earticle

영어

Multi-core processors have become the dominant market trend because they provide a great opportunity in increasing processing performance by exploiting various parallelisms. In JPEG (Joint Photographic Experts Group) compression, color space conversion is one of the major kernels known as a computationally expensive module. This paper presents a fast solution for color space conversion with multi-core parallel computation. For this purpose, we utilize Threading Building Blocks (TBB), a runtime library based on C++, and OpenMP (Open Multi-processing), a shared programming language. A RGB image is transformed into a luminance-chrominance color space such as YCbCr. The implementation results show that parallel implementations achieve greater performance improvement regarding processing speed compared with the serial implementation.