Show simple item record

dc.contributor.advisorLu, Doctor Weiping
dc.contributor.advisorDuncan, Professor Rory R.
dc.contributor.authorWilson, Rhodri S.
dc.date.accessioned2019-01-25T16:01:39Z
dc.date.available2019-01-25T16:01:39Z
dc.date.issued2017-05
dc.identifier.urihttp://hdl.handle.net/10399/3421
dc.description.abstractParticle tracking is an essential tool for the study of dynamics of biological processes. The dynamics of these processes happens in three-dimensional (3D) space as the biological structures themselves are 3D. The focus of this thesis is on the development of single particle tracking methods for analysis of the dynamics of biological processes through the use of image processing techniques. Firstly, introduced is a novel particle tracking method that works with two-dimensional (2D) image data. This method uses the theory of Haar-like features for particle detection and trajectory linking is achieved using a combination of three Kalman filters within an interacting multiple models framework. The trajectory linking process utilises an extended state space variable which better describe the morphology and intensity profiles of the particles under investigation at their current position. This tracking method is validated using both 2D synthetically generated images as well as 2D experimentally collected images. It is shown that this method outperforms 14 other stateof-the-art methods. Next this method is used to analyse the dynamics of fluorescently labelled particles using a live-cell fluorescence microscopy technique, specifically a variant of the super-resolution (SR) method PALM, spt-PALM. From this application, conclusions about the organisation of the proteins under investigation at the cell membrane are drawn. Introduced next is a second particle tracking method which is highly efficient and capable of working with both 2D and 3D image data. This method uses a novel Haar-inspired feature for particle detection, drawing inspiration from the type of particles to be detected which are typically circular in 2D space and spherical in 3D image space. Trajectory linking in this method utilises a global nearest neighbour methodology incorporating both motion models to describe the motion of the particles under investigation and a further extended state space variable describing many more aspects of the particles to be linked. This method is validated using a variety of both 2D and 3D synthetic image data. The methods performance is compared with 14 other state-of-the-art methods showing it to be one of the best overall performing methods. Finally, analysis tools to study a SR image restoration method developed by our research group, referred to as Translation Microscopy (TRAM) are investigated [1]. TRAM can be implemented on any standardised microscope and deliver an improvement in resolution of up to 7-fold. However, the results from TRAM and other SR imaging methods require specialised tools to validate and analyse them. Tools have been developed to validate that TRAM performs correctly using a specially designed ground truth. Furthermore, through analysis of results on a biological sample corroborate other published results based on the size of biological structures, showing again that TRAM performs as expected.en_US
dc.description.sponsorshipEPSCRen_US
dc.language.isoenen_US
dc.publisherHeriot-Watt Universityen_US
dc.publisherEngineering and Physical Sciencesen_US
dc.rightsAll items in ROS are protected by the Creative Commons copyright license (http://creativecommons.org/licenses/by-nc-nd/2.5/scotland/), with some rights reserved.
dc.titleQuantitative analysis of microscopyen_US
dc.typeThesisen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record