Block data flow implementation of the 2-D discrete wavelet transform

Implements the 2D discrete wavelet transform (DWT) for real time image processing applications. The algorithms for the 2D wavelet transform can also be viewed as subband filter banks. The authors propose a block data flow architecture (BDFA) for the 2D product separable filter which is used for the DWT. The BDFA uses block data processing and the block data flow paradigm to minimize system management overhead due to the communication protocol. The authors propose a 2D array structured BDFA which is an extended architecture of the BDFA for the 2D filter. The approach uses both data partitioning and algorithm partitioning to achieve high efficiency and high throughput. The authors compare their approach to existing approaches such as the systolic implementation of the wavelet transform. The proposed method is suitable for real time image processing in the sense that it achieves high throughput and high efficiency using a small number of processors.<
>