KR20220021493A - Method and apparatus for improving resolution based on image segmentation - Google Patents

Abstract

An object of the present invention is to propose a method and device for reducing time required to improve resolution of an image while improving resolution to a target image quality. According to an embodiment of the present invention, a method and for improving resolution based on image segmentation comprises: a step of obtaining a std value indicating a degree of data distribution for each arbitrary image; a step of classifying the images into image groups according to the numerical value of the std value; and a step of selecting an image group that meets a criteria among a plurality of image groups as image data for learning and executing learning of an AI image processing model.

Description

The method and apparatus for improving resolution based on image segmentation TECHNICAL FIELD

The present invention relates to a method and apparatus for improving the resolution of an image based on an image segmentation method.

A super-resolution image restoration method using deep learning and a convolution neural network (CNN) learns a network through supervised learning. In this case, general clear images are used as target images or label concepts for supervised learning, and images whose resolution is lowered through downsampling are used as input low-resolution images.

The learning process is as follows. At each iteration, the input low-resolution image or image is converted to a super-resolution image or image by passing through a network of convolutional and upsampling layers, the super-resolution image is compared to the target image to measure the difference, and the difference is reduced to reduce the difference. back-propagate to change the weight of the convolution layer. As the learning progresses, the difference value between the target images of the super-resolution image for the training data gradually decreases, and the more the data is diversified, the more the learning effect is generalized.

In such a super-resolution image restoration method using deep learning and CNN, in order to improve the resolution restoration performance, there is a method of increasing the number of convolutional layers to have a deeper dimensional neural network structure. However, in the case of this method, there is a problem that the computational load of data required in the learning step and the test step is large and the time required for the computation is long.

Accordingly, there has been a demand for a new method for reducing the data load required for calculation and for performing restoration to a target resolution while reducing the time required for resolution improvement.

It is an object of the present invention to propose a method and apparatus for performing high-quality resolution restoration in a simple manner without performing excessive data operations such as using a plurality of GPUs to improve image resolution. .

Another object of the present invention is to propose a method and apparatus for reducing the time required to improve the resolution of an image while improving the resolution to a target image quality.

In order to achieve the above object, the image segmentation-based resolution improvement method according to an embodiment of the present invention includes the steps of obtaining a std value indicating the degree of data distribution for each arbitrary image, and the image according to the numerical value of the std value. It may include the steps of classifying the images by image group and performing learning of the AI image processing model by selecting an image group that meets the criteria among a plurality of image groups as image data for training.

And when the learning of the AI image processing model is completed, the image division-based resolution improvement method of the present invention further includes an image processing step of inputting a target image requiring resolution improvement into the AI image processing model to perform an AI operation can do.

And the image processing step is a step of dividing the target image into an arbitrary patch size in order to input the target image for which resolution improvement is required to the AI image processing model, and inputting the divided target image into the AI image processing model to AI It may be configured including the step of performing the calculation and the step of generating final result data by concatenating the divided result data calculated after the AI calculation.

In the step of dividing the target image into arbitrary patch sizes, the target image may be divided into sizes corresponding to the patch sizes of the image data for training input in the training step of the AI image processing model. In addition, when dividing the target image, it may be set to divide the patch size according to the presence or absence of an object included in the target image and the size of the object.

In the step of performing the learning of the AI image processing model, the range of the std value selected as the image data for learning may be set differently in response to the property of the image to be improved in resolution.

In the present invention, in performing image resolution improvement, excessive data operation such as using a plurality of GPUs is required in the prior art, but the resolution is performed by dividing the image according to the standard, restoring the resolution of each divided image, and then joining the image. Since restoration is performed, the burden of excessive data operation may be reduced.

In addition, in the present invention, in order to improve image resolution with the same quality, the image is divided according to a predetermined criterion, the resolution is improved, and the divided images with improved resolution are joined later, so the convolution layer of the AI model is There is an advantage in that the number does not need to be increased excessively, so the time required for training the AI model and calculating the resolution improvement can be shortened.

