Chi-Wang Ho

Spatio-temporal boundary matching algorithm for temporal error concealment

In this paper, a novel temporal error concealment algorithm, called spatio-temporal boundary matching algorithm (STBMA), is proposed to recover the information lost in the video transmission. Different from the classical boundary matching algorithm (BMA), which just considers the spatial smoothness property, the proposed algorithm introduces a new distortion function to exploit both the spatial and temporal smoothness properties to recover the lost motion vector (MV) from candidates. The new distortion function involves two terms: spatial distortion term and temporal distortion term. Since both the spatial and temporal smoothness properties are involved, the proposed method can better minimize the distortion of the recovered block and recover more accurate MV. The proposed algorithm has been tested on H.264 reference software JM 9.0. The experimental results demonstrate the proposed algorithm can obtain better PSNR performance and visual quality, compared with BMA which is adopted in H.264

Motion Estimation for H.264/AVC using Programmable Graphics Hardware

We present an efficient implementation of motion estimation (ME) for H.264/AVC using programmable graphics hardware. The cost function for ME in H.264/AVC depends on the motion vector (MV) predictor which is the median MV of three neighboring coded blocks. Previous implementations assume no dependency among adjacent blocks, which is not true for H.264/AVC, they also perform unsatisfactorily because of their low arithmetic intensity, which is defined as operation per word transferred. To overcome the dependency problem, we introduce a new implementation which performs ME on block-by-block basis. Moreover, we can adjust the arithmetic intensity easily to optimize the performance on different graphics cards. Experimental results show that our implementation is substantially faster (by 10 times) than our SIMD optimized CPU implementation

Enhanced predictive motion vector field adaptive search technique (E-PMVFAST)-based on future MV prediction

Motion estimation (ME) is a core part of most modern video coding standard, and it directly affects the compression efficiency and visual quality of a video. If full search (FS) algorithm is used, ME could takes over 70% of computational power. Many algorithms such as TSS and PMVFAST, have been developed to achieve great speed up for ME. In this paper, a new algorithm enhanced-PMVFAST (E-PMVFAST) is proposed, which performs better than most if not all other existing algorithms in terms of speed up factor while keeping the similar PSNR with FS. Our experiments have also verified the robustness of the proposed E-PMVFAST algorithm.

Fast mode decision and motion estimation for H.264 (FMDME)

The H.264 video coding standard achieves highest coding gain. In addition to common tools in previous standards such as H.263 and MPEG4, it supports multiple frames, multiple block sizes, and 1/4 pixel motion estimation, deblocking filter, intra-prediction, and so on. However, full search of for all block modes and reference frames requires heavy computational complexity. The algorithm predictive motion vector field adaptive search technique (PMVFAST) (Tourapis, 2000) was previously accepted into MPEG standard to achieve hundreds of time of speed up. FMBME (Chang, 2004), and some other algorithms (Yanfei Shen, 2004) have been developed to do fast multiple block size motion estimation. For multiple reference frame motion estimation, FMFME (Chang, 2003), and some other algorithms (Xiang Li, 2004) have also been developed. In this paper a new algorithm is presented named as fast mode decision and motion estimation (FMDME). This algorithm achieves high speed up factor by involving fast skip mode checking, fast multiple block size motion estimation, fast multiple reference frame selection, fast motion estimation, and fast intra-block skipping for P frame in H.264

Low-Complexity Rate Control for Efficient H.263 to H.264/AVC Video Transcoding

Rate control is a complicated problem in the H.264/AVC coding standard, extra computation is usually needed for the existing rate control schemes to estimate the complexity of frames or macroblocks (MBs). However, during transcoding, information from preceded video could be used to simplify the rate control. In this paper, we propose a low-complexity rate control scheme for transcoding from H.263 to H.264/AVC. The relationship between the rate of the pre-coded video and both the rate and distortion of the transcoded video are studied. By using only the rate information from the preceded video, we introduce a row-layer bit allocation and perform average rate shaping across a row of MBs. Estimation error diffusion is also introduced. The proposed scheme has sufficiently lower computational complexity than other methods as there is no explicit complexity measurement of MBs and complicated parameters updating. Experimental results show the effectiveness of the proposed scheme.

Content adaptive watermarking using a 2-stage predictor

Digital watermarking is one of the solutions to protect intellectual properties and copyright by hiding information, such as a random sequence or a logo, into digital media. In this paper, a new watermarking scheme is proposed. The embedding process takes place in the spatial domain. A bi-level logo is embedded into digital media by comparing the absolute difference between the original pixel value and the predicted pixel value, which is computed and predicted by using the concept of activity measurement, followed by spatial varying filter (SVF). The logo will be extracted from a possibly corrupted image, without the help of original uncorrupted image. The new proposed algorithm can withstand the geometric attacks as well as common signal processing, and has a high payload capacity.

Improved refinement search for H.263 to H.264/AVC transcoding based on the minimum cost tendency search

An improved refinement search method for transcoding from H.263 to H.264/AVC is proposed in this paper. Many existing motion re-estimation methods refine the input motion vector (MV) with a small search range, which is usually input MV biased. Motion estimation (ME) in H.263 usually does not consider the rate required for coding the MV, and hence, the input MV may incur a large cost in H.264/AVC. To overcome this problem, we introduce a refinement search method, called minimum cost tendency search (MCTS), which takes the difference between the cost functions for ME in H.263 and H.264/AVC into consideration. The input MV and the predictor MV are used as two anchor points. The proposed MCTS starts searching from the anchor point with a higher cost to another. Finally, the best point is chosen as the center for further refinement. The performance of MCTS is evaluated by comparing with full search, FME in JM software and refinement scheme using small diamond pattern around the input MV (RSD). Experimental results show the proposed MCTS performs more stable than FME and RSD over a wide range of output video quality

Rate Control Based on Zero-Residue Pre-Selection for Video Transcoding

A common issue in video transcoding for heterogeneous network environment is to efficiently and accurately reduce the bit-rate such that the distortion is minimized under a given rate constraint. To convert the bit-rate of an encoded video to match the channel capacity, in general, re-quantization is done on the DCT coefficients with larger quantization step size. Most existing rate control algorithms for video transcoding in the literature calculate quantization parameters (QPs) of macroblocks (MBs) based on a relationship between certain properties of coded video and bit-rate. They reduce the computational complexity by simplifying the R-D model and reusing the statistics information of input video. In this paper, we propose a zero-residue pre-selection (ZRPS) mechanism to select only a portion of MBs to apply the rate control in video transcoding. TMN-8 is used to evaluate the impact of ZRPS. Experimental results show that, as compared to the original TMN-8 rate control scheme, TMN-8 with ZRPS achieves up to 1.60 dB gain, in term of PSNR, and requires less than 50% of the computational complexity compared to TMN-8, depending on the characteristics of the video content