KR102516812B1 - Method and system for optimization of image encode quality - Google Patents

Abstract

A method and system for optimizing image compression quality using machine learning are disclosed. The image compression optimization method includes: estimating an optimization quality that satisfies a target peak signal to noise ratio (PSNR) for an input image through a machine learning model that has learned image quality parameters and image characteristics; and outputting the image file encoded with the optimized quality as a compressed file for the input image.

Description

Method and system for optimizing image compression quality using machine learning {METHOD AND SYSTEM FOR OPTIMIZATION OF IMAGE ENCODE QUALITY}

The description below relates to image compression optimization techniques.

Among web contents transmitted over a network, image contents have the greatest influence on page loading speed.

The JPEG (Joint Photographic Experts Group) format is a lossy compression method and is widely used in a web environment because it can express 24-bit image pixels and has a high compression rate compared to other formats. In general, among the image formats distributed on the web, JPEG occupies a high proportion.

As an example of a JPEG compression method, Korea Patent Publication No. 10-2002-0035726 (published on May 15, 2002) "Still Image Compression and Restoration Method" compresses an input image signal using wavelet transform A technique is disclosed.

As the JPEG image format occupies a high proportion in the web environment, an optimization technique for improving the compression rate of JPEG images is required. The image-based service platform provides JAQ (JPEG adaptive quality) technology, a JPEG compression optimization method, as a JPEG encoding option.

The JAQ technology sets a criterion for not deteriorating image quality to the naked eye (PSNR 45.2db) and repeats the process of estimating the optimal parameters by lowering the quality parameters of JPEG.

Existing JAQ technology estimates optimal quality parameters through an iterative process, but especially in high-resolution images, processing time decreases.

Provided is a method and system capable of speeding up JAQ using SIMD (single instruction multiple data) dedicated to image processing.

We provide a method and system that can speed up JAQ through a single inference model without an iteration process by applying a machine learning (ML) framework.

A method for optimizing image compression executed in a computer device, the computer device comprising at least one processor configured to execute computer readable instructions contained in a memory, the method for optimizing image compression performed by the at least one processor , estimating an optimization quality that satisfies a target peak signal to noise ratio (PSNR) for an input image through a machine learning model that has learned image quality parameters and image characteristics; and outputting, by the at least one processor, an image file encoded with the optimized quality as a compressed file for the input image.

According to one aspect, the machine learning model may include a learning pipeline for inferring quality data corresponding to correct answer data and approximate values by taking a feature extractor for analyzing detailed components of an image as an input.

According to another aspect, the machine learning model may include a learning pipeline that refines learning data through data normalization and then learns a support vector regressor (SVR) model using the refined data.

According to another aspect, the SVR model may extract training metadata using a hold-out validation method or a grid-search validation method.

According to another aspect, the machine learning model may include an inference pipeline for estimating a quality parameter to be used for encoding the input image.

According to another aspect, the machine learning model may provide a normalization factor extracted from the learning data and learning metadata extracted from the SVR model as learning parameters of the inference pipeline.

According to another aspect, the estimating may include configuring the input image into a feature histogram using a uniform local binary pattern (U-LBP) method.

According to another aspect, the estimating step may further include adjusting the size of the input image to a predetermined size, before the step of constructing the feature histogram.

According to another aspect, single instruction multiple data (SIMD) may be applied to at least one machine learning component included in the machine learning model.

According to another aspect, the SIMD may be applied to at least one of an image resize process, a local binary pattern (LBP) construction process, and an image quality inference process as the machine learning component.

It provides a computer program stored in a computer readable recording medium to execute the image compression optimization method on the computer device.

A computer device, comprising at least one processor configured to execute computer readable instructions contained in a memory, wherein the at least one processor is configured to generate information about an input image through a machine learning model that has learned image quality parameters and image characteristics. estimating an optimization quality that satisfies a target PSNR; and outputting the image file encoded with the optimized quality as a compressed file for the input image.

According to embodiments of the present invention, processing time for image compression optimization can be improved by using SIMD to speed up an arithmetic function that is a bottleneck in JAQ.