1 is a diagram illustrating a configuration of a resolution improving apparatus according to an embodiment of the present invention.
2 is a diagram illustrating a configuration of a learning support unit according to an embodiment of the present invention.
3 is a diagram illustrating the configuration of an AI image processing unit according to an embodiment of the present invention.
4 is a diagram to help understanding of an image group according to an embodiment of the present invention.
5 is a diagram illustrating an operation of performing resolution improvement through image segmentation according to an embodiment of the present invention.
6 is a flowchart illustrating a sequence of a learning process of an AI model according to an embodiment of the present invention.
7 is a flowchart illustrating an image resolution improvement process through an AI model according to an embodiment of the present invention.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail 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, and only the present embodiments allow the disclosure of the present invention to be complete, and those of ordinary skill in the art to which the present invention pertains. It is provided to fully understand the scope of the present invention to those skilled in the art, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing the embodiments and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, “comprises” and/or “comprising” does not exclude the presence or addition of one or more other components in addition to the stated components. Like reference numerals refer to like elements throughout, 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 elements, these elements 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 component mentioned below may be the second component within the spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used herein will have the meaning commonly understood by those of ordinary skill in the art to which this invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless specifically defined explicitly.

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

Spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe the correlation between a component and other components. Spatially relative terms should be understood as terms including different directions of components during use or operation in addition to the directions shown in the drawings. For example, when a component shown in the drawing is turned over, a component described as “beneath” or “beneath” of another component may be placed “above” of the other component. can Accordingly, the exemplary term “below” may include both directions 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 refers to all types of hardware devices including at least one processor, and may be understood as encompassing software configurations operating in the corresponding hardware device according to embodiments. For example, a computer may be understood to include, but is not limited to, smart phones, tablet PCs, desktops, notebooks, and user clients and applications running on each device.

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

1 is a diagram illustrating a configuration of a resolution improving apparatus according to an embodiment of the present invention.

As shown in FIG. 1 , the resolution improving apparatus 100 according to an embodiment of the present invention may include a learning support unit 110 and an AI image processing unit 120 .

The learning support unit 110 may support learning of an AI calculation model for improving the resolution of a target image. The AI computation model may be a deep learning model including at least three or more convolutional layers, and a high-resolution target image defined as GT (ground truth) and a corresponding low-resolution version image (LR). It may include an SR (Super Resolution) model, which is a method that is learned to restore a low-resolution image (LR) to GT by inputting in pairs.

In addition, the learning support unit 110 selects an image for learning and may additionally perform a series of processes to learn the AI computation model using the selected image, which will be described with reference to FIG. 2 .

2 is a diagram illustrating a configuration of a learning support unit according to an embodiment of the present invention.

As shown in FIG. 2 , the learning support unit 110 may include a learning image classification unit 111 and a learning image selection unit 112 .

First, the training image classifying unit 111 may classify the training images to be trained on the AI operation model by group. In this case, the criterion for classifying the image may be based on the std value. For a description thereof, reference will be made to FIG. 4 .

4 is a diagram to help understand an image group according to an embodiment of the present invention.

As shown in FIG. 4 , arbitrary images have a std value (meaning standard deviation), which is a value indicating the degree of distribution of data, and the training image classifier 111 according to an embodiment of the present invention The std value may be checked, and an operation of classifying an image group may be performed based on the std value.

The std value contains an approximate structure of data, and it can be understood that the complexity of the image increases as the size of the std value increases. At this time, when an arbitrary patch size is determined, the training image classification unit 111 may divide a specific image into a predetermined patch size. there is.

The training image classifying unit 111 may classify the images for which the std value is calculated into an image group corresponding to the size of the image. Although an example is shown in FIG. 4 , the std value may be calculated in a more diverse range than shown in FIG. 4 . However, in the present specification, for better understanding, only images having std values in the range of 0 to 50 will be referred to.

As shown in FIG. 4 , the training image classifying unit 111 may classify an image group according to a range of std values. For example, images having std values of 0 to 50 are selected from 0 to 10, 10 to It can be classified into a total of five image groups, such as 20, 20-30, 30-40, 40-50.

When the images are classified into groups according to the std value, the training image selection unit 112 may select image data for training to be input to the AI computation model. In this case, the learning image selection unit 112 may select images for learning to be input to the AI computation model, and may perform a learning operation by inputting the images to the AI computation model.

