E. Oruklu and J. Saniie (USA)
Ultrasound, target detection, wavelets, DCT, WHT,multiplierless.
In this paper we present subband decoding processors for frequency-diverse ultrasonic imaging and target detection in high scattering noise (i.e., clutter echoes). Our design is based on the development of multiplierless architectures that successfully deal with the high computational load of real-time applications while minimizing area and power consumption. Target detection algorithms are based on the premise that clutter echoes exhibit randomness and are more sensitive to frequency shifts than target echoes. Therefore, subband decomposition methods such as discrete cosine transform (DCT), discrete wavelet transform (DWT) and Walsh Hadamard transform (WHT) can be used for differentiating the target information from the clutter echoes. In this study, DCT, DWT and WHT are examined with multiplierless architectures for subband decomposition of the ultrasonic data. We have analyzed multiplierless hardware architectures for each of these transforms. In particular, we synthesized DWT with a reduced adder graph (RAGn) algorithm. WHT is inherently multiplication free and requires only addition and subtraction operations. The DCT with binDCT method also offers multiplierless subband decoding.
Important Links:
Go Back
