A Bayesian Framework for the Local Configuration of Retinal Junctions

Retinal images contain forests of mutually intersecting and overlapping venous and arterial vascular trees. The geometry of these trees shows adaptation to vascular diseases including diabetes, stroke and hypertension. Segmentation of the retinal vascular network is complicated by inconsistent vessel contrast, fuzzy edges, variable image quality, media opacities, complex intersections and overlaps. This paper presents a Bayesian approach to resolving the configuration of vascular junctions to correctly construct the vascular trees. A probabilistic model of vascular joints (terminals, bridges and bifurcations) and their configuration in junctions is built, and Maximum A Posteriori (MAP) estimation used to select most likely configurations. The model is built using a reference set of 3010 joints extracted from the DRIVE public domain vascular segmentation dataset, and evaluated on 3435 joints from the DRIVE test set, demonstrating an accuracy of 95.2%.

[1]  Julien Jomier,et al.  Automatic Vascular Tree Formation Using the Mahalanobis Distance , 2005, MICCAI.

[2]  B. Wasan,et al.  Vascular network changes in the retina with age and hypertension , 1995, Journal of hypertension.

[3]  Touseef Ahmad Qureshi,et al.  A manually-labeled, artery/vein classified benchmark for the DRIVE dataset , 2013, Proceedings of the 26th IEEE International Symposium on Computer-Based Medical Systems.

[4]  Laurent D. Cohen,et al.  Grouping Connected Components using Minimal Path Techniques , 2002 .

[5]  P. Gregson,et al.  Automated grading of venous beading. , 1995, Computers and biomedical research, an international journal.

[6]  Bashir Al-Diri,et al.  Automated analysis of retinal vascular network connectivity , 2010, Comput. Medical Imaging Graph..

[7]  Guido Gerig,et al.  3D Graph Description of the Intracerebral Vasculature from Segmented MRA and Tests of Accuracy by Comparison with X-ray Angiograms , 1999, IPMI.

[8]  Touseef Ahmad Qureshi,et al.  Automatic localization of the optic disc in retinal fundus images using multiple features , 2012, 2012 IEEE 12th International Conference on Bioinformatics & Bioengineering (BIBE).

[9]  Stephen A. Davis,et al.  Retina Verification System Based on Biometric Graph Matching , 2013, IEEE Transactions on Image Processing.

[10]  P. Choyke,et al.  Gray-scale skeletonization of small vessels in magnetic resonance angiography , 2000, IEEE Transactions on Medical Imaging.

[11]  Bernhard Preim,et al.  Analysis of vasculature for liver surgical planning , 2002, IEEE Transactions on Medical Imaging.

[12]  Andrew Hunter,et al.  An Active Contour Model for Segmenting and Measuring Retinal Vessels , 2009, IEEE Transactions on Medical Imaging.

[13]  George Azzopardi,et al.  Detection of Retinal Vascular Bifurcations by Trainable V4-Like Filters , 2011, CAIP.

[14]  Chia-Ling Tsai,et al.  Model-based method for improving the accuracy and repeatability of estimating vascular bifurcations and crossovers from retinal fundus images , 2004, IEEE Transactions on Information Technology in Biomedicine.

[15]  Max A. Viergever,et al.  Ridge-based vessel segmentation in color images of the retina , 2004, IEEE Transactions on Medical Imaging.