Adaptive frequency compounding for speckle reduction

Speckle pattern in medical ultrasound images is generally regarded as one of crucial factors degrading contrast and signal-to-noise ratio (SNR). To mitigate the pattern, many commercial ultrasound scanners employ the frequency compounding method in which a radio-frequency (RF) spectrum is divided into two or more sub-bands and subsequently envelopes are extracted from the sub-bands. A compounding image is finally produced by adding the envelopes. To maximize the compounding effect, the spectrum downshift along imaging depth caused by the frequency-dependent attenuation should be considered when the bandwidth and the center frequency of each sub-band are determined. For this, the paper proposes an adaptive frequency compounding method in which spectral characteristics of RF signals are analyzed. The estimated spectral characteristics are used for the quadrature demodulation (QDM), determination of sub-band width along depth (AFC-SED). The performances of the proposed method were evaluated in terms of the speckle's signal to noise ratio (SSNR) and the contrast-to-noise ratio (CNR). Compared with Conventional frequency compounding method (CFC), SSNR and CNR were increased by 0.44 and 0.86 in the case of in-vitro experiment and those were 0.14 and 0.55 for in-vivo experiment. If depth-dependent weighting factors are applied to the proposed method (WAFC-SED), SSNR and CNR were more improved in both cases. The experimental results indicate that the proposed method is more suitable to improve SSNR and CNR than CFC.