Maximum likelihood estimation of scattering strength for high range resolution ultrasound imaging

Abstract

The range spatial resolution in ultrasound imaging is an important index determining image quality and basically determined by the frequency response of the ultrasonic transducer. In the present study, a new element-domain signal processing technique based on the maximum likelihood method was developed for improvement of the range spatial resolution in ultrasonic imaging. In the present study, the point spread function (PSF) of the ultrasonic system employed was suppressed to estimate the scattering strength of a target. An echo from a 15-micron stainless wire in water was used as the range PSF. An echo signal from a target, which is received by a transducer element, can be modeled as the sum of noise and convolution of the scattering strength of a target and PSF. The likelihood of an element echo signal can be obtained by assuming that noise is Gaussian. The scattering strength of a target was estimated so that the likelihood was maximized. The PSF of the system was obtained from a 15-micron fine wire in water, and the proposed method was applied to the ultrasonic data obtained by ultrafast plane wave imaging without compounding (1302 fps). The range spatial resolution, which was determined by the width at half maximum of an echo from the wire, was improved significantly from 0.212 mm to 0.081 mm by the proposed method. Also, the proposed method was applied to in vivo imaging of a human carotid artery. In the B-mode image obtained by the proposed method, echoes from the lumen-intima interfaces became sharper than those in the conventional B-mode image. Also, the speckle pattern was suppressed because the interference among echoes was reduced by shortening the ultrasonic pulse length. As a result, the lateral continuity of the image was improved. Furthermore, undesirable echoes in the lumen, which were considered to be generated from sidelobes, were also suppressed.