.png){kind=icon}
.png){kind=icon}
Return
Low-Complexity 2D FIR Filter Design Using Spectral Approximation and Parallel Polyphase Processing
.png){kind=icon}
PRE
AI
DOI：10.1007/s00034-026-03503-8.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Efficient design of two-dimensional (2D) filters is increasingly vital in fields such as real-time image processing and biomedical signal analysis. In this work, a new 2D filter design methodology based on 2D polyphase decom- position, which significantly reduces the implementation complexity in terms of multipliers and adders along with reducing the root mean square error is proposed. Here, the 2D filter kernel is partitioned into 2D polyphase components, which are then approximated using least square approximation in two directions and the obtained coefficients are represented in terms of 2D Farrow structure. Then, appropriately utilizing the fractional delays and interpolation technique, the 2D filter kernel is reproduced. The examples for the design of 2D circular and elliptical filters are provided to demonstrate the hardware efficiency of the proposed approach. It is shown that the proposed methodology results in 2D filters with around 20.39% reduction in the complexity of implementation, when compared to the latest existing one dimensional (1D) polyphase structure based 2D filter design methodology.
Keywords:
Polyphase decomposition
Spectral approximation
Farrow structure
Parallel processing
Journal
C
IF:
2
Papers:
318
Citations:
0