Segmentation of vessel structures from photoacoustic images with reliability assessment

Pasi Raumonen; Tanja Tarvainen

doi:10.1364/BOE.9.002887

Segmentation of vessel structures from photoacoustic images with reliability assessment

Biomed Opt Express. 2018 Jun 4;9(7):2887-2904. doi: 10.1364/BOE.9.002887. eCollection 2018 Jul 1.

Authors

Pasi Raumonen¹, Tanja Tarvainen^{2

3}

Affiliations

¹ Laboratory of Mathematics, Tampere University of Technology, PO Box 527, 33101 Tampere, Finland.
² Department of Applied Physics, University of Eastern Finland, P.O. Box 1627, 70211 Kuopio, Finland.
³ Department of Computer Science, University College London, Gower Street, London WC1E 6BT, UK.

Abstract

Photoacoustic imaging enables the imaging of soft biological tissue with combined optical contrast and ultrasound resolution. One of the targets of interest is tissue vasculature. However, the photoacoustic images may not directly provide the information on, for example, vasculature structure. Therefore, the images are improved by reducing noise and artefacts, and processed for better visualisation of the target of interest. In this work, we present a new segmentation method of photoacoustic images that also straightforwardly produces assessments of its reliability. The segmentation depends on parameters which have a natural tendency to increase the reliability as the parameter values monotonically change. The reliability is assessed by counting classifications of image voxels with different parameter values. The resulting segmentation with reliability offers new ways and tools to analyse photoacoustic images and new possibilities for utilising them as anatomical priors in further computations. Our MATLAB implementation of the method is available as an open-source software package [P. Raumonen, Matlab, 2018].

Keywords: (100.3008) Image recognition, algorithms and filters; (100.3190) Inverse problems; (100.6950) Tomographic image processing; (110.5120) Photoacoustic imaging.