Shixin Sun

New rate-distortion modeling and efficient rate control for H.264/AVC video coding

Rate control (RC) is crucial in controlling compression bit rates and qualities for networked video applications. In this paper, we propose a new rate-distortion (R-D) model and an efficient rate control scheme for H.264/AVC video coding, which elegantly resolve the inter-dependency problem between rate-distortion optimization and rate control by eliminating the need of coding complexity prediction for an inter-frame. The objective is to achieve accurate bit rate, obtain optimal video quality while reducing quality variations and simultaneously handling buffer fullness effectively. The proposed algorithm encapsulates a number of new features, including a coding complexity measure for intra-frames, a rate-distortion model, an accurate quantization parameter (QP) estimation for intra-frames, an incremental quantization parameter calculation method for inter-frames, a proportional+integral+derivative (PID) buffer controller, and an intelligent bit-allocation-balancing technique. Our experimental results demonstrate that the proposed scheme outperforms the JVT-G012 solution by providing accurate rate regulation, effectively reducing frame skipping, and finally improving coding quality by up to 1.80dB.

PID-Based Bit Allocation Strategy for H.264/AVC Rate Control

To achieve the best visual quality under the minimum bit rate and the limited buffer size, the rate control allocates appropriate and smooth bits to each frame. This brief proposes an effective bit-allocation strategy for the H.264/Advanced Video Coding rate control. Based on the different characteristics of intraframes and interframes, we introduce the different bit-allocation approaches for them, respectively. A proportional-integer-derivative controller is adopted to minimize the deviation between the target buffer level and the current buffer fullness. To avoid buffer overflow or underflow, a novel setting method for the bit-allocation boundary is presented. Experimental results demonstrate that the proposed bit allocation strategy achieves smooth target bits while better buffer control and visual quality are derived.

Image restoration using variational PDE-based neural network

Abstract In this paper we present two novel image restoration algorithms based on a modified Hopfield neural network and variational partial differential equations (PDE). The first algorithm is based on a harmonic model and the other is based on a total variation model. Both algorithms can restore the degraded images and preserve the edges. The experimental results show that our algorithms are better than other known neural network-based restoration algorithms, especially when images mainly consist of edges, such as letters.

New rate-complexity-quantization modeling and efficient rate control for H.264/AVC

In this paper, we propose a novel rate control scheme for H.264/AVC standard, including a new coding complexity measure for intra-frames, a new rate-complexity-quantization (R-C-Q) model, an accurate quantization parameter (QP) estimation for intra-frames, an incremental-control-based QP calculation for inter-frames, and a bit-allocation-balancing technique. Our experimental results demonstrate that, the proposed scheme outperforms the JVT-G012 solution by providing more accurate QP prediction, reducing frame skipping, depressing quality fluctuations, and finally, improving coding quality.

Incremental rate control for H.264 AVC scalable extension

The emerging H.264 Scalable Video Coding (H.264/SVC) requires the rate control algorithm to regulate the output bit rate of all the coarse-grain-scalability, temporal, spatial and combined enhancement layers. In order to address this topic, in this research, we propose an incremental rate control algorithm for H.264/SVC to control each layer’s encoding rate close to the target bit rate. The proposed algorithm introduces a number of efficient methods. First, based on our previous work on H.264/AVC rate control, a Rate-Complexity-Quantization (R-C-Q) model is extended in scalable video coding. Second, a complexity measure for Intra-frames based on their gradient and histogram information is used to precisely determine Quantization Parameters (QPs) for Intra-frames using the R-C-Q model. Third, we adopt an incremental approach to compute QPs of inter-frames. Fourth, a Proportional + Integral + Derivative (PID) buffer controller is presented to provide robust buffer control for each layer of H.264/SVC bitstream. Finally the QPs for hierarchical B-frames are adaptively decided by their neighbor inter-frames. Our extensive simulation results demonstrate that, our algorithm outperforms JVT-W043 rate control algorithm, adopted in the H.264/SVC reference software, by providing more accurate output bit rate for each layer, maintaining stable buffer fullness, reducing frame skipping finally, improving the overall coding quality.

