 |
Dr. Bashir Al-Diri (PhD., MSc., BSc., PGCE, FHEA) University of Lincoln, School of Computer Science Email: baldiri@lincoln.ac.uk |
|
Biography Dr. Al-Diri’s PhD research is concerned with the development of computer vision algorithms to analyse retinal blood vessels for the characterization of diabetic retinopathy. His work is unique in integrating automated segmentation of vessels with automated and accurate measurement, which is the foundation of this proposal. Dr. Al-Diri has developed a robust fully-automated system for retinal vascular segmentation and measurement, which provides a unique combination of good segmentation and superior measurement performance. It is thus uniquely well-suited to act as a base for research into the diagnosis of vascular diseases that cause measurable changes to the geometry of retinal vessels. The system is robust, and can accurately locate vessel edges under difficult conditions, including noisy blurred edges, closely parallel vessels, light reflections from vessels, and very fine vessels, and yields precise vessel width measurements, with sub-pixel average width errors. Dr. Al-Diri was the first author to undertake a thorough analysis of the performance of retinal vessel measurement algorithms, developing with the clinical partner, Sunderland Eye Infirmary, the REVIEW dataset (Retinal Vessel Image set for Estimation of Widths; http://ReviewDB.lincoln.ac.uk); which is now publicly available and is rapidly becoming the standard benchmark for retinal vascular measurement algorithms. He developed an edge marking algorithm used to process these segments to produce vessel profiles which provides more precise measurements. He has developed an algorithm that forms a retinal vessel graph by analysing the potential connectivity of segmented retinal vessels using implicit cost functions to resolve the configuration of local sets of segment ends to determining the network connectivity. Dr. Al-Diri has developed a computerized tool for the manual measurement of retinal bifurcation features, designed for use in investigating correlations between measurement features and clinical conditions and generating reference dataset measurements. This shows better agreement with theoretical predictions than other alternative available semi-manual tools for vascular width estimation. Dr. Al-Diri has subsequently developed an automated procedure for measuring retinal bifurcation geometries, allowing the rapid measurement of large numbers of retinal bifurcations. Its measurements are not subject to unconscious “confirmation bias” of an operator due to differing observer perceptions of the edge location. Dr. Al-Diri research includes computer vision, medical image analysis, automated surveillance, artificial intelligence, speech recognition, language corpus and lexical analysis, and applications of active contour models. |
|
|
Retinal Segments |
 |