According to various embodiments, the learning image selection unit 112 may learn several AI computation models for each pre-classified image group when performing a learning operation by inputting image data for learning into the AI computation model. For example, the training image selector 112 may input the first group (eg, an image group having a std value of 0 to 10) to the AI computation model to learn it, and generate the first model as a result of the learning. Similarly, the training image selection unit 112 may generate various types of result models, such as the second model and the third model, as a result of inputting each image group to the AI computation model and learning it.

In addition, when selecting image data for training, the training image selection unit 112 may determine different training image selection criteria according to target image properties of the corresponding AI model. Specifically, for example, when classifying according to the properties of the target image, the learning image selector 112 may classify the image into an examination image and a daily image, and apply different learning criteria to each.

The image for examination may mean images taken for examination such as X-ray and CT, and the images for such examination have a relatively narrow distribution range of std values. For everyday images taken through the std value, the distribution range of the std value may be calculated to be relatively wide.

In addition, the training image selection unit 112 sets the range of the std value selected as image data for training when learning the AI model targeting the inspection image, for example, to 10 to 50, and learns the AI model for everyday images. For example, the std value range selected as image data for learning may be set to 0 to 100. In this case, the range of the std value selected as the image data for training is merely an example, and the range of the actual std value may be more limited or may have a wider range.

Alternatively, the learning image selection unit 112 is an image to be selected for learning according to the properties of the image to be improved in image quality (eg, bio X-ray, bio ultrasound image, CT image, industrial X-ray, industrial general camera image, etc.) It is possible to change the range of data structure values (std values) of , but increase the selection of image groups based on the data structure values (std) of the images to be selected for learning as the average complexity of the target image for image quality improvement increases. For example, if the AI model to be created is a model for improving the resolution of X-ray images for bio, the image data to be selected for training is 3 image groups (eg, std 20~30, 30~40, 40~50). ), if the AI model to be created is a model for improving the resolution of general industrial camera images, there are 6 image groups (eg, std 20~30, 30~40, 40~50, 50~60, 60~). 70, 70-80) may be selected as image data for learning. As described above, the range of the data structure value std selected as the learning image may be increased or decreased according to the properties of the target image for image quality improvement.

According to various embodiments, when the balance of the training image input in the learning step is broken, the image improvement performance of the AI model may be deteriorated. In order to prevent this problem, the training image selection unit 112 according to an embodiment of the present invention may select only images having an std value within a reference range as image data for training. To this end, the super-resolution model learning unit 121a may exclude data having an extreme std value from the learning step. For example, the std value of the provided images is checked in the range of 0 to 100, When it is assumed that 10 groups (eg, 1 group: std 0-10, 2 group: 10-20…10 group: 90-100) are designated according to the size of the std value, the super-resolution model learning unit 121a is an extreme value Images corresponding to an image group (eg, 1 group, 10 groups) having a predetermined std value considered to be may be removed in the learning step.

As described above, the training image selection unit 112 may set the std reference value of the image to be selected as the training image data according to the property of the image to be improved in resolution (whether it is an image for inspection or not).

In addition, the AI model provided by the resolution improving apparatus 100 according to an embodiment of the present invention may be configured in a form including a plurality of lower AI models corresponding to each std value. For convenience of explanation, the upper AI model is referred to as an integrated model and the lower AI model is referred to as a divided model. .

In addition, the resolution improving apparatus 100 may train each segmented model as an image model having a specific std value. In this case, the number of image groups input as learning data is set differently according to the properties of the target image, and, accordingly, The number of split models included in the integrated model can be adjusted.

For example, the resolution improving apparatus 100 sets the number of image groups input for learning to be less than or equal to a reference value (eg, 4) to generate an AI model targeting an image for examination such as X-ray, The number of divided models may also be set to four corresponding to the number of image groups. On the other hand, the resolution improving apparatus 100 may increase and set the number of image groups to be input for learning in order to generate an AI model for everyday images, for example, set to 10, and the number of divided models corresponding to the number of image groups It can also be set to 10.

In other words, the resolution improvement apparatus 100 increases the std distribution range of the image set for learning, the number of image groups, and the number of segmentation models to be generated correspondingly as the complexity of the image to be improved increases. can

Hereinafter, with reference to FIG. 3 , an operation of improving the resolution of an image through an operation using an AI model will be described in detail.

