Related Experiment Video
Updated: Jul 13, 2026

An Experimental Protocol for Assessing the Performance of New Ultrasound Probes Based on CMUT Technology in Application to Brain Imaging
Published on: September 24, 2017
The challenges in creating reference maps for verification of ultrasound images
J E Wilhjelm1, S K Jespersen, E Falk
1CADUS, Ørsted DTU, Building 348, DK-2800 Kgs. Lyngby, Denmark. jw@oersted.dtu.dk <jw@oersted.dtu.dk>
Abstract:
In the development of new medical imaging techniques, references to which the images can be compared are necessary if one wants to assess how precise the images are. This is especially interesting in diagnostic ultrasound where a number of artefacts influence the image. The reference can either be derived from a phantom with precisely known properties and geometry, from the specifications of a computer phantom (simulated images) or from evaluation of biological tissue. The third approach can be conducted with other medical imaging modalities (CT, MRI, etc.) or "destructive testing" involving histology. In this paper, aspects of the latter method is considered in detail. Formalin fixed tissue is moulded into an agar block containing at set of fiducial markers. The block is scanned with ultrasound. Both tissue and fiducial markers are imaged. The block is afterwards sliced at the location of the fiducial markers. The slices are then photographed and analyzed histologically. From this data, reference maps with similar geometry as the ultrasound images can be created. Ideally, for each pixel in the ultrasound image, these reference maps indicate tissue type, such as collagen poor tissue, collagen rich tissue, etc. Many of the sources of error as well as the challenges with such a method are discussed.

