Quadratic Rate-distortion Model Research Articles

실시간 영상 부호화를 위한 H.264/AVC의 비트율 제어 하드웨어 설계

In this paper, the hardware design of rate control for real-time video encoded is proposed. In the proposed method, a quadratic rate distortion model with high-computational complexity is not used when quantization parameter values are being decided. Instead, for low-computational complexity, average complexity weight values of frames are used to calculate QP. For high speed and low computational prediction, the MAD is predicted based on the coded basic unit, using spacial and temporal correlation in sequences. The rate control is designed with the hardware for fast QP decision. In the proposed method, a quadratic rate distortion model with high-computational complexity is not used when quantization parameter values are being decided. Instead, for low-computational complexity, average complexity weight values of frames are used to calculate QP. In addition, the rate control is designed with the hardware for fast QP decision. The execution cycle and gate count of the proposed architecture were reduced about 65% and 85% respectively compared with those of previous architecture. The proposed RC was implemented using Verilog HDL and synthesized with UMC standard cell library. The synthesis result shows that the gate count of the architecture is about 19.1k with 108MHz clock frequency.

HEVC의 계층적 부호화 구조를 고려한 LCU 단위의 비트율 제어 기법

본 논문에서는 현재 표준화가 진행중인 HEVC(High Efficiency Video Coding)의 고정 비트율(CBR: Constant Bitrate) 부호화를 위한 비트율 제어(rate control) 기법을 다룬다. 기존의 H.264/AVC의 비트율 제어 기법에서는 계층적 부호화 구조의 특성을 고려하지 못한 제어 기법을 사용하여 정확한 비트율 제어에 어려움이 있다. 현재 HEVC의 참조 소프트웨어인 HM(HEVC test Model)에는 CBR 부호화를 위한 비트율 제어 알고리즘이 제공되지 않은 상태이며, 추후 비트율 제어 모듈이 추가될 것으로 예상된다. 이때 HM의 공통실험조건에 정의된 임의접근(Random Access: RA) 모드의 계층적 부호화 구조에서의 보다 정확한 비트율 제어가 요구된다. 따라서, 본 논문에서는 계층적 부호화 구조에서의 효율적인 비트율 제어를 위해 2차 비트율-왜곡(Rate-Distortion: R-D) 모델 기반의 시간계층 및 프레임 타입에 따른 비트율 특성을 반영한 비트율 제어 기법을 제시한다. 또한, 제어 단위에 따른 비트 변동율과 화질 손실 간에 발생하는 상충관계를 고려하여 비트율 제어 단위를 프레임 단위와 HEVC의 기본 부호화 단위(Coding Unit: CU)로 설정하고 모의실험을 통해 계층적 부호화 구조를 위해 제시된 기법의 비트율 제어 성능을 확인한다. In this paper, a method of rate control for constant bitrate (CBR) coding of High Efficiency Video Coding (HEVC) is addressed. The existing rate control of H.264/AVC may not provide exact rate control in the case of hierarchical coding structure since it doesn't consider the characteristics of the hierarchical coding structure. It is expected that a rate control is added to the reference software called HM for CBR encoding in the near future. More accurate rate control may be required in a hierarchical structure of random access (RA) mode defined in the common test condition of HM. In this paper, we propose a method of rate control based on quadratic Rate-Distortion (R-D) model considering temporal layers and frame types in hierarchical coding structure for efficient rate control. In the consideration of the trade-off relationship between the bit fluctuation and the average PSNR, both of frame and coding unit (CU) are set as the basic unit of rate control. The performance of the proposed rate control method is verified by simulations along with the trade-off relationships for the both cases of basic unit.

Bit allocation and rate control algorithm for MVC

Since the current multi-view video coding (MVC) software does not contain any rate control technique, this paper proposes a rate control algorithm for MVC based on the quadratic rate-distortion (R–D) model. The proposed algorithm classifies each picture into six frame types based on the relation between disparity prediction and temporal prediction estimation and also improves the estimation accuracy of the mean absolute difference (MAD). The proposed method allocates the bits and controls the rate for inter-view, frame layer and basic unit layer based on the analysis of the previously coded information. Compared to the multi-view video coding with fixed quantisation parameter, the proposed scheme achieves up to 0·25 dB improvements in peak signal-to-noise ratio (PSNR). Meanwhile, it can efficiently control the bit-rate with an average rate control error of 0·54%.

