Inversion elasticity reconstruction of soft tissue using split-and-merge strategy from strain map of ultrasound image sequence

摘要

Progress in ultrasound elasticity imaging (UEI) shows promise as a new way for early detection of cancers by assessing elastic characteristics of soft tissue. However its diagnostic value is reduced by the inherent limitation of spatial and contrast resolution in ultrasound images. We propose a finite-element based nonlinear scheme to estimate the elasticity distribution of soft tissue. A 3D freehand ultrasound imaging system is used to deform tissue with certain contact forces at the same time as imaging. Image displacements are estimated. A split-and-merge strategy is presented for elasticity reconstruction from an image sequence. The L/sub 2/-error between the target and reconstructed Young's modulus is found to be less than 5% for a single inclusion model even with 20% noise level. This result is significantly better than previously reported work. Experimental results on a gelatine phantom are used to illustrate the properties with this method.