According to the embodiments of the present invention, processing time for image compression optimization is reduced by applying a machine learning framework to speed up JAQ through a single inference model by learning the JPEG quality parameters and image feature values estimated from JAQ without an iterative process. can improve

1 is a block diagram for explaining an example of an internal configuration of a computer system according to an embodiment of the present invention.
2 to 5 show exemplary diagrams for explaining an example of a process of finding an optimization quality that satisfies a target PSNR.
6 shows a JAQ speedup method using machine learning in one embodiment of the present invention.
7 shows an example of a learning data set configuration according to an embodiment of the present invention.
8 is an exemplary diagram for explaining a cross-validation method according to an embodiment of the present invention.
9 is an exemplary diagram for explaining a process of generating an LBP image according to an embodiment of the present invention.
10 illustrates a sequence for applying JAQ using machine learning to an image processing server (IPS) according to an embodiment of the present invention.

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

Embodiments of the present invention relate to image compression optimization techniques.

Embodiments, including those specifically disclosed herein, can speed up computational functions that are bottlenecks in JAQ.

The present embodiments relate to a technology for speeding up JAQ, which is a JPEG compression optimization technology, and can be applied to an image processing server (IPS) of an image-based service platform.

1 is a block diagram illustrating an example of a computer system according to one embodiment of the present invention. For example, an image compression optimization system according to embodiments of the present invention may be implemented by the computer system 100 shown in FIG. 1 .

As shown in FIG. 1, a computer system 100 is a component for executing an image compression optimization method according to embodiments of the present invention, and includes a memory 110, a processor 120, a communication interface 130, and an input/output. Interface 140 may be included.

The memory 110 is a computer-readable recording medium and may include a random access memory (RAM), a read only memory (ROM), and a permanent mass storage device such as a disk drive. Here, a non-perishable mass storage device such as a ROM and a disk drive may be included in the computer system 100 as a separate permanent storage device separate from the memory 110 . Also, an operating system and at least one program code may be stored in the memory 110 . These software components may be loaded into the memory 110 from a recording medium readable by a separate computer from the memory 110 . The separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD/CD-ROM drive, and a memory card. In another embodiment, software components may be loaded into the memory 110 through the communication interface 130 rather than a computer-readable recording medium. For example, software components may be loaded into memory 110 of computer system 100 based on a computer program installed by files received over network 160 .

The processor 120 may be configured to process commands of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 120 by memory 110 or communication interface 130 . For example, processor 120 may be configured to execute received instructions according to program codes stored in a recording device such as memory 110 .

Communication interface 130 may provide functionality for computer system 100 to communicate with other devices via network 160 . For example, a request, command, data, file, etc. generated according to a program code stored in a recording device such as the memory 110 by the processor 120 of the computer system 100 is transferred to a network ( 160) to other devices. Conversely, signals, commands, data, files, etc. from other devices may be received into the computer system 100 via the communication interface 130 of the computer system 100 via the network 160 . Signals, commands, data, etc. received through the communication interface 130 may be transmitted to the processor 120 or memory 110, and files, etc. may be stored in the computer system 100. permanent storage).

The communication method is not limited, and may include not only a communication method utilizing a communication network (eg, a mobile communication network, wired Internet, wireless Internet, and broadcasting network) that the network 160 may include, but also short-distance wired/wireless communication between devices. there is. For example, the network 160 may include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , one or more arbitrary networks such as the Internet. In addition, the network 160 may include any one or more of network topologies including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, and the like. Not limited.

The input/output interface 140 may be a means for interface with the input/output device 150 . For example, the input device may include devices such as a microphone, keyboard, camera, or mouse, and the output device may include devices such as a display and a speaker. As another example, the input/output interface 140 may be a means for interface with a device in which functions for input and output are integrated into one, such as a touch screen. The input/output device 150 may be configured as one device with the computer system 100 .

Also, in other embodiments, computer system 100 may include fewer or more elements than those of FIG. 1 . However, there is no need to clearly show most of the prior art components. For example, the computer system 100 may be implemented to include at least some of the aforementioned input/output devices 150 or may further include other components such as a transceiver, a camera, various sensors, and a database.

