Jae Ho Kim

Fast full-search motion estimation based on multilevel successive elimination algorithm

An improved fast full-search algorithm for motion estimation is proposed. The computational process of the sum of absolute difference and the multilevel successive elimination algorithm (MSEA) are theoretically analysed. Based on these analyzes, a new detection criterion and an estimation of the starting level are proposed in order to reduce the computations required for judging invalid candidate blocks. The simulation results show that the proposed algorithm achieves a computational timesaving of about 40% compared with the MSEA.

New edge-enhanced error diffusion algorithm based on the error sum criterion

new edge-enhanced error diffusion algorithm, based on Eschbachs algorithm, is proposed. Thick-edged artifacts as well as small edge-enhancement effects for the bright or dark pixel values are observed in the previous algorithm. By analyzing the phenomena, a new improved algorithm is proposed by using the diffused error sum and input pixel value. An input pixel is classified into a normal- or edge-region pixel based on the error sum criterion. A new error calculation is then employed for the edge region pixel, while conventional error calculation is used for the normal-region pixel. The proposed method requires only a few additional calculations and provides edge-enhanced binary output images. The edges are in fluenced less by the brightness offset, and thick-edged artifacts are reduced.

A novel fast fractal decoding algorithm

A novel fast fractal decoding algorithm is proposed. In the proposed algorithm, continuation of the decoding process is judged according to an image convergence criterion (ICC), instead of a predefined number of iterations. A selective decoding excludes the converged range blocks by measuring a block convergence criterion (BCC) of the range block. Initial image estimation is used and one-buffer decoding is applied to accelerate the convergence speed. Without loss of quality, the reconstructed image is obtained by computation corresponding to nearly three iterations of conventional decoding.

Block wavelet transform for sub-image coding/decoding

Wavelet transform is used efficiently for high compression ratio image coding. It is useful for a high resolution and color document image processing system. However, a large amount of memory is required in wavelet decompression for a large image. Conventional wavelet transform does not allow reconstructing a sub-region of the image. Therefore, it is hard to use wavelet transform in a color document image processing system. In this paper, a block wavelet transform method is proposed. An image is divided into blocks and the samples of the adjacent block are used to transform one block for removing the edge artifact. The required number of samples of the adjacent block are derived. By using the proposed method, the memory requirement for wavelet coding/decoding is reduced to that of the block size. Any targeted region of an image can be compressed or reconstructed. The proposed method can be used for a color document image processing system.

Fast fractal decoding algorithm with convergence criteria

In fractal decoding procedures, the reconstructed image is obtained by means of a predefined number of iterations with an arbitrary initial image. A novel fast decoding algorithm with convergence criteria is proposed. It is composed of a selective decoding of range blocks with a block convergence criterion (BCC), estimation of an initial image, and one-buffer decoding. In the proposed algorithm, continuation of the decoding process is judged by an image convergence criterion (ICC), not by the predefined number of iterations. Without redundant iterations and loss of quality, the reconstructed image is obtained by the amount of calculations corresponding to only three to four iterations of the general decoding procedure.

Image coding using wavelet transform and human visual system

In this paper, we propose a new image compression technique using wavelet transform and human visually estimated noise sensitivities. These consist of frequency, background brightness, and edge height sensitivities. The background brightness sensitivity for each quantizing point is modeled by a quadratic function. The edge sensitivity for each quantizing point is modeled by a non-linear function. The minimum value becomes background brightness sensitivity and edge height sensitivity is multiplied by the frequency sensitivity for determining the quantization step size. Quantization step sizes are calculated by using coefficients of lowest frequency band which are coded losslessly. Therefore, in the proposed method, information to specify quantization step size for higher frequency band, is not needed. The coefficients of high frequency bands are arithmetically coded in horizontal and vertical directions depending on the edge direction. Compared with previous human visual systems based image compression methods, the proposed method shows improved image quality for the same compression ratio with less computational cost.

A hybrid motion-vector coding scheme based on an estimation of the locality for motion-vector difference

The bit portion for motion-vector (MV) information becomes significant in very-low bitrate video coding as texture coding has been improved. Especially, the existing methods, such as minimum bitrate prediction (MBP)-based coding, do not work well for slow-moving video sequences, which are popular in video conferencing and videophone applications. It is because the characteristics of the video sequences are not fully exploited for the MV encoding. In this paper, we propose a new hybrid encoding scheme for MV. A new parameter called a skip rate (SR) is defined for estimating the degree of the locality for the MV difference in the proposed scheme. In addition, we suggest a joint coding with a new single codeword combined with the horizontal and vertical codewords. Either the joint coding or the MBP-based coding is determined by the SR. The simulation results show that the proposed scheme enhances the coding performance for various video sequences without additional memory.

An improved multilevel successive elimination algorithm for fast full-search motion estimation

It is well known that a lot of computations are required for motion estimation in video encoding. Recently, the multilevel successive elimination algorithm (MSEA) has been proposed to reduce the computational complexity of block matching for motion estimation without degradation of estimation accuracy. In the MSEA, the block matching is hierarchically carried out and the invalid candidate block is successively eliminated. In this paper, we propose an improved MSEA based on a new efficient decision condition. The number of the calculations to eliminate the invalid block can be reduced by the new efficient decision condition in the proposed algorithm. The simulation results show that the proposed algorithm achieves about 30-40% computational saving as compared with the MSEA.

New edge-enhanced error diffusion based on the error sum criteria

In this paper, a new edge enhanced error diffusion algorithm which is based on the Eschbachs algorithm is proposed. Thick edge artifacts with large edge enhancing factor as well as less edge enhancement effects for the bright or dark pixel values are observed in the previous algorithm. By analyzing the phenomena, a new improved algorithm is proposed by using the diffused error sum and input pixel value. An input pixel is classified into a normal- or edge-region pixel based on the error sum criteria. Then, a new error calculation is employed for the edge-region pixel, while conventional error calculation is used for the normal- region pixel. The proposed method requires only a few additional calculations, and provides edge enhanced binary output images. The edge is less influenced by the brightness offset and thick edge artifacts are reduces.

Fast multiple run_before decoding method for efficient implementation of an H.264/advanced video coding context-adaptive variable length coding decoder

Abstract. We propose a fast new multiple run_before decoding method in context-adaptive variable length coding (CAVLC). The transform coefficients are coded using CAVLC, in which the run_before symbols are generated for a 4×4 block input. To speed up the CAVLC decoding, the run_before symbols need to be decoded in parallel. We implemented a new CAVLC table for simultaneous decoding of up to three run_befores. The simulation results show a Total Speed-up Factor of 205%∼144% over various resolutions and quantization steps.