KR20230103372A - Gan-based artificial intelligence hologram image restoration method and system - Google Patents

Abstract

A GAN-based artificial intelligence hologram image restoration method and system are disclosed. A GAN-based artificial intelligence holographic image restoration system according to an embodiment of the present invention includes an input unit for receiving a single holographic image in which optical field information is missing; and an output unit generating a restored image by applying a GAN algorithm to the holographic image input to the input unit.

Description

GAN-based artificial intelligence hologram image restoration method and system thereof

The present invention relates to an image restoration method and system using a GAN network, and more particularly, to a hologram image acquired from a scanner is converted into a vector in a latent space and stored, and then the converted image is converted into a GAN generator. A method and system for returning an image in which phase information is restored by inputting it to a network are provided.

An artificial intelligence system is a computer system that implements human-level intelligence, and unlike existing rule-based smart systems, machines learn, judge, and become smarter on their own. The more AI systems are used, the higher the recognition rate and the more accurate understanding of user preferences, so existing rule-based smart systems are gradually being replaced by deep learning systems.

Artificial intelligence technology consists of machine learning (deep learning) and element technologies using machine learning. Machine learning is an algorithm technology that classifies/learns the characteristics of input data by itself, and element technology is a technology that simulates functions such as recognition and judgment of the human brain using machine learning algorithms such as deep learning. It consists of technical fields such as understanding, inference/prediction, knowledge expression, and motion control.

The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding is a technology for recognizing and applying/processing human language/text, and includes natural language processing, machine translation, dialogue systems, question and answering, voice recognition/synthesis, and the like. Visual understanding is a technology for recognizing and processing objects like human vision, and includes object recognition, object tracking, image search, person recognition, scene understanding, space understanding, image improvement, and the like. Inference/prediction is a technique of reasoning and predicting logically by judging information, and includes knowledge/probability-based reasoning, optimization prediction, preference-based planning, and recommendation. Knowledge expression is a technology that automatically processes human experience information into knowledge data, and includes knowledge construction (data creation/classification) and knowledge management (data use). Motion control is a technology for controlling the autonomous driving of a vehicle and the movement of a robot, and includes motion control (navigation, collision, driving), manipulation control (action control), and the like.

On the other hand, with the development of multimedia technology and computer technology, image generation technology using a deep learning technique is being developed, but there are problems in that the learning speed is remarkably slow, the computational complexity is high, and the resolution of the generated image is low.

In this regard, Korea Patent Publication No. 10-2018-0004989, a prior art, combines a neural network circuit and a parallel processor to analyze an existing image and quickly perform a process of converting it into an alternative image. Deep learning Based image processing technology and its method are disclosed, and Korean Patent Registration No. 10-2132690 applies a face recognition pattern technology to select a high-resolution image restoration algorithm with the best restoration performance and a face recognition pattern technology to achieve ultra-high resolution Disclosed is an ultra-high resolution image restoration system in which an image is reconstructed, and when applied to a monitoring system, the high-resolution reconstructed image is similar to the original input image, and the high-resolution face recognition accuracy is closer to that of the original image.

However, conventional high-resolution image restoration systems have disadvantages in that they require a large amount of memory due to a large amount of data used as learning data, even if a computational process is complicated and deep learning technology is applied.

Therefore, a GAN-based artificial intelligence hologram image restoration method and system capable of restoring a hologram image with missing optical information into a flat image, but restoring a high-resolution image by selecting phase information and color information of the hologram image according to the user's request, and the system thereof research is needed on

The present invention provides a GAN-based artificial intelligence holographic image restoration method and system capable of generating a high-resolution image in which phase information, color information, etc. are restored from a holographic image with missing optical information by applying a GAN-based artificial intelligence system. It has its purpose.

