KR102163007B1 - Method and Apparatus for Adjusted Prediction Mode Decision Scheme to Speed Up of Low Contrast Video Encoding - Google Patents

Abstract

A method and apparatus for optimizing prediction mode determination to improve compression speed of low-light images are presented. The prediction mode determination optimization adjustment method for improving the compression speed of a low-illuminance image proposed in the present invention includes determining an optimal mode through a coding unit, determining whether the optimal mode is a SKIP mode, and the optimal mode is not a SKIP mode. In this case, comparing the SKIP mode cost with the lambda-adjusted cost of the optimal mode, and if the SKIP mode cost is not less than the lambda-adjusting cost of the optimal mode, performing splitting and encoding.

Description

{Method and Apparatus for Adjusted Prediction Mode Decision Scheme to Speed Up of Low Contrast Video Encoding}

The present invention relates to a method and apparatus for optimizing prediction mode determination for improving a compression speed of a low-illuminance image.

In order to store video information in various video-based surveillance systems, video compression is essentially used. Video encoders of the H.264 and HEVC standards, which are widely used, have very high compression rates, but also high computational complexity and power consumption. In order to improve this, various complexity reduction studies have been conducted. Existing studies have suggested techniques that consider the situation with sufficient illumination. However, since the surveillance system operates 24 hours a day, both bright and dark images are acquired. Since the time without lighting accounts for more than 40% of the day, it is necessary to study the encoding algorithm that actively considers the characteristics of low-light images. In general, encoding algorithms have advantages in compression rate and processing speed when they are highly correlated in space and time. If the characteristics of an image whose spatiotemporal correlation increases due to low illuminance are utilized, the encoding time in the low illuminance image may be further reduced, and as a result, power consumption may be reduced.

An object of the present invention is to provide an encoding algorithm for further reducing the encoding time and consequently reducing power consumption in low-light images, and a method and apparatus for optimizing prediction mode determination for improving the compression speed of low-light images.

In one aspect, the prediction mode determination optimization adjustment method for improving the compression speed of a low-illumination image proposed by the present invention includes determining an optimal mode through a coding unit, determining whether the optimal mode is a SKIP mode, and the optimal mode. Is not in the SKIP mode, comparing the SKIP mode cost with the lambda-adjusted cost of the optimal mode, and if the SKIP mode cost is not less than the lambda-adjusting cost of the optimal mode, splitting and encoding are performed. Includes steps.

When the optimal mode is the SKIP mode, the prediction process is terminated without further segmenting the CU using an ECU (Early Coding Unit termination) algorithm.

If the optimal mode is not the SKIP mode, the step of comparing the SKIP mode cost and the lambda-adjusted cost of the optimal mode calculates the SKIP mode cost using the following equation,

Here, D stands for distortion, R stands for rate, and λ stands for lambda multiplier.

In order to improve the extrusion rate by increasing the encoding speed, λ is adjusted in the low-illuminance image, and the SKIP mode cost and the lambda-adjustment cost of the optimal mode are recalculated as the corrected λ value.

In another aspect, the apparatus for determining an optimization mode for determining a prediction mode for improving a compression speed of a low-illuminance image proposed by the present invention determines an optimal mode through a coding unit, and a determination unit that determines whether the optimal mode is a SKIP mode, the When the optimal mode is not the SKIP mode, a comparison unit that compares the SKIP mode cost and the lambda-adjusted cost of the optimal mode, and when the SKIP mode cost is not less than the lambda-adjusted cost of the optimal mode, splitting and encoding It includes an encoding unit that performs.

According to embodiments of the present invention, an encoding algorithm and a method and apparatus for optimizing a prediction mode determination for improving a compression speed of a low-illuminance image may further reduce an encoding time in a low-light image and consequently reduce power consumption.

1 is a flowchart of a prediction mode determination optimization adjustment algorithm for improving a compression speed of a low-light image according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a configuration of a prediction mode determination optimization and adjustment apparatus for improving a compression speed of a low-light image according to an embodiment of the present invention.
3 is a graph showing PSNR according to bitrate for each algorithm 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.

The overall brightness of a low-light image is dark, and it is difficult to clearly grasp the shape and color information of an object. As a result, the speed of the smooth area increases and the spatial complexity decreases. Existing HEVC encoding speed improvement algorithms focus on a normal brightness image. A typical example is the speed-up algorithm built into the HEVC reference software. Among them, ECU termination (Early Coding Unit termination) is an algorithm that terminates the prediction process without further segmenting the CU when the optimal prediction mode is the SKIP mode during prediction mode execution. As the speed determined by SKIP mode increases, the encoding speed also increases. In the optimal mode of HEVC, the compression cost is determined in consideration of the rate and distortion. The prediction mode with the least cost is determined as the optimal mode. The equation for calculating the prediction mode cost is as follows.