3 is a diagram illustrating the configuration of an AI image processing unit according to an embodiment of the present invention.

As shown in FIG. 3 , the AI image processing unit 120 according to an embodiment of the present invention may include a patch size determination unit 121, an image division unit 122, an AI operation processing unit 123, and an image joining unit 124.

Prior to a description of each component of the AI image processing unit 120, an overall operation of the AI image processing unit 120 for upgrading the resolution of a target image will be first described with reference to FIG. 5 for better understanding.

5 is a diagram illustrating an operation of performing resolution improvement through image segmentation according to an embodiment of the present invention.

As shown in FIG. 5 , the AI image processing unit 120 divides the original image that is the target of resolution improvement, classifies the divided images according to each std value, and distributes them into the corresponding AI model (segment model). can do. As shown in FIG. 5 , assuming that the AI model according to an embodiment is configured to include 11 divided models, the divided images are classified into 11 types based on the std value, and corresponding to the std values of each image. AI calculations for resolution improvement by being input to the segmentation model can be processed individually. Since the result data calculated from each segmentation model will still be in a segmented form, the AI image processing unit 120 can calculate the final result data by concatenating the segmentation result data calculated through each segmentation model.

3 again, the patch size determiner 121 divides the image before inputting the target image to the AI model. At this time, the target image is divided into a preset patch size. At this time, the image is The size to be split is determined.

According to various embodiments, the patch size determiner 121 may determine the patch size as, for example, 24*24, 48*48, or the like.

Also, according to various embodiments, the patch size determiner 121 may set the patch size differently according to the properties of the target image, the presence or absence of a main object in the target image, and the properties of the main object in the target image.

Due to the nature of the present invention, since the resolution is restored by dividing the image and then splicing it, the resolution may inevitably be lowered in the splicing area. Therefore, in order to minimize this problem, it is important to divide the image so that the main object is not divided in the area where the main object is displayed. Accordingly, in consideration of this problem, the patch size determiner 121 may adjust the patch size in consideration of the presence or absence of the main object, the location of the main object, or the size of the main object.

More specifically, the patch size determiner 121 may first determine the presence or absence of a main object, draw a virtual dividing line in consideration of the location and size of the main object, and determine whether the main object is to be divided. . Thereafter, the patch size determiner 121 may determine a size such that the main object is not divided or is divided to a minimum among settable patch sizes.

Also, the image dividing unit 122 may divide the target image according to the patch size determined by the patch size determining unit 121 . In addition, the image segmentation unit 122 may obtain an std value of the segmented image and classify the image group according to the confirmed std value.

In addition, the AI operation processing unit 123 may perform an AI operation by inputting the classified image group into a segmentation model corresponding to each std value to calculate result data. At this time, since the result data calculated from the segmentation model will have improved resolution or a segmented image, a process of rejoining the segmented image like the original image is required.

Finally, the image splicing unit 124 may perform a splicing operation for calculating final result data based on the primary result value (an image with improved resolution but still segmented state) calculated by the AI operation processing unit 123 .

According to an embodiment of the present invention, at the time of image segmentation, a position value required for determining the position of the segmented image in the original image may be provided so that the segmented images can be rejoined later. Accordingly, each of the divided images may be in a state in which a position value is assigned, and the image junction unit 124 may recombine the divided image again based on the position value to restore the same as the original image.

6 is a flowchart illustrating a sequence of a learning process of an AI model according to an embodiment of the present invention.

As shown in FIG. 6 , the resolution improving apparatus 100 according to an embodiment of the present invention may perform operation 605 of checking the std value of one image. At this time, the std value means the standard deviation, and contains the approximate structure of the data as a value indicating the degree of data distribution.

In addition, when the std value is calculated in operation 605, the resolution improving apparatus 100 may perform operation 610 of classifying images into groups according to the calculated std value. For example, the resolution improving apparatus 100 may classify the std value of the image into a preset unit (eg, 10). Accordingly, images having a std value of 0 or more and less than 10 may be classified as a first group, and images having a std value of 10 or more and less than 20 may be classified as a second group.