In addition, networks for restoring the phase information, color information, and resolution of the holographic image are connected in series, and an estimated value for the restored image is calculated based on the output value of each network, and the estimated value is fed forward to the input of the output unit, so that the total The purpose is to provide a GAN-based artificial intelligence holographic image restoration method and system capable of improving restoration accuracy.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems to be solved by the present invention that are not mentioned here are to those of ordinary skill in the art to which the present invention belongs from the description below. will be clearly understood.

A GAN-based artificial intelligence holographic image restoration method and system according to an embodiment of the present invention applies a GAN algorithm to an input unit receiving a single holographic image from which optical field information is missing and a holographic image input to the input unit to obtain a restored image. Include an output that generates

In addition, the input unit is characterized in that it analyzes the data distribution of the hologram image to extract feature information, compresses the hologram image in the form of a feature map based on the extracted feature information, and stores it.

In addition, the output unit restores the hologram image compressed in the form of the feature map in the input unit, and the phase information restoration unit restores the phase information of the hologram image, the color information restoration unit restores the color information of the hologram image, and the hologram image. It includes a high-resolution restoration unit that restores the resolution of the hologram image and a restoration setting unit that receives at least one restoration setting of phase information, color information, and resolution for the hologram image, and the phase information restoration unit, the color information restoration unit, and the high-resolution restoration unit are serially It is characterized by being connected to.

The GAN-based artificial intelligence holographic image restoration method and system of the present invention have the effect of generating a high-resolution image in which phase information, color information, etc. are restored from a holographic image in which optical information is missing by applying a GAN-based artificial intelligence system. have

In addition, networks for restoring the phase information, color information, and resolution of the holographic image are connected in series, and an estimated value for the restored image is calculated based on the output value of each network, and the estimated value is fed forward to the input of the output unit, so that the total It has an effect of improving restoration accuracy.

1 is a block diagram of a GAN-based artificial intelligence holographic image restoration system according to an embodiment of the present invention.
2 is a diagram for explaining an output unit of a GAN-based artificial intelligence holographic image restoration system according to an embodiment of the present invention.

The specific details, including the problems to be solved, the means for solving the problems, and the effects of the invention for the present invention as described above are included in the embodiments and drawings to be described below. Advantages and features of the present invention, and methods for achieving them, will become clear with reference to the detailed description of the following embodiments in conjunction with the accompanying drawings.

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

1 is a configuration diagram of a GAN-based artificial intelligence holographic image restoration system according to an embodiment of the present invention, and FIG. 2 is a diagram for explaining an output unit of the GAN-based artificial intelligence holographic image restoration system according to an embodiment of the present invention. am.

<Example 1>

Referring to FIG. 1 , a GAN-based artificial intelligence holographic image restoration system 100 according to an embodiment of the present invention may include an input unit 110 and an output unit 120.

More specifically, the input unit 110 receives a single holographic image from which optical field information is missing, and the output unit 120 applies a GAN algorithm to the holographic image input to the input unit 110 to restore the image. can create

In addition, the input unit 110 may extract feature information by analyzing data distribution of the hologram image, and compress and store the hologram image in a feature map form based on the extracted feature information.

In this case, when the hologram image is a pathology image, the characteristic information may include various factors such as epidermal cells, nucleus size, texture, and lesion area.

Meanwhile, referring to FIG. 2 , the output unit 120 restores the hologram image compressed in the form of a feature map by the input unit 110, and the phase information restoration unit 121 for restoring phase information of the hologram image. ), a color information restoration unit 122 for restoring color information with respect to the hologram image, a high resolution restoration unit 123 for restoring the resolution of the holographic image, and the phase information, the color information and the phase information for the hologram image. It may include a restoration setting unit 124 that receives a restoration setting of at least one of the resolutions.

Also, the phase information restorer 121, the color information restorer 122, and the high resolution restorer 123 may be connected in series.

For example, even if the phase information restorer 121, the color information restorer 122, and the high resolution restorer 123 are connected in series, at least based on the restoration setting input to the restoration setting unit 124. A single restoration operation may be performed.