In this embodiment, a technique for improving the processing time of JAQ, which is a JPEG compression optimization method, is provided.

First, a JAQ speedup technique using SIMD processing will be described.

SIMD is a type of parallel processing, in which multiple values are computed simultaneously with a single instruction. SIMD is mainly used in a multimedia field such as a graphic card, and IPP can be used as an example of SIMD below.

2 to 5 show exemplary diagrams for explaining an example of a process of finding an optimization quality that satisfies a target PSNR.

JAQ technology is a process for optimizing the compression quality required for an input image, and finds an optimization quality that satisfies a target PSNR (Peak Signal to Noise Ratio) within a quality range set based on compression quality. Encoding is performed by finding a quality similar to the compression quality of the image and at the same time having a PSNR equal to or greater than the target value. Optimization quality is sought within a lower quality range than compression quality.

PSNR is a significant index for comparing different images, but it can be said that it is not the same index as the quality that humans distinguish in that the human eye cannot distinguish the difference between the two images when the PSNR is below a certain level. Accordingly, the file size can be reduced through quality optimization by determining a PSNR (ie, a target PSNR) whose quality is indistinguishable from the human eye and reducing the compression quality required for the image to a quality that satisfies the target PSNR.

The processor 120 sets the maximum and minimum values of the quality range for finding the optimal quality. The maximum value is set to the compression quality required for the image. The minimum value is set to a quality value lower than the compression quality set to the maximum value. For example, the smallest value among quality values applicable to image compression may be set as the minimum value of the quality range. Assuming that the quality range applicable to image compression is 62 to 100 and the quality required by the user is 94, the quality range for finding the optimal quality may be set to 62 to 94.

Referring to FIG. 2, assuming that the quality range applicable to image compression is 62 to 100 and the quality required by the user is 94, the maximum value of the quality range for finding the optimal quality can be 94 and the minimum value can be 62. there is.

Optimization quality that satisfies the target PSNR can be found in the quality range of 62 to 94. For example, assume a target PSNR of 45.2.

Referring to FIG. 3 , in Step1, the processor 120 encodes an image with a quality of 78 corresponding to the average value of the quality range 62 to 94, and then extracts the PSNR through decoding. If the PSNR for the quality of 78 is 32, it is smaller than the target PSNR of 45.2, so the minimum value of the quality range is updated to 78 and the quality search process continues.

In Step 2, the processor 120 encodes the image with a quality of 86 corresponding to the average value of the quality range of 78 to 94, and then extracts the PSNR through decoding. If the PSNR for the quality of 86 is 38, it is smaller than the target PSNR of 45.2, so the minimum value of the quality range is updated to 86, and the quality search process continues.

In Step 3, the processor 120 encodes the image with a quality of 90 corresponding to the average value of the quality range 86 to 94, and then extracts the PSNR through decoding. If the PSNR for the quality of 90 is 44.1, it is smaller than the target PSNR of 45.2, so the minimum value of the quality range is updated to 90 and the quality search process continues.

In step 4, the processor 120 encodes the image with a quality of 92 corresponding to the average value of the quality range of 90 to 94, and then extracts the PSNR through decoding. If the PSNR for the quality of 92 is 45.7, the target PSNR is 45.2 or more. Therefore, the quality 92 is determined as an optimized quality similar to the quality 94 requested by the user and the target PSNR is 45.2 or more, and the quality search process and encoding process are finished.

The processor 120 may apply and output an image file encoded with an optimized quality as a final compressed file.

Therefore, by encoding the image by lowering the compression quality required for the image to a quality that satisfies the target PSNR, it is possible to visually obtain an image with a quality almost similar to the compression quality required for the image and at the same time reduce the compressed file size. there is.

Encoding is repeated to find the optimal quality. Referring to FIG. 4, one encoding process 400 includes an encoding process 410 for converting a JPEG file into YCbCr and converting the resulting value of the encoding process 410 into RGB. A decoding process 420, a YUV conversion process 430 converting the resultant value of the decoding process 420 into YUV, and a PSNR process 440 extracting PSNR using the resultant value of the YUV conversion process 430 includes

In the process of finding the optimal quality that satisfies the target PSNR, for example, as shown in FIG. 5, when a total of 4 encodings (Step1 to Step4) are performed, the encoding process 410 and the decoding process 420 for the image , YUV conversion process 430 and PSNR process 440 are repeated four times, respectively.

