Parallelism exploration for 3D high-efficiency video coding depth modeling mode one

摘要

This article presents a parallelism exploration over the depth modeling mode 1 (DMM-1) encoding algorithm of the 3D high-efficiency video coding (3D-HEVC) standard and applied the proposed solutions in a multicore central processing unit (CPU) and two graphics processor unities (GPU). The article evaluates efficient parallel algorithms for DMM-1, which also take advantage of simplifications proposed in our previous works. We demonstrate that DMM-1 can obtain a scalable speedup when running in systems with several available cores even when simplifications are being applied. Experimental results for 1920 x 1088 resolution videos show that the proposed parallel algorithms achieved up to 2 frames per second (fps) in a four-cores (with eight threads) CPU and more than 30 fps in two different GPUs. Therefore, the speedup attained with GPU enables real-time 3D-video encoding applying the proposed parallelism strategies together with the DMM-1 proposed simplifications.