ABSTRACT
Most retinopathies, for example, from hypertension or diabetes, could be diagnosed early and treated if an accurate and objective analysis of symptoms at their initial onset could be performed. The analysis should be accurate enough to detect minor pathological signs and objective enough to be able to compare results with accepted clinical standards and with results obtained from the same patient at different times. This latter requirement is
CONTENTS
5.1 Introduction .......................................................................................................................... 97 5.1.1 Related Work ............................................................................................................ 98 5.1.2 Motivation .................................................................................................................99 5.1.3 Our Contribution ................................................................................................... 100
5.2 Methods ............................................................................................................................... 100 5.2.1 Seed Finding ........................................................................................................... 100 5.2.2 Multidirectional Graph Search Tracking ............................................................ 102
5.2.2.1 Background .............................................................................................. 102 5.2.2.2 Initialization ............................................................................................. 103 5.2.2.3 Search ........................................................................................................ 104 5.2.2.4 Merging .................................................................................................... 104 5.2.2.5 Cost Function ........................................................................................... 104 5.2.2.6 Circular Path Fixing ................................................................................ 105
5.2.3 Vessel Caliber Extraction ...................................................................................... 106 5.2.3.1 Vessel Prole Setup ................................................................................. 106 5.2.3.2 Border Point Extraction .......................................................................... 106 5.2.3.3 Adaptive Correction ............................................................................... 106
5.3 Material ............................................................................................................................... 108 5.4 Results ................................................................................................................................. 109
5.4.1 Vessel Segmentation .............................................................................................. 109 5.4.2 Caliber Estimation ................................................................................................. 111
5.5 Discussion ........................................................................................................................... 112 5.6 Conclusion .......................................................................................................................... 114 Acknowledgment ........................................................................................................................ 114 References ..................................................................................................................................... 114
of paramount importance when assessing the effect of established therapeutic treatments and even more when evaluating in a quantitative way the efcacy of new drugs during their development.