In this embodiment, some calculation functions that are bottlenecks in JAQ in the above process can be speeded up with SIMD.

The processor 120 speeds up JAQ implemented in an iterative loop using SIMD, and the color conversion (RGB to Gray) process 410 to 430 and the PSNR calculation process are the bottlenecks in JAQ implementation. In step 440, elements having parallelism can be speeded up through SIMD processing.

For example, the processor 120 may speed up the color conversion process and the PSNR calculation process applied to JAQ using SIMD. When mean squared error (MSE) is Equation 1, PSNR can be defined as Equation 2.

[Equation 1]

[Equation 2]

In the process of obtaining the MSE, the processing time of the for-loop increases as the image size increases. SIMD provides NormDiff related functions, and MSE can be accelerated with SIMD.

The for-loop operation in the PSNR operation process can be replaced with the NormDiff function provided by SIMD, and the last derivation of PSNR defined in Equation 2 is used to eliminate unnecessary sqrt/log operations.

As an example, SIMD can be applied to the JAQ PSNR computation process in units of whole images (Table 1).

AS-IS Input: A, B // two images
Output: psnr
Procedure getPSNR(A, B)
MSE := 0
for pixelA, pixelB in image of A, B
MSE := pow(pixelA - pixelB, 2)
end for
MSE := sqrt(MSE / imageSize)
psnr := 20.0 * log10(255.0/MSE)
End procedure TO-BE Input: A, B // two imagesOutput: psnr
Procedure PSNR(A, B)
A := vectorized A
B := vectorized B
mse := normDiff_L2(A, B) / imageSize
psnr := 48.1308036 - 10.0 * log10(mse)
End procedure

Therefore, processing time can be improved through speeding up by applying SIMD to some processes implemented as iterative loops in JAQ. Next, a JAQ speeding up technique using machine learning will be described.

In this embodiment, it is possible to improve the processing speed through a single inference model by removing the repetitive process that is the bottleneck of JAQ and learning the JPEG quality parameter estimated from JAQ and the feature value of the input image.

6 shows a JAQ speedup method using machine learning in one embodiment of the present invention.

Referring to FIG. 6 , the processor 120 includes a learning pipeline 610 and an inference pipeline 620 like a general machine learning framework.

The processor 120 may estimate a quality parameter of an image through a feature extraction process 622, a data normalization process 623, and an inference process (regression analysis) 624. .

In the feature extraction process 622, a U-LBP (uniform local binary pattern) method representing details of an image may be applied, and a data normalization process 623, which is a data refinement process to improve inference performance, - A zero-mean normalization method may be applied. In the inference process 624, a support vector regressor (SVR) inference model for estimating a JPEG quality parameter by analyzing feature values of an image may be applied.

The learning pipeline 610 includes answer data and feature extractors like a general supervised learning method for SVR learning. For example, the quality parameter of the existing JAQ can be configured as correct answer data, and the feature extractor of the input can be configured as U-LBP. The learning pipeline 610 operates to infer a value approximate to the quality data estimated by the existing JAQ, which is correct answer data, by taking U-LBP for analyzing detailed components of an image as an input.

The learning pipeline 610 proceeds with the SVR model learning process 614 using the feature values refined through the data normalization process 612 after loading the training data 611 . The SVR model learning process 614 may apply a learning methodology used to extract the learning meta of SVR, for example, a hold-out validation method, a grid-search validation method, and the like. can be used

In detail, in the training data loading process 611, the processor 120 may load a training data set in a CSV file or other format. Referring to FIG. 7 , a training data set can be configured by separating based on a delimiter (,) in the CSV file, and the first index of the CSV file can be configured with label data.

In the data normalization process 612, the processor 120 may refine the loaded training data using a zero-average normalization method (Table 2).

