Development of an image processing tool for tracking fluorescent labelled probes in a micro fibrillar cellulose gel

摘要

This paper presents a tool for automatic tracking of the trajectories of fluorescent labelled latex particles (probes) in a micro fibrillar cellulose (MFC) gel. The image processing tool is based on the well-known tracking method, Lucas-Kanade [1]. Lucas-Kanade is template-based, which means that the studied particle is compared with a search area. The Lucas-Kanade optical flow computation solves the least square problem, i.e. finds a correlation between the template and the search area [2]. To make the tracking method more robust and to have a good local accuracy the pyramidal implementation of the Lucas-Kanade optical flow computation is used. Affine warp is used to handle changes in shape between two subsequent images.