KR102556544B1 - Method, apparatus and computer program for improving recognition ratio of artificial intelligence model through image histogram analysis and recognizing object using artificial intelligence model with improved recognition ratio - Google Patents

Abstract

An object recognition method, device, and computer program using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis are provided. In accordance with various embodiments of the present invention, the recognition rate of an artificial intelligence model through image histogram analysis is improved and an object recognition method using an artificial intelligence model with an improved recognition rate is performed by a computing device, which includes performing histogram analysis on an image. Step, based on a result of performing histogram analysis on the image, extracting one or more image layers from the image, and recognizing an object included in the image by analyzing the extracted one or more image layers. .

Description

Recognition rate improvement of artificial intelligence model through image histogram analysis and object recognition method, device and computer program using artificial intelligence model with improved recognition rate USING ARTIFICIAL INTELLIGENCE MODEL WITH IMPROVED RECOGNITION RATIO}

Various embodiments of the present invention relate to an object recognition method, apparatus, and computer program using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis.

Recently, research on detecting or recognizing an object in an image captured by a camera by applying technology using artificial intelligence and big data to a device equipped with a camera is being conducted. For example, the artificial intelligence-based object recognizer can be applied to devices equipped with cameras, such as self-driving cars, surveillance cameras, and drones.

When this AI-based object recognizer recognizes an object in an image captured by a camera at a recognition rate higher than a predetermined level, devices equipped with such a camera and object recognizer provide services such as autonomous driving based on the recognized object. it is possible

Apart from the existing business environment, the need for robots, self-driving cars, drones, etc. to identify images and read texts is emerging as people recognize in daily life.

In particular, as images to be recognized and processed are gradually upgraded (eg, high resolution, high capacity), there is a demand for designing and developing an optimal algorithm capable of quickly and accurately processing high-resolution and high-capacity images.

The problem to be solved by the present invention is to analyze the image to recognize the object included in the image, extract a plurality of image layouts from the image based on the histogram analysis result of the image, and use the extracted plurality of image layouts to recognize the object To provide an object recognition method, device, and computer program using an artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis that can more quickly and accurately recognize objects from high-resolution and high-capacity images by recognizing will be.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

An object recognition method using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to an embodiment of the present invention for solving the above problems is a method performed by a computing device, wherein the image Performing a histogram analysis on , extracting one or more image layers from the image based on a result of performing the histogram analysis on the image, and analyzing the extracted one or more image layers to object included in the image. It may include the step of recognizing.

In various embodiments, the extracting of one or more image layers may include a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values. and sequentially selecting N pixel values from among the plurality of pixel values, starting with the largest number of pixels, and extracting a plurality of image layers from the image based on the selected N pixel values. can do.

In various embodiments, the step of extracting one or more image layers may include a histogram graph generated based on a result of performing histogram analysis on the image - the histogram graph is based on a plurality of pixel values corresponding to a plurality of colors, respectively. Including a graph of the number of pixels - Calculate the sum of the number of pixels of the plurality of colors for each of the plurality of pixel values using , sequentially from the pixel value of which the calculated sum of the number of pixels is the largest among the plurality of pixel values. and extracting a plurality of image layers from the image based on the selected N pixel values.

In various embodiments, the performing of the histogram analysis on the image may include performing the histogram analysis on the image to obtain a first histogram graph - the first histogram graph according to a plurality of pixel values corresponding to a plurality of colors, respectively. A second histogram graph comprising a pixel number graph - generating a gray scale image by converting the image to gray scale, and performing histogram analysis on the generated gray scale image - the second histogram graph is including a graph of the number of pixels according to a plurality of grayscale values corresponding to one color; and the step of extracting one or more image layers includes the step of extracting the one or more image layers using the generated second histogram graph. Selecting N grayscale values sequentially from the grayscale value having the largest number of pixels from among the plurality of grayscale values, and using the generated first histogram graph, the selected N grayscale values The method may include selecting corresponding N pixel values and extracting a plurality of image layers from the image based on the selected N pixel values.

In various embodiments, the step of extracting one or more image layers may include a histogram graph generated based on a result of performing histogram analysis on the image - the histogram graph is based on a plurality of pixel values corresponding to a plurality of colors, respectively. Including a graph of the number of pixels - Calculate the difference between the maximum value and the minimum value of the number of pixels for each of the plurality of pixel values by using, and at least one of the plurality of pixel values in which the calculated difference is equal to or greater than a preset value. The method may include selecting a pixel value and extracting a plurality of image layers from the image based on the selected at least one pixel value.

In various embodiments, recognizing an object included in the image may include, when a plurality of image layers are extracted from the image, generating a single integrated image layer by merging the extracted plurality of image layers; The method may include extracting one or more feature values of the object by analyzing the generated integrated image layer, and recognizing the object using the extracted one or more feature values.

In various embodiments, recognizing an object included in the image may include, when a plurality of image layers are extracted from the image, one or more characteristic values of the object by analyzing each of the plurality of extracted image layers. The method may include extracting, individually estimating the object using one or more extracted feature values, and recognizing the object using information about the individually estimated object.

In various embodiments, recognizing an object included in the image may include, when a plurality of image layers are extracted from the image, extracting a plurality of feature values of the object by analyzing each of the plurality of extracted image layers. and recognizing the object by using only at least one feature value extracted in duplicate from two or more of the plurality of extracted feature values.