수학식(1)

Equation (1)

Here, D is the distortion, R is the rate, and λ is the lambda multiplier. As the quantization parameter (QP) increases, the more quantization is performed, the less detailed the image. In this case, since reducing the speed term has a greater effect on the compression efficiency, λ also increases in proportion to the QP value. It can be expressed as λ = 0.85 Х 2(QP-12)/3. This equation is often used, and constant values may vary slightly depending on the prediction mode.

In the existing H.264 or HEVC standard, λ is determined only by QP as shown in the above equation, and video content is not considered. There are studies that improve the compression rate by adjusting λ in consideration of image characteristics or HVS (Human Visual System), but the computational complexity required for determining the optimal mode is very high, so it is difficult to be widely applied.

In the present invention, a technique for improving an encoding speed by adjusting λ in a low-light image is proposed. For this, the determination conditions of the existing ECU algorithm were modified.

1 is a flowchart of a prediction mode determination optimization adjustment algorithm for improving a compression speed of a low-light image according to an embodiment of the present invention.

Determining an optimal mode through a coding unit (110), determining whether the optimal mode is a SKIP mode (120), if the optimal mode is not a SKIP mode, a SKIP mode cost and a lambda of the optimal mode -Comparing (130) the adjustment cost, and if the SKIP mode cost is not less than the lambda-adjustment cost of the optimal mode, performing split and encoding (140).

First, in step 110, the existing ECU condition is checked. After performing all the prediction modes in one CU, the optimal prediction mode is checked. In step 120, it is determined whether the optimal mode is the SKIP mode.

In step 130, if the optimal prediction mode is the SKIP mode, the CU splitting and the subsequent prediction process are terminated. The SKIP mode is a mode in which a prediction block is sufficiently similar to the current block, so that a value after transforming and quantizing the residual, which is the difference between the two blocks, becomes 0, so that no information needs to be sent. The judgment condition is that even if the speed becomes 0 at cost = distortion + λ * rate, if the distortion is sufficiently small and has the lowest cost than other prediction modes, it is judged that the SKIP mode may be determined.

If the existing ECU (Early CU Termination) condition is not satisfied, the cost of the optimal prediction mode and the SKIP mode is recalculated and compared with the modified λ value. The cost at this time is called lambda-adjusted cost. The existing ECU condition is when the current CU optimal prediction mode is the SKIP mode.

In step 140, if the SKIP mode cost is not less than the lambda-adjusted cost of the optimal mode, splitting and encoding are performed. The corrected λ is a value calculated as QP+N, not the current QP. The current QP is a quantization parameter for the current block determined by a user or determined from a bitrate control algorithm. In QP + N, N is a positive number, meaning that a slightly larger value is used.

The focus was on the fact that when a general video was encoded with high QP and when the illuminance was lowered, it showed similar video characteristics. If the cost of the SKIP mode is smaller than that of the optimal mode compared to the lambda-adjusted cost, the CU splitting is stopped.

FIG. 2 is a diagram illustrating a configuration of a prediction mode determination optimization and adjustment apparatus for improving a compression speed of a low-light image according to an embodiment of the present invention.

The apparatus 200 for optimizing the determination of a prediction mode for improving the compression speed of a low-illuminance image according to the present invention includes a determination unit 210, a comparison unit 220, and an encoding unit 230.

The determination unit 210 determines an optimal mode through the coding unit, and determines whether the optimal mode is a SKIP mode. When the optimal mode is the SKIP mode, the determination unit 210 terminates the prediction process without further dividing the CU using an early coding unit termination (ECU) algorithm.

When the optimal mode is not the SKIP mode, the comparison unit 220 compares the SKIP mode cost and the lambda-adjustment cost of the optimal mode. The comparison unit 220 calculates the SKIP mode cost using Equation (1) as described above. In order to improve the extrusion rate by increasing the encoding speed, λ is adjusted in the low-illuminance image, and the SKIP mode cost and the lambda-adjusted cost of the optimal mode are recalculated as a modified λ value.

When the SKIP mode cost is not less than the lambda-adjustment cost of the optimal mode, the encoding unit 230 performs splitting and encoding.

