low-quality blacklist when the image is classified as a receipt image through a deep learning-based image classification model
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying the image into a low-quality blacklist when a ratio occupied by a human object detected in the image through an object detection model in the entire area is greater than or equal to a threshold value;
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying the image into a low-quality blacklist when an object prediction probability value of a facial object detected in the image through an object detection model is greater than or equal to a threshold value
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying blur images among the images submitted to the review into a low-quality blacklist through blur image detection logic
A review image classification method comprising a. According to claim 1,
The classification step is
Classifying screenshot images among the images submitted to the review into a low-quality blacklist through a screenshot image detection logic.
A review image classification method comprising a. According to claim 2,
The review image classification method,
limiting, by the at least one processor, the exposure of the low quality image in exposing the review of the place;
Review image classification method further comprising a. According to claim 2,
The review image classification method,
recommending, by the at least one processor, at least one of a review of the place including the high-quality image or a reviewer who submitted the high-quality image as a review of the place;
Review image classification method further comprising a. A computer program stored in a computer-readable recording medium to execute the review image classification method according to any one of claims 1 to 10 on a computer. In a computer system,
at least one processor configured to execute computer readable instructions contained in memory;
including,
The at least one processor,
Receive images submitted as reviews for places;
Classifying the image based on information related to the place in the result of analyzing the image based on deep learning
Characterized by a computer system. According to claim 12,
The at least one processor,
Classifying images submitted to the review into high-quality images and low-quality images through a deep learning model learned with an object-oriented learning data set related to the place and a rule-based logic that serves as an informative criterion for the place
Characterized by a computer system. According to claim 12,
The at least one processor,
Classifying an image submitted to the review, including at least one object related to the place, as a high-quality image and classifying the remaining images as low-quality images.
Characterized by a computer system. According to claim 12,
The classification step is
Classifying the image as a low-quality blacklist when the image is classified as a receipt image through a deep learning-based image classification model
Characterized by a computer system. According to claim 12,
The classification step is
Classifying the image as a low-quality blacklist when the ratio of the human object detected in the image through the object detection model to the entire area is greater than or equal to a threshold value
Characterized by a computer system. According to claim 12,
The at least one processor,
Classifying the image into a low-quality blacklist when an object prediction probability value of a face object detected in the image through an object detection model is greater than or equal to a threshold value
Characterized by a computer system. According to claim 12,
The at least one processor,
Classifying blur images and screenshot images among images submitted to the review into a low-quality blacklist through blur image detection logic and screenshot image detection logic
Characterized by a computer system. According to claim 13,
The at least one processor,
Limiting the exposure of the low-quality image in exposing the review of the place
Characterized by a computer system. According to claim 13,
The at least one processor,
recommending at least one of a review of the place containing the high-quality image or a reviewer who submitted the high-quality image as a review of the place;
Characterized by a computer system.

Images