Input: trainData
Output: normalTrainData
Procedure zero_mean(trainData)
for i := 0 to trainData columns
mean, sigma := getMeanStdDev(trainData. column(i))
normalTrainData.column(i) := (trainData.column(i) - mean) / sigma
end for
End procedure

In the data separation process (split/shuffle) 613, the processor 120 may make the refined training data into data for verification (Table 3).

Input: totalTrainData, ratio
Output: trainData, testData
Procedure setTrainDataForValidation(totalTrainData, ratio)
totalTrainData := shuffle for train data
trainData, testData := splitData(totalTrainData, ratio) // split data
End procedure

In the verification process, which is the SVR model learning process 614, the processor 120 may estimate the hyperparameters of the SVR through a cross-validation method (eg, the cross-validation method shown in FIG. 8).

The learning pipeline 610 uses the normalization factors (average and variance of feature value data) extracted from the training data 611 and the SVR metadata extracted through the SVR model training process 614 to train the inference pipeline 620. provided as a parameter.

The inference process in the inference pipeline 620 applies the feature extraction process 622 and data normalization process 623 used in the training data construction like the learning pipeline 610 and uses it in the JPEG encoder 630 through SVR. Quality parameters can be estimated.

U-LBP used in the feature extraction process 622 is widely used in the field of computer vision, etc., and is a feature extractor that expresses detailed components of an image in 59 dimensions. Since the result of the JPEG quality parameter extracted from JAQ is operated according to the detailed component of the image, U-LBP can be used as a feature extractor in the inference pipeline 620.

U-LBP can configure the features of points, lines, and planes of an image as a 59-dimensional feature histogram. For example, referring to FIG. -Bit labels can be created. When the LBP image composed of a total of 256 labels for the input image is output, the LBP image can be created as a 59-dimensional histogram by changing the 256-dimensional to 59-dimensional.

Since U-LBP is not a scale-invariant feature extractor, the inference pipeline 620 is a preprocessing component to improve quality and processing time, and the resolution of the input image is appropriately tested through a test on the learning result. An image resizing process 621 for adjusting the size (eg, 512×512 resolution) may be added.

Depending on the embodiment, JAQ using machine learning can also apply SIMD to each machine learning component for an element that can be a bottleneck.

For example, SIMD resizing may be used in the image resizing process 621, that is, in the process of resizing an input image to a size of 512×512.

As another example, in the feature extraction process 622, the LBP construction process can be implemented with SIMD (Table 4).

Input : A // image
Output: histogram // feature result
Procedure Histogram (A)
LBPImage := GetLBPImageUsingSIMD(inputImage)
histogram :=GetHistogramUsingSIMD(LBPImage)
End procedure

As another example, SIMD optimization may be applied in the inference process 624 for obtaining the result of JAQ (Table 5). GetFastDotProduct calculates a general dot product through a loop statement, and GetFastDotProductUsingSIMD calculates a dot product using SIMD processing.

AS-IS Input : vectorA, vectorB
Output: scalarResult
Procedure Dotprod(vectorA, vectorB)
scalarResult := GetDotProduct(vectorA, vectorB)
End procedure TO-BE Input: vectorA, vectorB Output: scalarResult
Procedure Dotprod(vectorA, vectorB)
scalarResult := GetFastDotProductUsingSIMD(vectorA, vectorB)
End procedure

Therefore, processing time can be improved through speeding up by applying SIMD to machine learning components that can be bottlenecks in JAQ.

10 illustrates a sequence for applying JAQ using machine learning to an image processing server (IPS) according to an embodiment of the present invention.

Referring to FIG. 10, Reset() performs a function of initializing learning metadata to be used in a reasoning class (SVR) through regression analysis (first time performed) (1 to 2). Encode() of the image input class and GetJpegQuality() of the quality optimization class play a role in calling JAQ functions through each method (3~6). Internally, the JAQ function that estimates the quality parameters of the image is performed in the order of feature extraction process (FeatureExtractor), data normalization process (Normalizer), and inference process (Regressor) (7 to 12). Finally, image encoding (JpegEncoder) is performed using the quality parameter estimated through the JAQ function (13 to 14).