Thereafter, the resolution improving apparatus 100 may perform operation 615 of selecting an image group meeting a criterion as an image for training. If, for example, five image groups are generated, the resolution improving apparatus 100 may train five image groups in each segmentation model.

Thereafter, the resolution improving apparatus 100 may perform operation 620 of training the AI model with the selected training image. For example, when the resolution improving apparatus 100 learns by inputting five image groups into each segmented model, it may obtain five learned segmented models.

7 is a flowchart illustrating an image resolution improvement process through an AI model according to an embodiment of the present invention.

As shown in FIG. 7 , when the resolution improving apparatus 100 according to an embodiment of the present invention wants to improve the resolution by inputting the target image to the AI model, first, operation 705 of dividing the target image into arbitrary patch sizes is performed. can be done

In addition, the resolution improving apparatus 100 may perform operation 710 to check the std value of the divided image.

Thereafter, the resolution improving apparatus 100 may perform operation 715 of calculating the divided image by inputting the divided image into the AI model of the type corresponding to the std value of the divided image. When operation 715 is performed, a plurality of divided images may be distributed and input into a plurality of types of AI models according to std values. In addition, since the primary result calculated by passing through the AI model is in the form of a split image, the split image needs to be rejoined based on information at the time of splitting.

Accordingly, the resolution improving apparatus 100 may perform operation 720 of generating a final result image by joining the calculated division result images.

In summary, the resolution improving apparatus 100 divides the target image whose resolution is to be improved into a preset size, classifies the divided image according to the std value, and displays the divided image in an AI model corresponding to the std value of the divided image. to increase the resolution, and by combining the divided images with improved resolution, it is possible to calculate the resulting image in the undivided state with improved resolution.

In other words, the image segmentation-based resolution improvement method according to an embodiment of the present invention includes the steps of obtaining a std value indicating the degree of data distribution for each arbitrary image, and classifying the images into image groups according to the numerical value of the std value. and selecting an image group that meets the criteria among a plurality of image groups as image data for training, and performing learning of the AI image processing model.

And when the learning of the AI image processing model is completed, the image division-based resolution improvement method of the present invention further includes an image processing step of inputting a target image requiring resolution improvement into the AI image processing model to perform an AI operation can do.

And the image processing step is a step of dividing the target image into an arbitrary patch size in order to input the target image for which resolution improvement is required to the AI image processing model, and inputting the divided target image into the AI image processing model to AI It may be configured including the step of performing the calculation and the step of generating final result data by concatenating the divided result data calculated after the AI calculation.

In the step of dividing the target image into arbitrary patch sizes, the target image may be divided into sizes corresponding to the patch sizes of the image data for training input in the training step of the AI image processing model. In addition, when dividing the target image, it may be set to divide the patch size according to the presence or absence of an object included in the target image and the size of the object.

In the step of performing the learning of the AI image processing model, the range of the std value selected as the image data for learning may be set differently in response to the property of the image to be improved in resolution. For example, the property of the target image may be classified into an examination image having a relatively narrow std value distribution, such as X-ray and CT, and an everyday image having a relatively wide std value distribution. In the case of inspection images, as the distribution range of std values is limited compared to everyday images, the range of std values selected as image data for training of the AI model dedicated to inspection images is similar to that of the AI model for everyday images in the training stage. It may be limited compared to the std reference range of image data for training. For example, assuming the std value range is 0 to 100, the std value range of the training image dataset of the AI model for inspection is 10-40, the std of the training image dataset of the AI model for everyday images is The value range can be set from 0 to 100.

In addition, the resolution improving apparatus 100 according to an embodiment of the present invention may be configured to include a learning support unit and an AI image processing unit, wherein the learning support unit obtains a std value indicating a degree of data distribution for each arbitrary image, and the image A training image classification unit that generates an image group classified according to the numerical value of the std value, and a training image selection unit that selects images in the image group meeting the criteria as image data for training and performs learning of the AI image processing model, and , the AI image processing unit may be configured to perform an AI operation by inputting a target image requiring resolution improvement into the AI image processing model as the learning of the AI image processing model is completed by the learning support unit.

In addition, according to various embodiments of the present disclosure, although not shown in the drawings, the resolution improving apparatus 100 includes a communication unit for receiving image data from the outside, and a storage unit for storing image data received from the outside and a resolution-improved result value. may include more.