In various embodiments, recognizing an object included in the image may include, when a plurality of image layers are extracted from the image, analyzing each of the extracted plurality of image layers to extract one or more feature values, respectively; Comparing one or more extracted feature values to calculate a similarity between the extracted plurality of image layers, merging two or more image layers among the extracted plurality of image layers based on the calculated similarity, and merging the extracted image layers. and recognizing the object by analyzing two or more image layers.

In various embodiments, the performing of histogram analysis on the image may include generating a grayscale image by converting the image to grayscale, and performing a histogram analysis on the generated grayscale image to generate a second histogram. Performing correction on the image based on a graph - the second histogram graph including a graph of the number of pixels according to a plurality of grayscale values - and performing a histogram analysis on the corrected image to obtain a first histogram Generating a graph, and performing correction on the image, wherein a minimum gray scale value of the second histogram graph is greater than or equal to a reference minimum value or a maximum gray scale value of the second histogram graph is less than or equal to a reference maximum value. If the difference between the maximum value and the minimum value of the gray scale of the second histogram graph is equal to or less than a first reference difference value or greater than or equal to a second reference difference value, performing brightness correction on the image, contrast correction on the image It may include the step of performing.

In various embodiments, performing the histogram analysis on the image may include identifying an object region included in the image, extracting the identified object region from the image, and performing histogram analysis on the extracted object region. It may include the step of performing.

In various embodiments, performing the histogram analysis on the image may include identifying a background region included in the image, generating a mask layer corresponding to the identified background region, and using the created mask layer as the It may include combining images and performing histogram analysis on the image combined with the generated mask layer.

In order to solve the above problems, an object recognition apparatus using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to another embodiment of the present invention is loaded into a processor, a network interface, a memory, and the memory. (load) and includes a computer program executed by the processor, wherein the computer program includes instructions for performing histogram analysis on an image, based on a result of performing histogram analysis on the image, the image It may include an instruction for extracting one or more image layers from and an instruction for recognizing an object included in the image by analyzing the extracted one or more image layers.

A computer program according to another embodiment of the present invention for solving the above problems is combined with a computing device, performing a histogram analysis on an image, based on a result of performing the histogram analysis on the image, the image To execute an object recognition method using image histogram analysis, which includes extracting one or more image layers from and recognizing an object included in the image by analyzing the extracted one or more image layers, which can be read by a computing device. It can be stored on a recording medium.

Other specific details of the invention are included in the detailed description and drawings.

According to various embodiments of the present invention, an object included in the image is recognized by analyzing the image, a plurality of image layouts are extracted from the image based on a histogram analysis result of the image, and the plurality of image layouts are extracted. By recognizing the object, there is an advantage in that the object can be more quickly and accurately recognized from a high-resolution and high-capacity image.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram illustrating an object recognition system using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to an embodiment of the present invention.
2 is a hardware configuration diagram of an object recognition apparatus using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to another embodiment of the present invention.
3 is a flow chart of an object recognition method using an artificial intelligence model with improved recognition rate and improved recognition rate of an artificial intelligence model through image histogram analysis according to another embodiment of the present invention.
4 is a diagram illustrating a process of extracting, selecting, and merging image layouts from images, according to various embodiments.
5 is a diagram illustrating an example of a histogram graph generated as a result of performing histogram analysis on an image, in various embodiments.
6 is a flowchart illustrating a method of extracting an image layer from an image, in various embodiments.
7 is a diagram exemplarily illustrating a histogram graph generated by performing histogram analysis on a grayscale converted image, in various embodiments.

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, only these embodiments are intended to complete the disclosure of the present invention, and are common in the art to which the present invention belongs. It is provided to fully inform the person skilled in the art of the scope of the invention, and the invention is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements. Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

The term "unit" or "module" used in the specification means a hardware component such as software, FPGA or ASIC, and "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. A “unit” or “module” may be configured to reside in an addressable storage medium and may be configured to reproduce one or more processors. Thus, as an example, a “unit” or “module” may refer to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and "units" or "modules" may be combined into smaller numbers of components and "units" or "modules" or may be combined into additional components and "units" or "modules". can be further separated.

The spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe a component's correlation with other components. Spatially relative terms should be understood as including different orientations of elements in use or operation in addition to the orientations shown in the drawings. For example, if you flip a component that is shown in a drawing, a component described as "below" or "beneath" another component will be placed "above" the other component. can Thus, the exemplary term “below” may include directions of both below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

In this specification, a computer means any kind of hardware device including at least one processor, and may be understood as encompassing a software configuration operating in a corresponding hardware device according to an embodiment. For example, a computer may be understood as including a smartphone, a tablet PC, a desktop computer, a laptop computer, and user clients and applications running on each device, but is not limited thereto.

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

Although each step described in this specification is described as being performed by a computer, the subject of each step is not limited thereto, and at least a part of each step may be performed in different devices according to embodiments.

1 is a diagram illustrating an object recognition system using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to an embodiment of the present invention.

Referring to FIG. 1 , an object recognition system using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to an embodiment of the present invention includes an object recognition device 100 and a user terminal 200 , an external server 300 and a network 400.

Here, the recognition rate of the artificial intelligence model through image histogram analysis shown in FIG. 1 is improved and the object recognition system using the artificial intelligence model with improved recognition rate is according to an embodiment, and its components are in the embodiment shown in FIG. It is not limited and may be added, changed or deleted as needed.

In one embodiment, the object recognizing device 100 (hereinafter referred to as "computing device 100") may recognize an object included in an image by analyzing a specific image.

