Seishi Takamura

MPEG-2 one-pass variable bit rate control algorithm and its LSI implementation

MPEG-2, the international standard of the ISO/IEC, is a key technology for video coding and is widely used in digital television etc. Its most frequently used coding rate control is the constant bit rate (CBR) mode. CBRs bandwidth requirement is relatively low, but it needs further improvement in terms of the coding efficiency. On the other hand, the variable bit rate (VBR) mode may be able to enhance the picture quality in situations that are rich in bandwidth or in storage media such as DVD. We have already developed a one-chip MPEG-2 video encoder LSI (SuperENC) to cope with the broader range of digital image communication applications. This paper discusses a new one-pass VBR coding control algorithm based on mathematical relations among coding parameters such as quantization parameter, bit rate and decoded image distortion. Our aim is to solve the problems of conventional techniques and verify the algorithms effectiveness through hardware implementation. A one-pass VBR control function was implemented in the chips firmware. Not being a very complex algorithm, it is robust with respect to abrupt scene changes. The SNR gain in a TM-5-based CBR mode was 1.5 to 1.8 dB, and the bit rate reduction under the same subjective quality was 10-30 %.

Intra prediction with spatial gradients and multiple reference lines

This paper proposes an improved intra prediction method for H.264/AVC. The conventional method uses only adjoining pixels as reference pixels. The residual energy is increased if the decoded pixels are occluded, which would result in low coding efficiency. In addition, the conventional prediction method copies the same weighted prediction value to several of the pixels lying in the prediction direction. However, if the original image exhibits gradations, the residual energy might be increased which would also result in low coding efficiency. To handle these problems, we proposed, separately, two improved methods. One allows the use of distant pixels as well as adjoining pixels. The other generates prediction signals with a spatial gradient. This paper confirms the performance of the combination of the two methods. Simulation results show that the combined method outperforms the proposed methods, in isolation, at high bit-rates.

A study on an evolutionary pixel predictor and its properties

Evolutionary methods based on genetic programming (GP) enable dynamic algorithm generation, and have been successfully applied to many areas such as plant control, robot control, and stock market prediction. However, conventional image/video coding methods such as JPEG and H.264 all use fixed (non-dynamic) algorithms without exception. In this article, we introduce a GP-based image predictor that is specifically evolved for each input image. Experimental results demonstrate 2.9 % less entropy (overhead included) than CALICs gradient adjusted predictor.

Intra prediction with spatial gradient

Spatial intra prediction has been added recently to the latest video coding standard H.264/AVC. In the intra prediction of H.264/AVC, there are 9, 9 and 4 prediction modes for 4×4, 8×8 and 16×16 blocks, respectively. Prediction signals are generated by using one or several reference pixels. The value of a reference pixel is copied as the prediction value. In some prediction modes, we calculate a weighted mean by averaging several pixels. The same prediction value is copied to several of the pixels lying in the prediction direction. However, if original image has patterns like gradations, the residual energy could increase which would result in low coding efficiency. In this paper, we propose a new intra prediction that generates prediction signals with a spatial gradient to deal with this problem. Simulation results show that it improves the picture quality and reduce the bit-rate by about 0.14 dB and 1.0 % on average for CIF sequences, respectively. It is also confirmed that our method is effective at high bit-rates.

Lossless scalable video coding with H.264 compliant base layer

This paper proposes a two-layered SNR scalable video coding scheme having a lossless enhancement layer. The base bitstream is fully compliant with the H.264 standard, and the enhancement bitstream is derived from the orthogonal transform coefficients of H.264. In enhance bitstream coding, grid points that cannot be used by the original signal are eliminated to optimize the coding efficiency. We also utilize the numerical relationships between the coefficients to accelerate the process. Computer simulation confirms that the intra coding efficiency of proposed method is comparable to that of motion JPEG2000 reversible mode, while inter coding was achieved with bit reductions of 25 to 50%. A base layer optimization strategy is also described that is shown to yield reasonable overall performance.

Intra angular prediction with weight function and modification filter

Standardization of a new video coding standard, High Efficiency Video Coding (HEVC version 1), has been completed and its compression ratio is twice as much as that of H.264/AVC under the condition of the same subjective quality. In HEVC, many coding tools contribute to the improvement of the coding efficiency. One example is the intra angular prediction mode that has finer prediction directions compared to H.264/AVC. In this paper, we propose two modified intra angular prediction methods: (A) reference pixels that are closer to the predicted pixels than the original reference pixels are exploited, and (B) a blending filter similar to the filter used in the intra DC mode is enhanced and applied. In the proposed method, the predicted signals of the angular prediction mode are generated by blending the predicted signals of the selected angular prediction mode and those of the other mode which has the opposite prediction direction. The overall average coding gain against the HM9.0.1 RExt 1.0 anchor was about 0.5% for the combined method. The maximum coding gain was about 1.0%. The average encoding and decoding run-times were 104.1% and 103.4%, respectively.

Introduction to the Issue on Emerging Technologies for Video Compression

The ten papers in this special issue focus on emerging technologies for video compression. There are contributions related to four key research areas: 1) the merging HEVC standardization process; 2)advances in 3-D coding; 3) quality assessment for next generation codecs; and 4) new compression methods exploiting spatio-temporal content analysis and synthesis.

Adaptive down-sampling of frame-rate for high frame-rate video

Over the past decade, video acquisition rates, which had been 24 Hz (cinema), 30–60 Hz (webcam) or 50–60 Hz (SD/HDcam), has broken through to reach 1000 Hz. In order to display these high frame-rate video signals on current display devices in real time, they must be down-sampled first. This study proposes a down-sampling method suitable for high frame-rate video signals. It is designed with the goal of reducing the inter-frame prediction error. Our method can improve the PSNR of prediction signal by 0.13 [dB] to 0.23 [dB] compared to simple sub-sampling with constant interval.

Theoretical Modeling of Inter-Frame Prediction Error for High Frame-Rate Video Signal

Realistic representations using extremely high quality images are becoming increasingly popular. For example, digital cinemas can now display moving pictures composed of high-resolution digital images. Although these applications focus on increasing the spatial resolution only, higher frame-rates are being considered to achieve more realistic representations. Since increasing the frame-rate increases the total amount of information, efficient coding methods are required. However, its statistical properties are not clarified. This paper establishes for high frame-rate video a mathematical model of the relationship between frame-rate and bit-rate. A coding experiment confirms the validity of the mathematical model.

Constructing a uniform color space for visually lossless color representation and image coding

A uniform color space in terms of just-noticeable differences (JND) to the human eye is important in industry and image processing, as well as for visually lossless image coding and noise evaluations. The method of this paper builds a non-linear color space, by making use of structural analysis with various JND data sets. Through computer simulation, the color space obtained by this method is shown to be superior to conventional ones such as CIELUV in terms of uniformity, and attained a 3-bit reduction in representing chromaticity of RGB color data.