ABSTRACT
Nowadays, most areas in medicine make use of specialized computer systems for disease diagnosis assistance, providing clinicians with a tool for a more efficient processing of information, making it more reliable and reproducible. Ophthalmology is no exception as many computerized tools have been developed to diagnose, control, and follow retinal pathologies such as glaucoma, age-related macular degeneration, or neovascularization. These tools rely on image acquisition techniques that provide images with adequate quality and information. Among these techniques, it is important to highlight optical coherence tomography (OCT), a relatively recent modality that enables the acquisition of retinal images to a level not achieved before. OCT allows the acquisition of images with
CONTENTS
12.1 Introduction ........................................................................................................................ 175 12.2 Retinal Layer Detection in OCT Images ......................................................................... 177
12.2.1 Preprocessing ......................................................................................................... 177 12.2.2 Layer Detection ...................................................................................................... 178
12.2.2.1 3-D Graph Construction ......................................................................... 179 12.2.2.2 Minimum Closed Set Search ................................................................. 181 12.2.2.3 Reconstruction and Layer Storage ........................................................ 182
12.2.3 Representation ........................................................................................................ 182 12.3 Experiments and Results .................................................................................................. 183
12.3.1 Data Set .................................................................................................................... 183 12.3.2 Results in the 2-D Model ....................................................................................... 184 12.3.3 Results in the 3-D Model ...................................................................................... 185
12.4 Computational Optimizations to the 3-D Model .......................................................... 188 12.4.1 First Optimization ................................................................................................. 188 12.4.2 Second Optimization............................................................................................. 188 12.4.3 Performance Analysis ........................................................................................... 190
12.5 Conclusions ......................................................................................................................... 190 References ..................................................................................................................................... 191
differentiation of the layers of the retina. Layer distribution defines a number of surfaces in the retina, in particular, six surfaces of interest for several pathologies: Surface 1 corresponds to the inner limiting membrane (ILM). Surface 2 separates the nerve fiber layer (NFL) from the ganglion cell layer (GCL). Surface 3 corresponds to the separation between inner plexiform layer (IPL) and the inner nuclear layer (INL). Surface 4 separates the outer plexiform layer (OPL) from the outer nuclear layer (ONL). Surface 5 is the union of the inner segments (IS) and outer segments (OS). Surface 6 separates the OS from the retinal pigment epithelium (RPE).