Here, recognizing an object included in an image means whether an object is included in the image (existence of an object), and if the object is included in the image, properties of the object (eg, type, size, shape, color, etc. of the object) It may mean to identify, but is not limited thereto.

In various embodiments, the computing device 100 may extract one or more image layers from an image for the purpose of more quickly and accurately recognizing an object from a high-resolution, high-capacity image, and by analyzing the extracted one or more image layers, Objects included in images can be recognized.

In various embodiments, the computing device 100 may recognize an object included in an image by analyzing an image (or one or more image layers extracted from the image) using a pre-learned artificial intelligence model.

Here, the pre-learned artificial intelligence model is a pre-learned model using an image labeled with object information (eg, object attribute information, etc.) as training data, and a specific image (or specific image layer) It may be a model that extracts a recognition result of an object included in a specific image as result data using input data.

An artificial intelligence model (e.g., a neural network) is composed of one or more network functions, which may consist of a set of interconnected computational units, which may generally be referred to as ‘nodes’. These 'nodes' may also be referred to as 'neurons'. One or more network functions include at least one or more nodes. Nodes (or neurons) that make up one or more network functions can be interconnected by one or more 'links'.

In an artificial intelligence model, one or more nodes connected through a link may form a relative relationship of an input node and an output node. The concept of an input node and an output node is relative, and any node in an output node relationship with one node may have an input node relationship with another node, and vice versa. As described above, the input node to output node relationship can be created around the link. More than one output node can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of the output node may be determined based on data input to the input node. Here, a node interconnecting an input node and an output node may have a weight. The weight may be variable, and may be variable by a user or an algorithm in order to perform a function desired by the artificial intelligence model. For example, when one or more input nodes are interconnected by respective links to one output node, the output node is set to a link corresponding to values input to input nodes connected to the output node and respective input nodes. An output node value may be determined based on the weight.

As described above, in the AI model, one or more nodes are interconnected through one or more links to form an input node and output node relationship in the AI model. Characteristics of the artificial intelligence model may be determined according to the number of nodes and links in the artificial intelligence model, the relationship between the nodes and links, and the value of weight assigned to each link. For example, when there are two artificial intelligence models having the same number of nodes and links and different weight values between the links, the two artificial intelligence models may be recognized as different from each other.

Some of the nodes constituting the artificial intelligence model may constitute one layer based on distances from the first input node. For example, a set of nodes having a distance of n from the first input node may constitute n layers. The distance from the first input node may be defined by the minimum number of links that must be passed through to reach the corresponding node from the first input node. However, the definition of such a layer is arbitrary for explanation, and the order of a layer in an artificial intelligence model may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from a final output node.

An initial input node may refer to one or more nodes to which data is directly input without going through a link in relation to other nodes among nodes in the artificial intelligence model. Alternatively, in an artificial intelligence model network, in a relationship between nodes based on a link, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the artificial intelligence model. Also, the hidden node may refer to nodes constituting an artificial intelligence model other than the first input node and the last output node. An artificial intelligence model according to an embodiment of the present invention may have more nodes of an input layer than nodes of a hidden layer close to an output layer, and the number of nodes decreases as the number of nodes increases from the input layer to the hidden layer. can

AI models can include one or more hidden layers. A hidden node of a hidden layer may use outputs of previous layers and outputs of neighboring hidden nodes as inputs. The number of hidden nodes for each hidden layer may be the same or different. The number of nodes of the input layer may be determined based on the number of data fields of the input data and may be the same as or different from the number of hidden nodes. Input data input to the input layer may be operated by a hidden node of the hidden layer and may be output by a fully connected layer (FCL) that is an output layer.

In various embodiments, the artificial intelligence model may be a deep learning model.

A deep learning model (e.g., a deep neural network (DNN)) may refer to an artificial intelligence model including a plurality of hidden layers in addition to an input layer and an output layer. With a deep neural network, data It can identify latent structures, i.e., the latent structures of a photograph, text, video, audio, or music (e.g., what objects are in the photo; content and emotions, etc.).

Deep neural networks include convolutional neural networks (CNN), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted Boltzmann machines (RBMs). boltzmann machine), deep belief network (DBN), Q network, U network, Siamese network, etc., but is not limited thereto.

In various embodiments, the network function may include an autoencoder. Here, the auto encoder may be a type of artificial neural network for outputting output data similar to input data.

An auto-encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between input and output layers. The number of nodes of each layer may be reduced from the number of nodes of the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with the reduction from the bottleneck layer to the output layer (symmetrical to the input layer). Nodes of the dimensionality reduction layer and the dimensionality restoration layer may or may not be symmetrical. Also, an autoencoder can perform non-linear dimensionality reduction. The number of input layers and output layers may correspond to the number of remaining sensors after preprocessing of input data. In the auto-encoder structure, the number of hidden layer nodes included in the encoder may decrease as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, a sufficient amount of information may not be conveyed, so more than a certain number (e.g., more than half of the input layer, etc.) ) may be maintained.

The neural network may be trained using at least one of supervised learning, unsupervised learning, and semi-supervised learning. The learning of neural networks is to minimize errors in the output. More specifically, learning of the neural network repeatedly inputs training data to the neural network, calculates the output of the neural network for the training data and the error of the target, and converts the error of the neural network to the output of the neural network in a direction to reduce the error. This is the process of updating the weight of each node of the neural network by performing backpropagation from the layer to the input layer.

