Rapid, easy and reliable calibration for freehand 3D ultrasound

Abstract

This paper presents improvements to the plane-based technique for calibrating freehand 3D ultrasound systems. The improvements are designed to make it easier for inexperienced users to perform plane-based calibration and to know that they have got a reliable result. In particular, we enable the calibration to be performed using water at room temperature while producing a result that is valid for average soft tissue, and we show how it is possible to provide feedback on the reliability of the calibration using a metric based on the curvature of the calibration criterion function. We present comprehensive results showing that these innovations improve the precision of the calibration and offer useful feedback to the user.

Freehand 3D Ultrasound Calibration: A Review

“…In some rare cases, the scales may be supplied by the manufacturer (Boctor et al, 2003) or by accessing the raw ultrasound signals (Hsu et al, 2006), but this requires special arrangements with the supplier. Otherwise, the scales can be obtained explicitly by using the distance measurement tool provided by the ultrasound machines (Hsu et al, 2006).…”

“…Otherwise, the scales can be obtained explicitly by using the distance measurement tool provided by the ultrasound machines (Hsu et al, 2006).…”

“…From our experience of supporting medical physicists and clinicians, an incorrect calibration is often obtained because the phantom has not been scanned from a sufficiently diverse set of positions and orientations. Although an eigenvalue metric exists to detect whether the system of equations is under-constrained (Hsu et al, 2006), the user is still required to be sufficiently trained. Dandekar et al (2005) used two parallel wires to mimic a plane phantom.…”

“…This means that points near the centre of the image are still roughly correct, but errors measured by points towards the edges are more visible. Therefore, many papers in recent years quote calibration reproducibility for a point at a corner of the image (Blackall et al, 2000;Rousseau et al, 2005), points along the left and right edges of the image (Leotta, 2004) and the four corners of the image Gee et al, 2005;Hsu et al, 2006). Leotta (2004) and the Cambridge group Gee et al, 2005;Hsu et al, 2006) also quoted the spread in the centre of the image.…”

“…Before the stylus was developed enabling the evaluation of point reconstruction accuracy, many groups assessed accuracy by measuring the distances between objects (Leotta et al, 1997;Prager et al, 1998;Blackall et al, 2000;Boctor et al, 2003;Lindseth et al, 2003b;Leotta, 2004;Dandekar et al, 2005;Hsu et al, 2006;Krupa, 2006). This technique is popular because the experiment is easy to set up.…”