A feature-based coding mode prediction method is proposed to reduce the encoder complexity of the H.264 video coding standard in this work. The current H.264 reference codes employ exhaustive search to find the best m...
详细信息
ISBN:
(纸本)0819456586
A feature-based coding mode prediction method is proposed to reduce the encoder complexity of the H.264 video coding standard in this work. The current H.264 reference codes employ exhaustive search to find the best mode that optimizes the rate-distortion performance among all possible intra/inter predictive modes. To develop a fast binary intra/inter-modeprediction scheme, we consider the expected risk of choosing the wrong mode in a multi-dimensional feature space. The proposed algorithm consists of three steps. First, three features are extracted from the current macroblock. Second, the expected risk is calculated for the corresponding partition in the trained feature vector space, which is used to decide where the feature vector lies in one of the three regions depending on the expected risk, i.e. risk-free, risk-tolerable, and risk-intolerable regions. Finally, depending on the region to which the feature vector belongs, we can apply mechanisms of different complexities for final mode decision. For the likelihood estimation of the risk, parametric and non-parametric density estimation schemes are compared in terms of the rate-distortion-complexity tradeoff. It is demonstrated by experimental results that the proposed algorithm can save approximately 20-32% of the total encoding time of H.264 (JM7.3a) with little degradation in the rate-distortion performance.
This research investigates the forensics and anti-forensics of video editing operations, including shot insertion and deletion, which are common practices of content tampering. The proposed methodology is based on the...
详细信息
This research investigates the forensics and anti-forensics of video editing operations, including shot insertion and deletion, which are common practices of content tampering. The proposed methodology is based on the fact that excessive prediction residuals in the H.264/AVC encoding may appear in the inter-coded frames that were intra-coded. For the anti-forensics, the relationships of codingmodes in consecutive frames are employed to determine whether the intra-prediction can be applied during the re-encoding process of the tampered video. The correspondences of coding parameters and the bit-rates are then examined to predict the targeted distribution of quantization indices. Some nonzero indices are modified for satisfying the targeted distribution and generating a video that does not seem an edited one. Next, two forensic methods are presented. The first method evaluates the energy of resulting residuals after the deblocking filtering to reveal the editing operations that have been applied. The second method makes use of the rate control mechanism to check the quantization parameters. Although there exist certain limitations, the methods deserve more discussions under appropriate conditions. (C) 2015 Published by Elsevier Inc.
A fast feature-based intra/inter codingmode selection scheme for the H.264/AVC video coding standard is proposed in this paper. First, three features are extracted from a macroblock to form a feature space. Then, the...
详细信息
A fast feature-based intra/inter codingmode selection scheme for the H.264/AVC video coding standard is proposed in this paper. First, three features are extracted from a macroblock to form a feature space. Then, the feature space is partitioned into three regions, i.e., risk-free, risk-tolerable, and risk-intolerable regions, where the risk is calculated using the rate-distortion (RD) performance loss due to wrong mode decision as well as the probability distribution of inter/intra modes in the feature space obtained from an off-line training process. Depending on the region where the feature vector of a macroblock is located, we can apply mechanisms of different complexity for final mode decision. To calculate the likelihood function of the risk, both parametric and non-parametric density estimation schemes are developed to offer different rate-distortion-complexity tradeoffs. It is demonstrated by experimental results that the proposed algorithm can save approximately 20%-32% of the total encoding time of H.264 (JM7.3a) with little degradation in the rate-distortion performance.
暂无评论