Strain compounding: spatial resolution and performance on human images.

Abstract

Compounding has long been used to reduce speckle brightness variations by incoherently combining partially correlated measurements of the same image object. Conventionally, decorrelation between measurements is introduced by imaging from different spatial positions (i.e., spatial compounding) or within different frequency ranges (i.e., frequency compounding). An alternative compounding approach based on speckle decorrelation of an object under different strain conditions was recently proposed. The new approach was referred to as strain compounding. Speckle correlation of strain compounding was previously analyzed using simulations and radiofrequency (RF) data from a gelatin-based phantom. Results indicated that effective speckle reduction is possible at strains potentially achievable in certain clinical situations. In this paper, performance of strain compounding is further analyzed. Potential degradation in spatial resolution of strain compounding is compared to that of conventional approaches using computer simulations. Results show that strain compounding potentially can achieve the same level of speckle reduction, with less degradation in spatial resolution if a large compression can be applied. Performance of strain compounding is also tested using postdetection human images acquired by a commercial imaging system. Two-dimensional (2-D) speckle tracking is performed to correct for tissue motion resulting from external compression and to ensure that the images to be compounded are spatially matched. Thyroid images show that speckle brightness variations can be effectively reduced without significant degradation in spatial resolution. Speckle reduction in liver imaging is also demonstrated. However, speckle reduction is not as significant as in thyroid imaging, due to the fact that a large compression cannot be achieved.