3 is a graph showing PSNR according to bitrate for each algorithm according to an embodiment of the present invention.

In order to verify the performance of the algorithm proposed in the present invention, low-light images BasketballDrive_dark and BasketballDrive_dark were created using BasketballDrive and BQTerrace images. Encoding of 60 frames was performed for each video with the HM_16.10 reference software in the low-delay main setting. 3 shows the compression efficiency when the ECU and the proposed algorithm are applied compared to the reference software to which the speed improvement algorithm is not applied. For the proposed algorithm, the QP+3 setting was used. Table 1 shows TS (Time Saving) together with BDBR.

<Table 1>

It is shown that the proposed algorithm can achieve an encoding speed improvement with a very small reduction in compression efficiency.

According to embodiments of the present invention, an encoding algorithm and a method and apparatus for optimizing a prediction mode determination for improving a compression speed of a low-illuminance image may further reduce an encoding time in a low-light image and consequently reduce power consumption.

The apparatus described above may be implemented as a hardware component, a software component, and/or a combination of a hardware component and a software component. For example, the devices and components described in the embodiments include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It can be implemented using one or more general purpose computers or special purpose computers, such as a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications executed on the operating system. In addition, the processing device may access, store, manipulate, process, and generate data in response to the execution of software. For the convenience of understanding, although it is sometimes described that one processing device is used, one of ordinary skill in the art, the processing device is a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of these, configuring the processing unit to behave as desired or processed independently or collectively. You can command the device. Software and/or data may be interpreted by a processing device or to provide instructions or data to a processing device, of any type of machine, component, physical device, virtual equipment, computer storage medium or device. Can be embodyed in The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording 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 in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, 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. -A hardware device specially configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described by the limited embodiments and drawings, various modifications and variations are possible from the above description by those of ordinary skill in the art. For example, the described techniques are performed in a different order from the described method, and/or components such as a system, structure, device, circuit, etc. described are combined or combined in a form different from the described method, or other components Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and claims and equivalents fall within the scope of the claims to be described later.

Claims (8)

Determining an optimal mode through the coding unit;
Determining whether the optimal mode is a SKIP mode;
If the optimal mode is not a SKIP mode, comparing a SKIP mode cost and a lambda-adjusted cost of the optimal mode; And
If the SKIP mode cost is not less than the lambda-adjusted cost of the optimal mode, performing splitting and encoding
Including,
If the optimal mode is not the SKIP mode, comparing the SKIP mode cost and the lambda-adjusted cost of the optimal mode,
Calculate the SKIP mode cost using the following equation,

Where D is distortion, R is rate, λ is lambda multiplier,
When an image is encoded with a quantization parameter higher than a predetermined threshold value, a low illuminance image characteristic with an illuminance lower than a predetermined threshold value is used, and in order to increase the extrusion rate by increasing the encoding speed, adjust λ,
The SKIP mode cost and the lambda-adjusted cost of the optimal mode are recalculated and compared with the corrected λ value, and if the SKIP mode cost is less than the cost of the optimal mode, the CU division is stopped.
Prediction mode decision optimization adjustment method. The method of claim 1,
When the optimal mode is the SKIP mode, the prediction process is terminated without further segmenting the CU using an ECU (Early Coding Unit termination) algorithm.
Prediction mode decision optimization adjustment method. delete delete A determination unit determining an optimal mode through a coding unit and determining whether the optimal mode is a SKIP mode;
A comparison unit comparing a SKIP mode cost and a lambda-adjustment cost of the optimal mode when the optimal mode is not a SKIP mode; And
If the SKIP mode cost is not less than the lambda-adjusted cost of the optimal mode, the encoding unit performs splitting and encoding
Including,
The comparison unit,
Calculate the SKIP mode cost using the following equation,

Where D is distortion, R is rate, λ is lambda multiplier,
When an image is encoded with a quantization parameter higher than a predetermined threshold value, a low illuminance image characteristic with an illuminance lower than a predetermined threshold value is used, and in order to increase the extrusion rate by increasing the encoding speed, adjust λ,
The SKIP mode cost and the lambda-adjusted cost of the optimal mode are recalculated and compared with the corrected λ value, and if the SKIP mode cost is less than the cost of the optimal mode, the CU division is stopped.
Prediction mode decision optimization adjustment device. The method of claim 5,
The determination unit,
When the optimal mode is the SKIP mode, the prediction process is terminated without further segmenting the CU using an ECU (Early Coding Unit termination) algorithm.
Prediction mode decision optimization adjustment device. delete delete

Images