Incremental Rate Control for H.264 Scalable Video Coding

The emerging H.264 Scalable Video Coding (H.264/SVC) requires the rate control algorithm to regulate the output bit rate of all the coarse-grain-scalability, temporal, spatial and combined enhancement layers. In this research, we propose an incremental rate control algorithm for H.264/SVC. First, a Rate-Complexity-Quantization (R-C-Q) model is extended in scalable video coding based on our previous work on H.264/AVC. Second, a complexity measure for Intra-frames is used to precisely determine QPs (Quantization Parameters) for Intra-frames. Finally, we adopt an incremental approach to compute QPs of inter-frames and a Proportional + Integral + Derivative (PID) buffer controller to provide robust buffer control for each layer. Our simulation results demonstrate that, our algorithm outperforms JVT-W043 rate control algorithm by providing more accurate output bit rate for each layer, maintaining stable buffer fullness, reducing frame skipping and quality fluctuation, finally, improving the overall coding quality.

Frame complexity prediction for H.264/AVC rate control

Rate control regulates the output bit rate of a video encoder in order to obtain optimum visual quality within the available network bandwidth and to maintain buffer fullness within a specified tolerance range. In this paper, we propose a novel rate control scheme for H.264/AVC video compression with a number of new features. We first introduce a calculation approach of frame complexity based on the linear prediction theory. Then, we propose a joint Rate-Distortion model which is an integration of a liner rate-complexity model and an exponential rate-quantization model. Finally, we develop an effective target bit estimation approach. Experimental results show that, compared with JVT-W042, our scheme achieves more accurate rate regulation, provides robust buffer control, efficiently reduces frame skipping, and improves visual quality.

Effective intra-only rate control for H.264/AVC

Rate control in H.264/AVC aims to achieve the best tradeoff between encoding quality and bandwidth while satisfying the buffer restriction. Due to the improving efficiency of intra-only rate control, we propose an effective rate control scheme for intra-only encoding. The proposed scheme employs a novel Rate-Distortion (RD) model, a new complexity measure, a precise Quantization Parameter (QP) calculation method, and a simple but effective model adaptation mechanism for intra-frames. Experimental results demonstrate that, compared with JVT-W042, the proposed algorithm achieves higher precise bit estimation, provides more robust buffer control, and improves coding quality.

Letters: A fast approximate algorithm for training L1-SVMs in primal space

We propose a novel and fast algorithm to train support vector machines (SVMs) in primal space, which solves an approximate optimization of SVMs with the properties of unconstraint, continuity and twice differentiability by utilizing the Newton optimization technique. Further, we devise a special pre-extracting procedure to speed up the convergence of the algorithm by resorting to a high-quality initial solution. Theoretical studies show that the proposed algorithm produces an @?-approximate solution to standard SVMs and maintains low computational complexity. Experimental results on benchmark data sets demonstrate that our algorithm is much faster than the dual based method such as SVM^l^i^g^h^t while it achieves the similar test accuracy.

Adaptive initial quantization parameter selection for H.264/SVC rate control

Initial quantization parameter (QP) selection is critical for the overall performance of video rate control. In this paper, we propose an adaptive initial QP selection algorithm for H.264/AVC Scalable Extension (SVC) rate control. The proposed algorithm introduces a number of efficient methods, including a coding complexity measure for intra-frames, and a rate-complexity-quantization (R-C-Q) model to accurately determine initial QPs for H.264/SVC rate control. Our experimental results demonstrate that, the proposed initial QP selection algorithm outperforms the one of JVT-W043 rate control scheme, adopted in the H.264/SVC reference software, by providing more accurate initial QP prediction, reducing buffer overflow/underflow, depressing quality fluctuations, and finally, improves the overall coding quality by up to 0.48dB.