Kook-Yeol Yoo

A Temporal Error Concealment Method for H.264/AVC Using Motion Vector Recovery

Recently H.264/AVC is rapidly deployed in the mobile multimedia market, such as terrestrial and satellite DMB (digital multimedia broadcasting), and Internet multimedia streaming systems. To provide the better quality under the unreliable communication environments, we propose a motion vector (MV) recovery method for the temporal error concealment. The H.264/AVC adopts the various block sizes for the motion estimation and compensation, ranging from 16times16 to 4times4 block sizes. To increase the accuracy in the temporal error concealment, the 4times4 block size is used as the MV recovery unit. Flexible MB ordering (FMO) option, by which the neighboring MBs can be transmitted in the different packets, is used. The MVs of the lost MBs are recovered based on the MV tendency which is derived from the neighboring MVs. The simulation results show that the proposed method improves video quality up to 2.95dB and 2.45dB, compared with the H.264/AVC test model and Lagrange interpolation method, respectively.

Low complexity intra prediction algorithm for MPEG-2 to H.264/AVC transcoder

Transcoding MPEG-2 video into H.264/AVC format is highly necessary for the broad distribution of digital video contents to mobile terminals. Since H.264/AVC achieves great coding efficiency by employing computationally demanding algorithms, the computational complexity of the transcoder is one of technically challenging issues. The empirical results show that DCT energy trend in the decoded MB of MPEG-2 bitstream has strong correlation with the Intra prediction modes of H.264/AVC. Based on the relationship, the mode skipping rule of intra prediction mode decision in the H.264/AVC encoder part of the transcoder is proposed in this paper. The simulation results show that the proposed algorithm can achieve on average 72.36% reduction in computational complexity, compared with that of the conventional transcoder.

Fast Interframe mode decision algorithm based on mode mapping and MB activity for MPEG-2 to H.264/AVC transcoding

Recently the latest video coding standard H.264/AVC is widely used for the mobile and low bitrate video codec in the various multimedia terminals. On the other hand, the MPEG-2 MP@HL codec has become the center of digital video contents since it is the standard codec for the Digital TV (DTV). To provide the bridge between the contents in MPEG-2 and mobile terminals, the transcoding of MPEG-2 contents into H.264/AVC format is an inevitable technology in the digital video market. The main bottleneck in the process lies in the computational complexity. In H.264/AVC, the variable block size (VBS) mode decision (MD) is used in the Interframe for the improved performance in the motion compensated prediction. For the macroblock (MB) which cannot be accurately predicted with one motion vector (MV), it is partitioned into smaller blocks and predicted with different MVs. In addition, SKIP and Intra modes are also permitted in the Interframe MD of H.264/AVC to further ameliorate the encoding performance. With the VBS MD technology, the Inter prediction accuracy can be improved significantly. However, the incidental side-effect is the high computational complexity. In this paper, we propose a fast Interframe MD algorithm for MPEG-2 to H.264/AVC transcoding. The relationships between SKIP and Intra modes are detected at first to map these two kinds of modes directly from MPEG-2 to H.264/AVC. And then the MB activity will be scaled by the residual DCT energy obtained from the MPEG-2 decoding process to estimate the block sizes of the MB mode for H.264/AVC Interframe MD. In our proposed method, the original redundant candidate modes can be eliminated effectively, resulting in the reduction of the computational complexity. It can reduce about 85% Rate-to-Distortion Cost (RDCost) computing and 45% entire processing time compared with the well-known cascaded transcoder while maintaining the video quality.

A study on the PAR reduction by hybrid algorithm based on the PTS and SLM techniques

An application of orthogonal frequency division multiplexing (OFDM) signal for high-bit rate transmission is limited by the high peak-to-average power ration (PAR). Recently, two techniques for improving the PAR statistics of OFDM signal have been proposed: selective mapping (SLM) and partial transmit sequences (PTS) methods. However, the complexity of these two methods has limited its PAR reduction capability. In this paper we analyze the problems of the conventional SLM and PTS schemes and propose a hybrid algorithm to reduce the complexity PAR reduction system by using the SLM method and the PTS method with the different phase searching algorithms. The results obtained indicate that the complexity of the system can be significantly reduced when the PAR statistics degradation is negligible. Also, we propose a threshold control algorithm (TCA) for the further reducing processing time in the scheme. We show that the hybrid system with the TCA has an additional capability for the use of a large number of side information bits resulting in better PAR reduction even compared to the PTS and SLM.

Robust hand segmentation and tracking to illumination variation

In this paper, we propose a hand segmentation and tracking algorithm under environment with illumination variation. The initial HSV histogram is computed by the rough HSV ranges of human skin and changed region. The histogram is updated by using the HSV pixel values in the current hand region and changed region. For the accurate update of histogram, other skin region is not used during histogram update. The empirical results show that the proposed method accurately tracks the predefined hand and effectively rules out other skin regions, such as the other hand or face.

A fast video mixing method for multiparty video conference

In this paper, we propose a fast video mixing method for reducing the computational complexity in the MCU (Multipoint Control Unit) used in the video conferencing. The conventional mixing method is based on the pixel-domain transcoder, of which computational complexity is linearly increased as the number of participants is increased. Basically the method requires many decoders and one huge encoder to mix the multiple bitstreams. To reduce the computational complexity, we propose a hybrid mixing method based on the syntax-based bitstream modification and pixel-domain transcoder. The proposed method reduces the computational complexity about 45% at the improved quality, compared with the conventional mixing method based on the pixel-domain transcoder.

Statistical Analysis and Derivation of Intra MB Mode Decision Rules for MPEG-2 to H.264/AVC Transcoding

In this paper, the DCT coefficients from the MPEG-2 decoder are used for the intra mode decision of H.264/AVC encoder. Based on the spatial activity analysis with the coefficients, early mode decision method between Intra16times16 and Intra4times4 modes and the Intra prediction direction are proposed. The empirical results show that the proposed method reduces about 70% computational complexity reduction compared with conventional method at the slight reduction in PSNR.

A fast full search algorithm for variable block-based motion estimation of h.264

In this paper, we propose a novel fast motion estimation algorithm based on successive elimination algorithm (SEA) which can dramatically reduce complexity of the variable block size motion estimation in H.264 encoder. The proposed method applies the conventional SEA in the hierarchical manner to the seven block modes. That is, the proposed algorithm can remove the unnecessary computation of SAD by means of the process that the previous minimum SAD is compared to a current bound value which is obtained by accumulating current sum norms and reused SAD of 4x4 blocks for the bigger block sizes than 4x4. As a result, we have tighter bound in the inequality between SAD and sum norm than the bound in the ordinary SEA. If the basic size of the block is smaller than 4x4, the bound will become tighter but it also causes to increase computational complexity, especially addition operations for sum norm. Compared with fast full search algorithm of JM of H.264, our algorithm saves 60 to 70% of computation on average for several image sequences.

Syntax-based mixing method of H.263 coded video bitstreams

This work presents a syntax-based mixing method, not using traditional transcoding methods, to remove the re-encoding process. The proposed algorithm only changes the syntax element and lossless encoded parameters, such as motion vectors, without loss of quality, at the expense of a slight increase of bitrate.

Laplacian model-based inverse quantization for DCT-based image codec system

Quantization is an essential step which leads to compression in discrete cosine transform (DCT) domain. In this paper, we show how a statistically non-optimal uniform quantizer can be improved by employing an efficient reconstruction method. For this method, we estimate the probability distribution function (PDF) of original DCT coefficients in a decoder. The proposed method can be used practically in any applications where uniform quantizers are used.