On Optimizing H. 264/AVC Rate Control by Improving R-D Model and Incorporating HVS Characteristics

The state-of-the-art JVT-G012 rate control algorithm of H.264 is improved from two aspects. First, the quadratic rate-distortion (R-D) model is modified based on both empirical observations and theoretical analysis. Second, based on the existing physiological and psychological research findings of human vision, the rate control algorithm is optimized by incorporating the main characteristics of the human visual system (HVS) such as contrast sensitivity, multichannel theory, and masking effect. Experiments are conducted, and experimental results show that the improved algorithm can simultaneously enhance the overall subjective visual quality and improve the rate control precision effectively.

Frame Coding Complexity Prediction in H.264/AVC Rate Control

H.264/AVC video coding standard inherited the quadratic rate-distortion model of VM8, and proposed a linear tracking model to predict Mean Absolute Difference of the current frame. Since Rate-Distortion Optimization is introduced, the frame coding complexity MAD is predicted in H.264/AVC rate control. However, any single-mode prediction approach of frame coding complexity has its shortages due to unexpected changes of video source. In this paper, we induct the frame coding complexity based on extensive experiments, and propose an optimized choice approach to predict the frame coding complexity. Simulation results demonstrate that the novel approach for frame coding complexity gains better precision than that of joint model in H.264/AVC reference software.

Frame-layer bit allocation for multi-view video coding based on frame complexity estimation

Current multi-view video coding (MVC) reference model in joint video team (JVT) does not provide efficient rate control schemes. This paper presents a rate control algorithm for MVC by improving the quadratic rate-distortion (R-D) model. We reasonably allocate bit-rate among views based on the correlation analysis. The proposed algorithm consists of three levels to control the rate bits more accurately, of which the frame layer allocates bits according to the frame complexity and the temporal activity. Extensive experiments show that the proposed algorithm can control the bit rate efficiently.

A Rate Control Scheme of the Even Low Bit-Rate Video Encoder

Rate control plays an important role in transmitting low-delay and high-quality images over the channel of very low bandwidth. The rate control algorithm in MPEG-4 or H.26X only defined the rate control model of P-frame, and did not introduce the rate control model of I-frame as it supposed that only the first frame is an I-frame, the others are all P-frames. However, in practical applications, a certain number of I-frames have to be inserted to meet the demand for fault-tolerant transmission and information retrieval. Oriented toward the application demand of remote surveillance system based on PSTN, a rate control strategy is proposed for encoding video at even low bit-rate in this paper. The strategy includes a rate-control algorithm for I-frame and introduces a quadratic rate distortion model, which solves the problem of the rate control used at very low bit-rate preferably. The rate distortion model of I-frame is designed referring to that of P-frame. It mainly involves three processes: initial quantization parameter calculation, objective bit number allocation, and objective quantization parameter calculation. TM5 of MPEG-2 proposed the target bit-number allocation model of I, P, and B frame, but the algorithm for the model is a bit complex. Based on the quadratic rate distortion model, a novel algorithm for the target bit-number allocation is presented in this paper. In order to keep the stability of image quality, VM8 limits the change range of the new quantization parameter from −0.25 to 0.25 times of the previous frame. As a result, if the previous frame's quantization parameter is small but the current frame's quantization parameter is calculated to be large, the current frame's quantization parameter should have been corrected with a small value, which will cause excessive coding bits, and worsen the transmission delay. In this paper, we present a correction algorithm for the computation of P-frame's quantization parameter. The results of experiments show that the proposed scheme reduces the delay of encoder buffering greatly, and improves the quality of reconstructed images clearly at the same time. In terms of the system's transmission delay, our scheme uses reasonable algorithms for computing the quantization parameters of I-frame and P-frame, and adjusting the GOP length dynamically. The transmission delays are significantly decreased and the buffer overflows are avoided. In terms of the reconstructed image quality, the test sequence's PSNRs with our scheme are higher than VM8's at all the bit rates, and the coded bit-number of our scheme is less than VM8's at the same bit rate. In terms of the deviation of output bit-streams, the average deviation of our algorithm is much less than VM8's. The less the deviation of the output bit-stream is, the more appropriate it is to be transmitted at the given bandwidth, and the less the transmission delay is.

