Superimposed Boundary Strength Matrix Method for Decision Process of In-Loop Deblocking Filter
Jae Gon Kim, Jin Woo Hwang, Jun‐Dong Cho · Hokkaido University Collection of Scholarly and Academic Papers (Hokkaido University) · 2009
This paper presents the so-called "superimposed boundary strength matrix method" for the decision process for an in-loop deblocking filter. In this paper, we reduce the number of branch operations so as to cope with computational complexity, and such as pipeline stalls. Our method is compared with e.g. JM reference software, and verified that our method meets the specification of JM reference such as boundary strength values and PSNR. Simulation results show that the proposed method can achieve performance enhancement, ranging from 64.4% to 80.3%, within reasonable processing time. In addition to this, highly parallelized processing components can be obtained, and if the proposed method is applied to the parallel processing hardware implementation, performance enhancement will be apparent by exploiting parallelism included in our superimposed BS matrix method.