Update to Uncertainty Quantification and Multiparameter reconstruction in breast imaging

The main objective of WP2 is the scientific validation of the technology and the development of an imaging software prototype based on its principles. Hereafter, our experts explain how the first two task have evolved over the past few months.

Uncertainty quantification in images

Our uncertainty estimator is designed to measure different sources of error in Ultrasound Computed Tomography (US-CT) images. Examples of errors include variable density, attenuation, and transducer position; all of which cause bright/dark spots in the US-CT sound speed image. These spots are not a true representation of the sound speed, and the uncertainty estimates should help us to locate the erroneous spots.

We have shown that the uncertainty estimate is capable of this task. By introducing different types of error, we notice that imaging errors are correctly quantified by uncertainty. For example, when the circle has a high density, a hyper-intensive region forms on the boundary between tissues. This region is correctly identified as erroneous by the corresponding uncertainty estimate, which has a ring of high uncertainty where the sound speed is too high. We find this is also true in the case of attenuation and transducer position errors.

Multiparameter Reconstruction of Breast Images

The propagation of ultrasound waves in biological tissue is subject to complex phenomena. In other words, ultrasound beams become highly distorted due to the heterogeneities and irregularities of the tissue. These variations are related to physical properties of the tissues which will provide doctors with further information about the breast and possible lesions detected.

Over the past months, we have validated the multiparameter algorithm through several canonical experiments that helped us to see the precision of the algorithm based on each parameter: speed of sound, density and attenuation factor; and completed the study on the performance of the Fourier Pseudo-Spectral algorithm with very positive results.

Another important task over the past few months has been to provide a novel approach for the source modelling in order to characterize the non-isotropic behaviour of a real Tranducer Array System (TAS). This solution is not aiming only at reproducing the specific radiation pattern generated by an isolated TAS, but also at impressing its position and orientation for each of them in the 3D USCT III system.

Authors: Oscar Bates (ICL), Carlos Spa (BSC), Cristina Durán (FW) and Natalia Gutierrez (FW).