First, in the case of teacher learning, each learning data is labeled with the correct answer (ie, labeled learning data), and in the case of comparative teacher learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of teacher learning regarding data classification may be data in which each learning data is labeled with a category. Labeled training data is input to the neural network, and an error may be calculated by comparing an output (category) of the neural network and a label of the training data.

Next, in the case of comparative history learning for data classification, an error may be calculated by comparing input learning data with a neural network output. The calculated error is back-propagated in a reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back-propagation. The amount of change in the connection weight of each updated node may be determined according to a learning rate. The neural network's computation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of iterations of the learning cycle of the neural network. For example, a high learning rate may be used in the early stage of neural network training to increase efficiency by allowing the neural network to quickly obtain a certain level of performance, and a low learning rate may be used in the late stage to increase accuracy.

In neural network learning, generally, training data can be a subset of real data (ie, data to be processed using the trained neural network), and therefore, errors for training data are reduced, but errors for real data are reduced. There may be incremental learning cycles. Overfitting is a phenomenon in which errors for actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that has learned a cat by showing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing the error of machine learning algorithms. Various optimization methods can be used to prevent such overfitting. In order to prevent overfitting, methods such as increasing training data, regularization, and omitting some nodes of a network in the process of learning may be applied.

In various embodiments, the computing device 100 may build an artificial intelligence model with an improved recognition rate by merging at least one image layer related to an object among a plurality of image layers extracted from a specific image and learning the artificial intelligence model. , By analyzing an image (or an image layer extracted from an image) through an artificial intelligence model with improved recognition rate, there is an advantage in that an object can be recognized more quickly and accurately.

In various embodiments, the computing device 100 may be connected to the user terminal 200 through the network 400 and may provide an object recognition result extracted as the user terminal 200 analyzes an image.

Here, the user terminal 200 includes navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA(Code Division Multiple Access)-2000, W-CDMA(W-Code Division Multiple Access), Wibro(Wireless Broadband Internet) Terminal, Smartphone, Smartpad, Tablet PC (Tablet PC), but is not limited thereto.

Also, here, the network 400 may refer to a connection structure capable of exchanging information between nodes such as a plurality of terminals and servers. For example, the network 400 includes a local area network (LAN), a wide area network (WAN), a world wide web (WWW), a wired and wireless data communication network, a telephone network, a wired and wireless television communication network, and the like. can do.

In addition, here, the wireless data communication networks are 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), 5GPP (5th Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi (Wi-Fi) Fi), Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), RF (Radio Frequency), Bluetooth network, A near-field communication (NFC) network, a satellite broadcasting network, an analog broadcasting network, a digital multimedia broadcasting (DMB) network, etc. may be included, but is not limited thereto.

In one embodiment, the external server 300 may be connected to the computing device 100 through the network 400, and various information and data necessary for the computing device 100 to perform an object recognition method using image histogram analysis. may be stored and managed, or various information and data generated as the computing device 100 performs an object recognition method using image histogram analysis may be received, stored, and managed. For example, the external server 300 may be a storage server provided separately outside the computing device 100, but is not limited thereto. Hereinafter, with reference to FIG. 2 , the hardware configuration of the computing device 100 that performs the object recognition method using image histogram analysis will be described.

2 is a hardware configuration diagram of an object recognition apparatus using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to another embodiment of the present invention.

Referring to FIG. 2 , in various embodiments, a computing device 100 includes one or more processors 110, a memory 120 for loading a computer program 151 executed by the processor 110, and a bus ( 130), a communication interface 140, and a storage 150 for storing the computer program 151. Here, in FIG. 2, only components related to the embodiment of the present invention are shown. Therefore, those skilled in the art to which the present invention pertains can know that other general-purpose components may be further included in addition to the components shown in FIG. 2 .

The processor 110 controls the overall operation of each component of the computing device 100 . The processor 110 includes a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphic Processing Unit (GPU), or any type of processor well known in the art of the present invention. It can be.

Also, the processor 110 may perform an operation for at least one application or program for executing a method according to embodiments of the present invention, and the computing device 100 may include one or more processors.

In various embodiments, the processor 110 may temporarily and/or permanently store signals (or data) processed in the processor 110 (RAM: Random Access Memory, not shown) and ROM (ROM: Read -Only Memory, not shown) may be further included. In addition, the processor 110 may be implemented in the form of a system on chip (SoC) including at least one of a graphics processing unit, RAM, and ROM.

Memory 120 stores various data, commands and/or information. Memory 120 may load computer program 151 from storage 150 to execute methods/operations according to various embodiments of the present invention. When the computer program 151 is loaded into the memory 120, the processor 110 may perform the method/operation by executing one or more instructions constituting the computer program 151. The memory 120 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The bus 130 provides a communication function between components of the computing device 100 . The bus 130 may be implemented in various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 140 supports wired and wireless Internet communication of the computing device 100 . Also, the communication interface 140 may support various communication methods other than internet communication. To this end, the communication interface 140 may include a communication module well known in the art. In some embodiments, communication interface 140 may be omitted.

The storage 150 may non-temporarily store the computer program 151 . When the recognition rate of the artificial intelligence model is improved through image histogram analysis through the computing device 100 and the object recognition process using the artificial intelligence model with the improved recognition rate is performed, the storage 150 stores the recognition rate of the artificial intelligence model through image histogram analysis. It is possible to store various types of information necessary to provide an object recognition method process using an artificial intelligence model with an improved recognition rate.

The storage 150 may be a non-volatile memory such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, or the like, a hard disk, a removable disk, or a device well known in the art. It may be configured to include any known type of computer-readable recording medium.