At this time, since the data input to the output unit 120 uses data compressed in the form of a feature map by the input unit 110, the phase information restoration unit 121, the color information restoration unit 122, and the high-resolution restoration are used. Even if the unit 123 is connected in series, the total amount of data calculation performed by the output unit 120 can be reduced, and accordingly, the data can be processed quickly.

In addition, since the phase information restorer 121, the color information restorer 122, and the high-resolution restorer 123 are configured in series instead of in parallel, one restorer 121 of the output unit 120 , 122, 123), even if an error occurs, it has the advantage of being easy to recover data before the error occurs.

Meanwhile, an estimated value for the reconstructed image is calculated based on the output values of the phase information restorer 121, the color information restorer 122, and the high resolution restorer 123, and the calculated estimate is calculated by the output unit. It can be fed forward to the input of (120).

Therefore, based on the estimated value, the output unit 120 can generate a more accurate original image.

For example, the output unit 120 generates a reconstructed image by applying a GAN algorithm, and the GAN algorithm may include a generator and a discriminator.

The creator creates a restored image similar to the hologram image input to the input unit 110, and the discriminator serves to distinguish the restored image by comparing the restored image generated by the creator with the hologram image (original image). carry out

At this time, the creator learns to create a restored image similar to the original image enough to deceive the discriminator, and the discriminator learns to more accurately distinguish the restored image created by the creator.

Therefore, the GAN-based artificial intelligence holographic image restoration system 100 can generate a high-resolution reconstructed image by the GAN algorithm in which the generator and the discriminator compete and learn.

According to the effects of the present invention as described above, by applying a GAN-based artificial intelligence system, GAN-based artificial intelligence holographic image restoration capable of generating a high-resolution image in which phase information, color information, etc. are restored from a holographic image with missing optical information. A method and its system may be provided.

In addition, networks for restoring the phase information, color information, and resolution of the holographic image are connected in series, and an estimated value for the restored image is calculated based on the output value of each network, and the estimated value is fed forward to the input of the output unit, so that the total A GAN-based artificial intelligence holographic image restoration method and system capable of improving restoration accuracy may be provided.

In addition, the GAN-based artificial intelligence holographic image restoration method according to an embodiment of the present invention may be recorded in a computer readable medium containing program instructions for performing various computer-implemented operations. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The medium may include program instructions specially designed and configured for the present invention, or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

As described above, although one embodiment of the present invention has been described by means of limited embodiments and drawings, one embodiment of the present invention is not limited to the above-described embodiments, which is based on common knowledge in the field to which the present invention belongs. Those who have it can make various modifications and variations from these materials. Therefore, one embodiment of the present invention should be grasped only by the claims described below, and all equivalent or equivalent modifications thereof will be said to belong to the scope of the present invention.

110: input unit
120: output unit 121: phase information restoration unit
122: color information restoration unit
123: high-resolution restoration unit
124: restoration setting unit

Claims (3)

an input unit for receiving a single holographic image in which optical field information is missing; and
an output unit generating a restored image by applying a GAN algorithm to the holographic image input to the input unit;
A GAN-based artificial intelligence holographic image restoration system including According to claim 1,
the input unit,
GAN-based artificial intelligence holographic image restoration system, characterized in that for extracting feature information by analyzing the data distribution of the hologram image, and compressing and storing the hologram image in the form of a feature map based on the extracted feature information. According to claim 2,
the output unit,
Restoring the hologram image compressed in the form of the feature map in the input unit,
a phase information restoration unit restoring phase information of the holographic image;
a color information restoration unit restoring color information of the holographic image;
a high-resolution restoration unit restoring a resolution of the holographic image; and
A restoration setting unit that receives at least one restoration setting of the phase information, the color information, and the resolution with respect to the holographic image;

The GAN-based artificial intelligence holographic image restoration system, characterized in that the phase information restoration unit, the color information restoration unit and the high-resolution restoration unit are connected in series.

Images