Computational model for multifocal imaging in optical projection tomography and numerical analysis of all-in-focus fusion in tomographic image reconstruction

Abstract

<p>Optical projection tomography (OPT) is a noninvasive 3D imaging method that has been used to study small biological samples. In OPT samples can be mounted in hydrogel scaffold mimicking real life extracellular matrix, and hence grown in all natural dimensions. In optical imaging systems, focusing lenses are required for image acquisition. Due to these lenses, particles at a certain distance from objectives — in the focal plane of the lens — are captured accurately and the further a particle is from the focal plane the blurrier it is captured in the resulting image. To compensate this limitation, multifocal OPT is implemented, where images from each angle are taken with multiple focal planes at different distances. From these images, parts in focus are detected and combined into a single image using all-in-focus fusion algorithm. In this work we present computational way of modeling multifocal imaging and use the presented model to assess the performance of two different all-in-focus fusion methods.</p>