Computer program 151 may include one or more instructions that when loaded into memory 120 cause processor 110 to perform methods/operations in accordance with various embodiments of the invention. That is, the processor 110 may perform the method/operation according to various embodiments of the present disclosure by executing the one or more instructions.

In one embodiment, the computer program 151 may include performing histogram analysis on the image, extracting one or more image layers from the image based on a result of performing the histogram analysis on the image, and extracting one or more image layers. It may include one or more instructions for performing an object recognition method using an artificial intelligence model with improved recognition rate and improved recognition rate of an artificial intelligence model through image histogram analysis, which includes the step of recognizing an object included in an image by analyzing .

Steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented in a software module executed by hardware, or implemented by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any form of computer readable recording medium well known in the art to which the present invention pertains.

Components of the present invention may be implemented as a program (or application) to be executed in combination with a computer, which is hardware, and stored in a medium. Components of the present invention may be implemented as software programming or software elements, and similarly, embodiments may include various algorithms implemented as data structures, processes, routines, or combinations of other programming constructs, such as C, C++ , Java (Java), can be implemented in a programming or scripting language such as assembler (assembler). Functional aspects may be implemented in an algorithm running on one or more processors. Hereinafter, with reference to FIGS. 3 to 7 , an improvement in the recognition rate of the artificial intelligence model through image histogram analysis performed by the computing device 100 and an object recognition method using the artificial intelligence model having an improved recognition rate will be described.

3 is a flow chart of an object recognition method using an artificial intelligence model with an improved recognition rate and an improved recognition rate of an artificial intelligence model through image histogram analysis according to another embodiment of the present invention, and FIG. 4 is an image from an image in various embodiments. It is a diagram showing the process of extracting, selecting, and merging layouts by way of example.

Referring to FIGS. 3 and 4 , in step S110 , the computing device 100 may perform histogram analysis on an image (eg, a target image to be recognized as an object).

Here, the histogram analysis of the image is to analyze the image and count the number of pixels included in the image according to the pixel value to express the pixel value distribution for the image in the form of a graph. It may mean calculating the number of pixels for each pixel value and generating a graph of the number of pixels according to a plurality of pixel values using the calculated number, but is not limited thereto.

In various embodiments, the computing device 100 may generate a histogram graph (eg, FIG. 5 ) as a result of performing the histogram analysis as the histogram analysis is performed on the image.

Here, various techniques (eg, image histogram analysis using Open CV, etc.) are known for a method of performing histogram analysis on an image, and these known techniques can be selectively used. The specific method of performing the analysis is not limited.

In various embodiments, the computing device 100 may identify an object region included in an image, extract the object region from the image, and perform histogram analysis using only the extracted object region. For example, the computing device 100 may identify an object region included in the image by analyzing the image using an image analysis model, or may receive at least a partial region of the image selected by the user as the object region, and extract only the object region from the image. Histogram analysis can be performed.

In various embodiments, when a plurality of objects are included in one image, the computing device 100 may identify a plurality of object regions respectively corresponding to the plurality of objects and extract each of the plurality of object regions from the image. , histogram analysis may be individually performed on each of the extracted object regions.

In this case, the computing device 100 may perform histogram analysis by extracting only object regions corresponding to objects whose importance is greater than or equal to a preset importance among the plurality of objects, based on the importance of each of the plurality of objects. For example, when first to third objects are included in one image, and the importance of each of the first to third objects is first to third, the computing device 100 determines that the first to third objects are of first to third importance. Histogram analysis can be performed by extracting only object regions corresponding to each of the first and second objects of the first grade, and the object region corresponding to the third object of the third grade of importance is set as a background region to determine the histogram analysis target. can be excluded.

Here, the importance of each of the plurality of objects may be determined based on the size of the region corresponding to each of the plurality of objects, but is not limited thereto, and the importance of each of the plurality of objects may be set (or input) directly from the user. there is.

In various embodiments, when a plurality of objects are included in one image, the computing device 100 may extract one object region including all of the plurality of objects, and perform histogram analysis on the extracted one object region. can be done individually.

In this case, the computing device 100 may perform histogram analysis by extracting one object region including only objects whose importance is greater than or equal to a preset importance among the plurality of objects, based on the importance of each of the plurality of objects.

In various embodiments, the computing device 100 may identify a background region included in the image, remove the background region from the image, and perform histogram analysis on the image from which the background region has been removed. For example, the computing device 100 may identify a background area included in the image by analyzing the image using an image analysis model, or may select at least a partial area on the image as a background area from a user, and may use a mask layer corresponding to the background area. can be created and combined with the image, and histogram analysis can be performed on the image combined with the mask layer.

In various embodiments, the computing device 100 may histogram-analyze the grayscale image generated by converting the image to grayscale, and perform correction on the image based on the second histogram graph generated through this, , it is possible to perform histogram analysis on the corrected image.

Typically, as a result of performing histogram analysis on a grayscale image, when the grayscale values are generally distributed high, the brightness of the image is bright, and when the grayscale values are generally low, the brightness of the image is dark. .

That is, the computing device 100 considers that the size of the gray scale value is proportional to the brightness of the image, so that the minimum gray scale value of the second histogram graph is equal to or greater than the reference minimum value or the maximum gray scale value of the second histogram graph is the reference value. If it is less than the maximum value, that is, if it is determined that object recognition is adversely affected by the brightness of the image being too bright or too dark, brightness correction may be performed on the image.

