Abstract
Purpose: :
The ability to measure thickness of retinal layers has potential value for early detection of pathologies and disease monitoring. A new image processing segmentation algorithm was developed for automated detection of retinal layer boundaries and measurement of thickness of 6 retinal layers in optical coherence tomography (OCT) images.
Methods: :
OCT images were acquired with time and spectral domain OCT instruments in 15 visually normal healthy subjects. A dedicated software program was developed in Matlab for image processing and analysis. The image processing algorithm segmented the OCT image by the following steps: 1) alignment of A-scans; 2) gray-level mapping; 3) directional filtering; 4) edge detection; and 5) model-based decision making. Thickness profiles for 6 retinal layers were generated in normal subjects. Automated boundary detection and quantitative thickness measurements estimated using the algorithm were compared with measurements obtained from boundaries manually marked by 3 observers.
Results: :
On OCT images, 7 retinal layer boundaries were automatically identified by the algorithm. The root mean squared error (RMSE) between the manual and automatic boundary detection ranged between 3 and 10 microns. Thickness profiles were generated for 6 retinal layers: nerve fiber layer (NFL), inner plexiform layer and ganglion cell layer (IPL+GCL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer and photoreceptor inner segments (ONL+PIS), and photoreceptor outer segments (POS). The mean absolute values of differences between automated and manual thickness measurements were comparable to inter-observer differences, ranging between 3 and 4 microns. Thickness profiles of retinal layers corresponded with normal anatomy. Inner retinal thickness profiles demonstrated minimum thickness at the fovea. The OPL and ONL+PIS thickness profiles displayed a minimum and maximum thickness at the fovea, respectively. The POS thickness profile was relatively constant along the OCT scans through the fovea.
Conclusions: :
The application of this image processing technique is promising for investigating thickness changes of retinal layers due to disease progression and therapeutic intervention.
Keywords: image processing • imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)