Computer-Aided Analysis of Fluorescein Angiograms Using Color Leakage Maps / Eman Magdy Ahmed

Supervisor: Ahmed Fahmy

Thesis (M.A.)—Nile University, Egypt, 2013 .

"Includes bibliographical references"

Contents:
Chapter 1: INTRODUCTION ..................................................................................................... 1
1.1 Objectives ..................................................................................................................... 1
1.2 Motivation ..................................................................................................................... 2
1.3 Research Flow ............................................................................................................... 2
1.4 Datasets ......................................................................................................................... 3
1.5 Organization of thesis ................................................................................................... 4
Chapter 2: BACKGROUND ........................................................................................................ 5
2.1 Anatomy of Human Eye ............................................................................................... 5
2.1.1 Retina ................................................................................................................ 7
2.2 Retinal imaging ........................................................................................................... 10
2.2.1 Fundus Fluorescein Angiography ...................................................................... 10
2.2.1.1 Normal Fluorescein Angiograms ........................................................ 11
2.2.1.2 Abnormal Fluorescein Angiograms .................................................... 13
2.2.2 Optical Coherence Tomography ........................................................................ 16
2.2.2.1 Normal Optical Coherence Tomograms ............................................. 17
2.2.2.2 Abnormal Optical Coherence Tomograms ......................................... 18
2.3 Retinal Diseases .......................................................................................................... 19
2.3.1 Diabetic Macular Edema.................................................................................... 19
2.3.2 Age-related Macular Degeneration .................................................................... 21
2.4 Summary ..................................................................................................................... 25
Chapter 3: IMAGE PRE-PROCESSING ................................................................................. 27
3.1 Image Registration ......................................................................................................
3.1.1 Literature Review............................................................................................ 27
3.1.2 Methodology ................................................................................................... 29
3.1.3 Implementation ............................................................................................... 31
3.1.4 Experiments .................................................................................................... 32
3.2 Vessel Removal .......................................................................................................... 33
3.2.1 Literature Review............................................................................................ 33
3.2.2 Methodology ................................................................................................... 33
3.2.3 Implementation ............................................................................................... 35
3.2.4 Experiments .................................................................................................... 37
3.3 Macula Cropping ........................................................................................................ 38
3.3.1 Methodology ................................................................................................... 38
3.3.2 Implementation ............................................................................................... 38
3.3.3 Experiments .................................................................................................... 39
3.4 Intensity Normalization .............................................................................................. 40
3.4.1 Methodology ................................................................................................... 40
3.4.2 Implementation ............................................................................................... 40
3.4.3 Experiments .................................................................................................... 41
3.5 Summary ..................................................................................................................... 42
Chapter 4: LESION ENHANCEMENT ................................................................................... 43
4.1 Modeling the Early Macula ........................................................................................ 43
4.1.1 Literature Review............................................................................................ 44
4.1.2 Methodology ................................................................................................... 45
4.1.3 Implementation ............................................................................................... 46
4.1.4 Experiments .................................................................................................... 48
4.2 Subtraction .................................................................................................................. 49
4.2.1 Methodology ................................................................................................... 49
4.2.2 Implementation ............................................................................................... 49
4.2.3 Experiments .................................................................................................... 50
4.3 Summary ..................................................................................................................... 52
Chapter 5: COLOR MAPPING ................................................................................................ 53
5.1 Simulating OCT line scans (Re-sampling) .................................................................
5.1.1 Methodology ................................................................................................... 54
5.1.2 Implementation ............................................................................................... 54
5.1.3 Experiments .................................................................................................... 54
5.2 Simulating OCT image reconstruction ....................................................................... 55
5.2.1 Methodology ................................................................................................... 55
5.2.2 Implementation ............................................................................................... 56
5.2.3 Experiments .................................................................................................... 58
5.3 Mapping ...................................................................................................................... 59
5.3.1 Methodology ................................................................................................... 59
5.3.2 Implementation ............................................................................................... 60
5.3.3 Experiments .................................................................................................... 62
5.4 Summary ..................................................................................................................... 65
Chapter 6: RESULTS AND EVALUATION ........................................................................... 67
6.1 Results ......................................................................................................................... 67
6.2 Evaluation ................................................................................................................... 69
6.2.1 Correlation between FLM and OCT ............................................................... 69
6.2.2 Reproducibility of FLM .................................................................................. 72
Chapter 7: CONCLUSION AND FUTURE WORK ............................................................... 75
7.1 Conclusion .................................................................................................................. 75
7.2 Future Work ................................................................................................................

Abstract:
Fundus Fluorescein Angiography (FFA) is a standard screening and diagnostic technique for several retinal diseases. FFA is powerful and essential for detecting and treatment-guiding of vascular-related retinal diseases such as diabetic retinopathy, macular edema, macular degeneration, retinal detachment, vascular occlusion, and tumors. Nevertheless, current analysis methods of FFA images depend on subjective evaluation of the amount of fluorescein leakage. This evaluation is qualitatively performed by skilled observers. Accordingly, these methods are vulnerable to inter-/intra-observer variability and lack accuracy and reproducibility as well.
In this thesis, we present a new method for computer-aided analysis of FFA images. The method is based on generating objective color fluorescein leakage maps (FLM) that mimic the thickness maps produced by Optical Coherence Tomography (OCT). The generated FLM evaluates the fluorescein leakage and gives an indication about its severity. A complete system is introduced in this work to automatically generate FLM. It starts with pre-processing steps including: image registration, retinal blood vessels removal, cropping the macula and normalizing intensity between images. Then, modeling the early macula timeframes before subtracting it from late timeframes has been followed. Finally, simulating OCT image reconstruction has been presented to generate color FLMs. The proposed method was applied to 64 datasets of two different retinal diseases and the FLMs results were compared to the OCT thickness maps. The results show strong correlation between the FLMs and OCT thickness maps (r=0.8). The proposed method was found to be reproducible and robust to variability in the image acquisition times.

Text in English, abstracts in English.