In addition, typically, as a result of performing histogram analysis on grayscale images, if the grayscale values are generally evenly distributed, the contrast ratio of the corresponding image is high and clear, and if the grayscale values are generally narrowly distributed, the corresponding image The contrast ratio of the image may be low and may appear blurry.

That is, considering that the computing device 100 is proportional to the distribution range of gray scale values and the contrast ratio, the difference between the maximum gray scale value and the minimum value of the second histogram graph is equal to or less than the first reference difference value or the second reference difference value. If it is greater than the value, that is, if it is determined that object recognition is adversely affected by the contrast of the image being too high or too low, contrast correction may be performed on the image.

Here, performing brightness correction and contrast correction on the image is to more accurately recognize the object, and the computing device 100 extracts only the object area from the grayscale image or removes the background area and then performs histogram analysis. After extracting only the object area from the grayscale image or removing the background area, brightness correction and contrast correction may be performed on the original image based on a result of performing histogram analysis.

In step S120, the computing device 100 may extract one or more image layers from the image by using the histogram graph generated by performing the histogram analysis on the image in step S110. Hereinafter, it will be described in more detail with reference to FIG. 6 .

6 is a flowchart illustrating a method of extracting an image layer from an image, in various embodiments.

Referring to FIG. 6 , in step S210, the computing device 100 may select a pixel value for image layer extraction according to a preset criterion.

In various embodiments, the computing device 100 uses a graph of the number of pixels according to a plurality of pixel values, which is a histogram graph generated by performing histogram analysis on an image, from a pixel value having the largest number of pixels among a plurality of pixel values. It is possible to sequentially select N pixel values.

At this time, the computing device 100, as shown in FIG. 4, when the histogram graph includes a red (R) histogram graph, a green (G) histogram graph, and a blue (B) histogram graph, corresponding to each color. For the histogram graph, one or more pixel values may be individually selected sequentially from a pixel value having the largest number of pixels among a plurality of pixel values.

In various embodiments, the computing device 100 calculates the sum of pixel numbers for each of a plurality of pixel values by using a graph of the number of pixels according to a plurality of pixel values, which is a histogram graph generated by performing histogram analysis on an image. Among the plurality of pixel values, N pixel values may be sequentially selected from a pixel value having the largest sum of pixel numbers. For example, as shown in FIG. 4 , the computing device 100, when the histogram graph includes a red (R) histogram graph, a green (G) histogram graph, and a blue (B) histogram graph, each of a plurality of pixel values The sum of the number of pixels for each pixel value may be calculated by summing the numbers of red, green, and blue pixels for each pixel, and one or more pixel values may be sequentially selected from a pixel value having the largest sum of the calculated number of pixels.

In various embodiments, the computing device 100 may select a pixel value using a histogram graph generated by performing histogram analysis on a grayscale image generated by converting the image into a gray scale.

More specifically, first, the computing device 100 may generate a first histogram graph (eg, FIG. 5 ) by performing histogram analysis on the image.

Thereafter, the computing device 100 generates a grayscale image by converting the image into a grayscale, and performs a histogram analysis on the grayscale image to perform a second histogram graph (eg, a plurality of graphs corresponding to one color (black)). A graph of the number of pixels according to gray scale values) (eg, FIG. 7 ) may be generated. Here, the histogram analysis method for the grayscale image may be the same as the histogram analysis method for the above image, but is not limited thereto.

Thereafter, the computing device 100 may sequentially select N grayscale values starting from the grayscale value having the greatest number of pixels among the plurality of grayscale values using the second histogram graph, and using the first histogram graph to select N grayscale values. Among the pixel values, N pixel values corresponding to N gray scale values may be selected.

In various embodiments, the computing device 100 uses a graph of the number of pixels according to a plurality of pixel values, which is a histogram graph generated by performing histogram analysis on an image, to obtain a maximum value of the number of pixels for each of a plurality of pixel values. and a minimum value difference (for example, a difference value between the number of pixels of a first color having the largest number of pixels and the number of pixels of a second color having the smallest number of pixels for a specific pixel value), and calculating the number of pixels among a plurality of pixel values. At least one pixel value having a difference between a maximum value and a minimum value equal to or greater than a predetermined value may be selected.

In various embodiments, when a histogram analysis for each of the plurality of object regions is individually performed as a plurality of objects are included in one image, the computing device 100 individually performs histogram analysis for each of the plurality of object regions. A pixel value for each of a plurality of objects may be individually selected using a plurality of histogram graphs generated as the process is performed.

In various embodiments, the computing device 100 calculates one or more pixel values corresponding to an inflection point of the histogram graph by using a graph of the number of pixels according to a plurality of pixel values, which is a histogram graph generated by performing histogram analysis on an image. You can choose.

In various embodiments, the computing device 100 may identify an object region included in an image, extract a color of an object by analyzing the identified object region, and a plurality of graphs included in a histogram graph (eg, A pixel value may be selected using only a graph corresponding to the color of an extracted object among graphs of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors.

In various embodiments, the computing device 100 may identify an object region included in the image, and generate the object image and the background image by extracting the object region and the background region from the image, respectively, based on the identified object region. In addition, one or more first pixel values corresponding to the object and one or more second pixel values corresponding to the background may be selected using the histogram graph generated by performing histogram analysis on each of the object image and the background image.

At this time, if the color of the extracted object is two or more, the computing device 100 may select a pixel value using a graph corresponding to each of the two or more colors, but is not limited thereto, and selects the largest number of pixels among the two or more colors. A pixel value can be selected using a graph corresponding to only one color.