The communication unit may use a network for data transmission/reception between the user device and the server, and the type of the network is not particularly limited. The network may be, for example, an IP (Internet Protocol) network that provides a large-capacity data transmission/reception service through Internet Protocol (IP) or an All IP network that integrates different IP networks. In addition, the network includes a wired network, a Wibro (Wireless Broadband) network, a mobile communication network including WCDMA, a High Speed Downlink Packet Access (HSDPA) network, and a Long Term Evolution (LTE) network including a mobile communication network, LTE advanced (LTE-A) ), a mobile communication network including 5G (Five Generation), a satellite communication network, and a Wi-Fi network, or a combination of at least one or more of them.

The storage unit may include, for example, an internal memory or an external memory. The internal memory includes, for example, a volatile memory (eg, dynamic RAM (DRAM), static RAM (SRAM), or synchronous dynamic RAM (SDRAM), etc.), non-volatile memory (eg, OTPROM (one time programmable ROM), programmable ROM (PROM), erasable and programmable ROM (EPROM), electrically erasable and programmable ROM (EEPROM), mask ROM, flash ROM, flash memory (such as NAND flash or NOR flash), hard drive, or a solid state drive (SSD).

External memory is a flash drive, for example, CF (compact flash), SD (secure digital), Micro-SD (micro secure digital), Mini-SD (mini secure digital), XD (extreme digital), It may further include a multi-media card (MMC) or a memory stick. The external memory may be functionally and/or physically connected to the electronic device through various interfaces.

Although the present invention has been described in detail with reference to the above-described examples, those skilled in the art can make modifications, changes, and modifications to the examples without departing from the scope of the present invention. In short, in order to achieve the intended effect of the present invention, it is not necessary to separately include all the functional blocks shown in the drawings or follow all the orders shown in the drawings. Note that it may fall within the scope.

100: resolution improvement device
110: Learning Support Department
111: learning image classification unit
112: learning image selection unit
120: AI image processing unit
121: patch size determining unit
122: image division part
123: AI operation processing unit
124: image joint

Claims (7)

obtaining a std value indicating a degree of data distribution for each arbitrary image;
classifying the images into image groups according to the numerical value of the std value; and
performing learning of the AI image processing model by selecting an image group that meets the criteria among a plurality of image groups as image data for training; Image segmentation-based resolution improvement method comprising a. 2. The method of claim 1
When the learning of the AI image processing model is completed, an image processing step of performing AI operation by inputting a target image for which resolution improvement is required to the AI image processing model; . 3. The method of claim 2,
The image processing step is
dividing the target image into arbitrary patch sizes to input the target image for which resolution improvement is required to the AI image processing model;
performing AI calculation by inputting the segmented target image to the AI image processing model; and
Image segmentation-based resolution improvement method comprising the; generating final result data by concatenating segmentation result data calculated after AI operation. 4. The method of claim 3,
The step of dividing the target image into an arbitrary patch size is
Image division-based resolution improvement method, characterized in that the target image is divided into a size corresponding to the patch size of the image data for training input in the learning step of the AI image processing model. 4. The method of claim 3
The step of dividing the target image into an arbitrary patch size is
An image segmentation-based resolution improvement method, characterized in that the segment is divided into patch sizes adjusted according to the presence or absence of an object included in the target image and the size of the object. The method of claim 1,
The step of performing the learning of the AI image processing model is
An image segmentation-based resolution improvement method, characterized in that a range of std values selected as image data for training is set differently in response to a property of an image to be improved in resolution. In the resolution improving device comprising a learning support unit and an AI image processing unit,
The learning support department
a learning image classification unit that obtains a std value indicating a degree of data distribution for each arbitrary image, and generates an image group in which the image is classified according to the numerical value of the std value; and
Including; a learning image selection unit that performs learning of the AI image processing model by selecting images in the image group that meet the criteria as image data for training;
The AI image processing unit
As the learning of the AI image processing model is completed by the learning support unit, the image division-based resolution improvement apparatus, characterized in that, by inputting the target image for which resolution improvement is required to the AI image processing model to perform AI operation.

Images