Effects of correlation window length and reflection amplitude on axial displacement measurements using cone phantoms

Correlation based displacement estimation techniques are widely used in the field of ultrasound elastography. Changes in the underlying echo pattern due to mechanical excitation allow the detection of local displacements. In the presence of bright reflections due to an acoustic impedance mismatch, a displacement estimation bias may occur resulting in image blurring. The goal of this study was to measure the effects of correlation window length and boundary reflections on axial displacement measurements using a cone phantom. A 2D MATLAB simulation was developed to model this system. Displacements were estimated using windowed normalized cross-correlation. Windows ranged in length from 0.2 mm to 2.9 mm (1.3 ¿ to 20 ¿ given f = 5.33 MHz and c = 1540 m/s). A line of scatterers were located on the edge of the wedge and assigned a higher amplitude in order to simulate a bright reflection. The location on the wedge axis where the displacement was half the maximum displacement at the base of the wedge was identified. The axial wedge width at that location was reported as the resolution. Axial resolution measurements were linearly proportional to window length. Resolution was approximately 0.8 mm given a 0.6 mm window length with a reflection that was 5 times the amplitude of the surrounding speckle, and 0.4 mm without a reflection for our set of parameters. A preliminary experiment was also performed in which a tissue-mimicking cone phantom was imaged using an Acoustic Radiation Force Impulse (ARFI) Imaging system to measure the effects of the reflective boundary as a function of correlation window length.