A High Performance Architecture for 3D Wavelet Transform

Discrete Wavelet Transform (DWT) is attracted great deal of attention and used in various image, video and signal processing applications. Even though lifting scheme is proposed to reduce the computational load of DWT, traditional hardware solutions based on lifting scheme for DWT suffer from high critical path and computational resources. Flipping structure has been proposed to resolve these problems in the lifting based hardware implementations for DWT. In this paper we propose a pipeline architecture based on flipping structure for 3D DWT. The proposed architecture has lower critical path and computational resources compared to the other proposed hardwares for 3D DWT in the literature. Moreover, the proposed architecture can be used to implement 1D and 2D DWT besides the 3D DWT.