An efficient low area implementation of 2-D DCT on FPGA

Abstract

This paper presents the design and implementation for 2-D discrete cosine transform (DCT) with the goal of achieving low area utilization and high-speed operation on FPGAs. The design is based on the row-column decomposition technique, which requires two successive 1-D DCT transforms and a transpose memory between them for storing and transposing the results of the first 1-D DCT. The proposed implementation of 2-D DCT is capable of compressing at least 70 images per second in 720×480 resolution on Xilinx Spartan 3E and 30 images per second in 1920×1080 resolution on Xilinx Virtex 7 FPGA. Consequently, the proposed 2-D DCT design and implementation can be very useful in various image and video compressing applications