Frequency Compounding of the Ultrasound Images Using Wavelet Packets Decomposition

In a standard ultrasound imaging, the radio-frequency (RF) backscattered signals are decimated to get a medical image. The RF signal carries information about the scattering structure. Analysis of the RF signals may get useful clinical information for the tissue characterization. In this paper, the approach that is wavelet packets (WP) decomposition of I/Q ultrasound signals was proposed to gain the medical image. Using this method, all the signals are decomposed to generate a set of sub-band I/Q signals, so the ultrasound echo signals are simultaneously frequency decomposed and decimated by two quadrature filters. Besides, we apply the wavelet-packet (WP) de-noising method to de-noise I/Q signals for enhancing the feature of tissue. Finally the sub-band images can be compounded to reduce the speckle. Application of this method proved a significant improvement to CNR. In the experiment, because of the wavelet packets decomposition with frequency compounding, the speckle will be reduced largely but the detail resolution of the image can be lower than the image with the conventional imaging method. To estimate the loss of the detail resolution, we can use the 3db width to compute the detail resolution in the end and we found the loss of detail resolution with the WP method is within the acceptable range.