An Efficient VLSI Architecture and FPGA Implementation of High-Speed and Low Power 2-D DWT for (9, 7) Wavelet Filter

摘要

This paper presents an efficient VLSI architecture of a high speed, low power 2-D Discrete Wavelet Transform computing. The proposed architecture, based on new and fast lifting scheme approach for (9, 7) filter in DWT, reduces the hardware complexity and memory accesses. Moreover, it has the ability of performing progressive computations by minimizing the buffering between the decomposition levels. The system is fully compatible with JPEG2000 standard. Our designs were realized in VHDL language and optimized in terms of throughput and memory requirements. The implementations are completely parameterized with respect to the size of the input image and the number of decomposition levels. The proposed architecture is verified by simulation and successfully implemented in a Cyclone II and Stratix III FPGAs, and the estimated frequency of operation is 350 MHz. The resulting computing rate is up to 48 frames (4096x2160) per second with 24 bpp. The architecture has regular structure, simple control flow, small embedded buffers and low power consumption. Thus, it is very suitable for new generation image compression systems, such as JPEG2000.