In step S220, the computing device 100 may extract an image layer from the image based on the pixel value selected in step S210.

In various embodiments, when N pixel values are selected from among a plurality of pixel values using the histogram graph, the computing device 100 may extract a plurality of image layers using the N pixel values.

For example, when the pixel values selected using the histogram graph are red (R) 230, green (G) 100, and blue (B) 70, all pixels having a color pixel value of 230 (eg, An image layer containing red pixel value 230, green pixel value 0 to 255, blue pixel value 0 to 255), all pixels with green pixel value 100 (e.g. red pixel value 0 to 255, green pixel value) An image layer containing 100, pixels with a blue pixel value of 0 to 255, and all pixels with a blue pixel value of 70 (e.g., pixels with a red pixel value of 0 to 255, a green pixel value of 0 to 255, and a blue pixel value of 70). s) can be extracted.

As another example, when the pixel values selected using the histogram graph are 230 red (R), 100 green (G), and 70 blue (B), the computing device 100 obtains a red pixel value of 230, a green pixel value of 100, and a blue pixel. One image layer containing pixels with a value of 70 can be extracted.

Again, referring to FIG. 3 , in step S130 , the computing device 100 may recognize an object included in the image using the image layer extracted through step S120 .

In various embodiments, the computing device 100 may recognize an object by integrating a plurality of image layers. For example, when a plurality of image layers are extracted from one image, the computing device 100 generates one integrated image layer by merging the plurality of image layers, and analyzes the integrated image layer to generate one or more characteristic values of an object. can be extracted, and an object can be recognized using one or more extracted feature values.

In various embodiments, the computing device 100 may recognize an object by individually using a plurality of image layers. For example, when a plurality of image layers are extracted from one image, the computing device 100 may extract one or more feature values by analyzing each of the plurality of image layers, and use the one or more feature values respectively extracted. Thus, the object can be individually estimated, and the object can be finally recognized using information about the individually estimated object.

In various embodiments, the computing device 100 may recognize an object using only feature values redundantly extracted from a plurality of image layers. For example, when a plurality of image layers are extracted from one image, the computing device 100 may extract a plurality of feature values of an object by analyzing each of the plurality of image layers, and may extract two of the plurality of extracted feature values. An object may be recognized using only at least one feature value extracted over and over again.

In various embodiments, the computing device 100 may recognize an object using only at least some image layers among a plurality of image layers.

More specifically, first, when a plurality of image layers are extracted from one image, the computing device 100 may extract one or more feature values by analyzing each of the plurality of image layers, and may extract one or more feature values respectively. A similarity between a plurality of image layers may be calculated by comparing feature values.

Here, the method of calculating the similarity between a plurality of image layers by comparing one or more extracted feature values is Jaccard Similarity, Cosine Similarity, and Euclidean Similarity for two or more feature values. A method of calculating at least one similarity among similarity and defining the calculated similarity as a similarity between image layers from which two or more feature values are respectively extracted may be applied, but is not limited thereto.

Thereafter, the computing device 100 may select two or more image layers having similarities greater than or equal to a first reference value based on the similarities between the plurality of image layers, merge the selected two or more image layers, and merge the two or more images. An object can be recognized based on the feature value extracted by analyzing the layer. However, the computing device 100 is not limited thereto, and the computing device 100 may recognize an object by merging image layers other than at least one image layer having a similarity less than a second reference value based on a similarity between a plurality of image layers.

In various embodiments, the computing device 100 uses a histogram graph generated by performing histogram analysis on each of the object image and the background image, and uses one or more first pixel values corresponding to the object and one or more first pixel values corresponding to the background. When 2 pixel values are selected, a plurality of first feature values may be extracted from a plurality of first image layers extracted based on one or more first pixel values, and a plurality of first feature values extracted based on one or more second pixel values. A plurality of second feature values may be extracted from the second image layer, and the object may be recognized using only at least one first feature value that does not overlap with the plurality of second feature values among the plurality of first feature values. .

As described above, the object recognition method using the artificial intelligence model with improved recognition rate and improved recognition rate of the artificial intelligence model through image histogram analysis according to various embodiments of the present invention extracts a plurality of image layers from a high-resolution image, and extracts the extracted It is possible to improve the object recognition rate through the artificial intelligence model by selectively merging at least one image layer related to the object among the plurality of image layers and recognizing the object by analyzing the merged image layer based on the artificial intelligence model. There is an advantage.

The recognition rate improvement of the artificial intelligence model through the above-described image histogram analysis and the object recognition method using the artificial intelligence model with improved recognition rate have been described with reference to the flow chart shown in the drawings. For a brief description, the recognition rate improvement of the artificial intelligence model through image histogram analysis and the object recognition method using the artificial intelligence model with improved recognition rate have been illustrated and described as a series of blocks, but the present invention is not limited to the order of the blocks, Some blocks may be performed in a different order or concurrently than those shown and performed herein. In addition, new blocks not described in the present specification and drawings may be added, or some blocks may be deleted or changed.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

100: object recognition device (computing device)
200: user terminal
300: external server
400: Network

Claims (14)