Adaptive rate control for H.264

This paper presents a rate control scheme for H.264 by introducing the concept of basic unit and a linear prediction model. The basic unit can be a macroblock (MB), a slice, or a frame. The linear model is used to predict the mean absolute differences (MADs) of the remaining basic units in the current stored picture by those of the co-located basic units in the previous stored picture. The target bits for the current stored picture are computed by adopting a fluid flow traffic model and linear tracking theory, and are further bounded by two values that are derived by taking the hypothetical reference decoder (HRD) into consideration. The remaining bits are allocated to the remaining basic units in the current stored picture according to their predicted MADs. The corresponding quantization parameter is computed by using a quadratic rate-distortion model. The rate distortion optimization (RDO) is then performed for all MBs in the current basic unit by the quantization parameter. Both constant bit rate and variable bit rate cases are studied. The average PSNR is improved by up to 0.8 dB for an encoder using our scheme compared to an encoder using a fixed quantization parameter. With our scheme, an H.264 encoder can be adaptive to time varying channel bandwidth that is available for the coding process.

Rate-constrained arbitrarily shaped video object coding with object-based rate control

The authors develop an object-based rate control algorithm for MPEG-4 arbitrarily shaped video object coding. There are three main contributions. An improved quadratic rate-distortion model for texture information is derived, which is more accurate for arbitrarily shaped video objects. Secondly, an accurate rate-distortion model for shape information is introduced. Combining the rate-distortion models for texture and shape information, an object level rate control is proposed. To achieve constant bit rate, the quantiser is adjusted based on buffer fullness. The tradeoff between picture quality and buffer fullness is considered. Compared with the rate control algorithm for the MPEG-4 verification model, the proposed scheme can achieve a more stable buffer without sacrificing picture quality.

Advanced rate control technologies for 3D-HDTV

This paper presents advanced rate control technologies for 3D-HDTV. We remodel the quadratic rate-distortion model for 3D multiview sequences based on their picture types. The proposed method consists of two levels for more accurate bits rate control. In a frame level, it sets quantization parameters based on HVS (human visual system) and reduces error bits by remodeling rate-quantization. In a MB level, rate control is activated to provide stricter buffer regulations and higher bit rate encoding based on the target bits and the quantization parameters calculated at the frame level. The proposed algorithm shows improvements in rate control as well as in PSNR compared to the conventional methods. It also provides more accurate rate control and higher image quality than TM5 and TMN8 at various bit rates.

Improved single-video-object rate control for MPEG-4

Rate control plays a central role in constant bit-rate video coding applications using MPEG-4. This paper considers single-video-object rate control for MPEG-4 and presents a new rate-control algorithm based on the quadratic rate-distortion (R-D) model. The major innovations are a novel constraint for the least-mean-square estimation of the model parameters of the R-D function, a new measure for the encoding complexity, novel quantizer control and an efficient frame skipping strategy. Simulation results show that the MPEG-4 coder, using the proposed algorithm, can achieve a higher PSNR than a coder using the conventional rate-control algorithm.

A new rate control scheme using quadratic rate distortion model

A new rate control scheme is used to calculate the target bit rate for each frame based on a quadratic formulation of the rate distortion function. The distortion measure is assumed to be the average quantization scale of a frame. The rate distortion function is modeled as a second-order function of the inverse of the distortion measure. We present a closed form solution for the target bit allocation which includes the MPEG-2 TM5 rate control scheme as a special case. The model parameters are estimated using statistical linear regression analysis. Since the estimation uses the past encoded frames of the same picture prediction type (I, P, B pictures), the proposed approach is a single pass rate control technique. Because of the improved accuracy of the rate distortion function, the fluctuations of the bit counts are significantly reduced by 20-65% in the standard deviation of the bit count while the picture quality remains the same. Thus, the buffer requirement is reduced at a small increase in complexity. This technique has been adopted by the MPEG committee as part of VM5.0 in November 1996.