As such, according to embodiments of the present invention, processing time for image compression optimization can be improved by speeding up the process that is the bottleneck of JAQ using machine learning or SIMD. Through JAQ's high-speed, it is possible to increase usability and save storage and network bandwidth by improving the responsiveness of the server in the service to which the JPEG image optimization technology of the photo cloud platform is applied.

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

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

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

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

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

Claims (20)

An image compression optimization method executed on a computer device,
The computer device includes at least one processor configured to execute computer readable instructions contained in a memory;
The image compression optimization method,
estimating, by the at least one processor, an optimization quality that satisfies a target peak signal to noise ratio (PSNR) for an input image through a machine learning model that has learned image quality parameters and image characteristics; and
Outputting, by the at least one processor, an image file encoded with the optimized quality as a compressed file for the input image.
including,
The machine learning model is a single inference model of a machine learning framework in which the process of iteration to find the optimized quality is removed, and quality data corresponding to correct answer data using a feature extractor that analyzes detailed components of an image as input including a pipeline that infers values approximated by
Image compression optimization method characterized by. delete According to claim 1,
The machine learning model includes a learning pipeline for learning a support vector regressor (SVR) model using the refined data after refining the learning data through data normalization
Image compression optimization method characterized by. According to claim 3,
The SVR model extracts learning metadata using a hold-out validation method or a grid-search validation method.
Image compression optimization method characterized by. According to claim 3,
The machine learning model includes an inference pipeline for estimating a quality parameter to be used for encoding the input image.
Image compression optimization method characterized by. According to claim 5,
In the machine learning model, a normalization factor extracted from the training data and learning metadata extracted from the SVR model are provided as learning parameters of the inference pipeline.
Image compression optimization method characterized by. According to claim 1,
The estimating step is
Constructing the input image into a feature histogram using a U-LBP (uniform local binary pattern) method
Image compression optimization method comprising a. According to claim 7,
The estimating step is
Before configuring the feature histogram, adjusting the size of the input image to a predetermined size
Image compression optimization method further comprising a. According to claim 1,
Applying SIMD (single instruction multiple data) to at least one machine learning component included in the machine learning model
Image compression optimization method characterized by. According to claim 9,
Applying the SIMD to at least one of an image resize process, a local binary pattern (LBP) construction process, and an image quality inference process as the machine learning component.
Image compression optimization method characterized by. A computer program stored in a computer readable recording medium to execute the image compression optimization method of any one of claims 1, 3 to 10 in the computer device. In a computer device,
at least one processor configured to execute computer readable instructions contained in memory;
including,
The at least one processor,
estimating an optimization quality that satisfies a target PSNR for an input image through a machine learning model that has learned image quality parameters and image features; and
Outputting an image file encoded with the optimized quality as a compressed file for the input image
processing,
The machine learning model is a single inference model of a machine learning framework in which the process of iteration to find the optimized quality is removed, and quality data corresponding to correct answer data using a feature extractor that analyzes detailed components of an image as input including a pipeline that infers values approximated by
Characterized by a computer device. delete According to claim 12,
The machine learning model includes a learning pipeline for learning the SVR model using the refined data after refining the learning data through data normalization
Characterized by a computer device. According to claim 14,
The SVR model extracts learning metadata using a cross-validation method or a grid search validation method.
Characterized by a computer device. According to claim 14,
The machine learning model includes an inference pipeline for estimating a quality parameter to be used for encoding the input image.
Characterized by a computer device. According to claim 16,
In the machine learning model, a normalization factor extracted from the training data and learning metadata extracted from the SVR model are provided as learning parameters of the inference pipeline.
Characterized by a computer device. According to claim 12,
The at least one processor,
Constructing the input image into a feature histogram using the U-LBP method
Characterized by a computer device. According to claim 18,
The at least one processor,
Adjusting the size of the input image to a predetermined size before constructing the feature histogram
Characterized by a computer device. According to claim 12,
As at least one machine learning component included in the machine learning model, SIMD is applied to at least one of an image resizing process, an LBP construction process, and an image quality inference process.
Characterized by a computer device.

Images