In a method performed by a computing device,
performing histogram analysis on the image;
extracting one or more image layers from the image based on a result of performing histogram analysis on the image; and
Recognizing an object included in the image by analyzing the extracted one or more image layers,
The step of extracting the one or more image layers,
Using a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors, the plurality of pixel values calculating the sum of the number of pixels of the plurality of colors for each color, and sequentially selecting N pixel values from among the plurality of pixel values, starting with a pixel value having the largest sum of the calculated number of pixels; and
Extracting a plurality of image layers from the image based on the selected N pixel values,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
The step of extracting the one or more image layers,
A pixel value having the largest number of pixels among the plurality of pixel values using a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values. sequentially selecting N pixel values from; and
Extracting a plurality of image layers from the image based on the selected N pixel values,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. delete According to claim 1,
Performing histogram analysis on the image,
generating a first histogram graph by performing histogram analysis on the image, wherein the first histogram graph includes a graph of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors; and
A grayscale image is generated by converting the image to grayscale, and a histogram analysis is performed on the generated grayscale image to obtain a second histogram graph - the second histogram graph is based on a plurality of grayscale values corresponding to one color. including a graph of the number of pixels according to;
The step of extracting the one or more image layers,
sequentially selecting N grayscale values from the grayscale values having the largest number of pixels among the plurality of grayscale values using the generated second histogram graph; and
N pixel values corresponding to the selected N grayscale values are selected from among the plurality of pixel values using the generated first histogram graph, and a plurality of image layers are selected from the image based on the selected N pixel values. Including the step of extracting,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
The step of extracting the one or more image layers,
Using a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors, the plurality of pixel values calculating a difference between a maximum value and a minimum value of the number of pixels for each pixel value, and selecting at least one pixel value having the calculated difference greater than or equal to a predetermined value among the plurality of pixel values; and
Extracting a plurality of image layers from the image based on the selected at least one pixel value,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Recognizing an object included in the image,
When a plurality of image layers are extracted from the image, generating a single integrated image layer by merging the extracted plurality of image layers;
extracting one or more feature values of the object by analyzing the generated integrated image layer; and
Recognizing the object using the extracted one or more feature values,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Recognizing an object included in the image,
When a plurality of image layers are extracted from the image, one or more feature values of the object are extracted by analyzing each of the extracted image layers, and the object is obtained using the one or more feature values respectively extracted. estimating individually; and
Recognizing the object using information about the individually estimated object,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Recognizing an object included in the image,
When a plurality of image layers are extracted from the image, a plurality of feature values of the object are extracted by analyzing each of the extracted image layers, and at least one of two or more of the extracted feature values is repeatedly extracted. Recognizing the object using only the feature value of
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Recognizing an object included in the image,
When a plurality of image layers are extracted from the image, each of the extracted image layers is analyzed to extract one or more feature values, and the extracted one or more feature values are compared to obtain a plurality of extracted images. Calculating a degree of similarity between layers; and
Merging two or more image layers among the plurality of extracted image layers based on the calculated similarity, and recognizing the object by analyzing the merged two or more image layers,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Performing histogram analysis on the image,
A second histogram graph generated by converting the image to gray scale to generate a gray scale image and performing histogram analysis on the generated gray scale image - the second histogram graph is the number of pixels according to a plurality of gray scale values performing a correction on the image based on - including a graph; and
Generating a first histogram graph by performing histogram analysis on the corrected image,
The step of correcting the image is,
performing brightness correction on the image when the minimum gray scale value of the second histogram graph is greater than or equal to the reference minimum value or the maximum gray scale value of the second histogram graph is less than or equal to the reference maximum value; and
Performing contrast correction on the image when the difference between the maximum and minimum grayscale values of the second histogram graph is less than or equal to a first reference difference value or greater than or equal to a second reference difference value,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Performing histogram analysis on the image,
identifying an object region included in the image; and
Extracting the identified object region from the image and performing histogram analysis on the extracted object region.
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. According to claim 1,
Performing histogram analysis on the image,
identifying a background region included in the image;
creating a mask layer corresponding to the identified background region and combining the created mask layer with the image; and
Including performing histogram analysis on the image combined with the generated mask layer,
Object recognition method using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. processor;
network interface;
Memory; and
A computer program loaded into the memory and executed by the processor,
The computer program,
instructions for performing histogram analysis on the image;
instructions for extracting one or more image layers from the image based on a result of performing histogram analysis on the image; and
An instruction for recognizing an object included in the image by analyzing the extracted one or more image layers,
The instructions for extracting the one or more image layers,
Using a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors, the plurality of pixel values an instruction for calculating a sum of the number of pixels of the plurality of colors for each color, and sequentially selecting N pixel values from among the plurality of pixel values, starting from a pixel value in which the calculated sum of the number of pixels is the largest; and
Including instructions for extracting a plurality of image layers from the image based on the selected N pixel values,
Object recognition device using artificial intelligence model with improved recognition rate and improved recognition rate of artificial intelligence model through image histogram analysis. Combined with a computing device,
performing histogram analysis on the image;
extracting one or more image layers from the image based on a result of performing histogram analysis on the image; and
Recognizing an object included in the image by analyzing the extracted one or more image layers,
The step of extracting the one or more image layers,
Using a histogram graph generated based on a result of performing histogram analysis on the image, wherein the histogram graph includes a graph of the number of pixels according to a plurality of pixel values corresponding to a plurality of colors, the plurality of pixel values calculating the sum of the number of pixels of the plurality of colors for each color, and sequentially selecting N pixel values from among the plurality of pixel values, starting with a pixel value having the largest sum of the calculated number of pixels; and
In order to improve the recognition rate of the artificial intelligence model through image histogram analysis, including the step of extracting a plurality of image layers from the image based on the selected N pixel values, and to execute an object recognition method using the artificial intelligence model with improved recognition rate. A computer program stored on a recording medium readable by a computing device.