Abstract:Discrete cosine transform (DCT) has been widely used in image/video coding systems, where zigzag scan is usually employed for DCT coefficient organization. However, due to local diversity of prediction errors, the fixed scan pattern, such as the traditional zigzag scan, is not efficient all the time for organizing the DCT transformed prediction errors. In this paper, the statistical distribution of prediction residuals is studied first, and two other scan patterns are introduced to complete the set of possible patterns. Then, a new adaptive scheme that adaptively chooses for each macroblock the best scan pattern to arrange the quantized DCT coefficients into arrays is presented. The best scan pattern is selected for each macroblock according to its own residual characteristics and hence leads to improved coding efficiency. Experimental results show that this macroblock-level adaptive scan scheme (MLASS) can always outperform the traditional zigzag one and the peak signal to noise ratio (PSNR) gain can be up to 0.65dB. Moreover, the extra two scan patterns do not increase memory requirement much. Additionally, a fast scan pattern decision algorithm is given for complexity reduction. This fast version of MLASS can keep almost the same coding efficiency while with much lower complexity.

Discrete Cosine Transform Coefficients Based Scan Format Conversion

FAST ALGORITHM FOR DCT DOMAIN IMAGE UP-CONVERSION

A novel high performance discrete cosine transform algorithm based on the stochastic computation

Macroblock-Level Adaptive Scan Scheme for Discrete Cosine Transform Coefficients

A fast scheme for converting DCT coefficients to H.264/AVC integer transform coefficients

A novel fast DCT coefficient scan architecture

Low-complexity Multidimensional DCT Approximations

New Cost-Effective VLSI Implementation of 2D Discrete Cosine Transform and Its Inverse

A Block-Dct Based Edge Detection Algorithm

An Algorithm for Computing DCT Using Improved Arithmetic Fourier Transform

A Novel Multiple Description Coding Frame Based on Reordered DCT Coefficients and SPIHT Algorithm

A Novel Embedded Coding Algorithm Based on the Reconstructed DCT Coefficients.

Extensions on low-complexity DCT approximations for larger blocklengths based on minimal angle similarity

Dynamic Bit-Width Adaptation in DCT: An Approach to Trade Off Image Quality and Computation Energy

Directional Discrete Cosine Transforms: A Theoretical Analysis

Low-complexity 8-point DCT Approximation Based on Angle Similarity for Image and Video Coding

Error-resistance and Low-complexity Integer Inverse Discrete Cosine Transform

On Arithmetic Cosine Transform Algorithm

Diagonal Discrete Cosine Transforms for Image Coding

An Energy-efficient Approximate DCT Design for Image Processing (Invited)

A Novel DS-Based Hybrid DCT for Low Bit-Rate Video Coding