Level-Set Method for Image Analysis of Schlemm's Canal and Trabecular Meshwork

Open-source deep learning-based automatic segmentation of mouse Schlemm's canal in optical coherence tomography images

The purpose of this study was to develop an automatic deep learning-based approach and corresponding free, open-source software to perform segmentation of the Schlemm's canal (SC) lumen in optical coherence tomography (OCT) scans of living mouse eyes. A novel convolutional neural network (CNN) for semantic segmentation grounded in a U-Net architecture was developed by incorporating a late fusion scheme, multi-scale input image pyramid, dilated residual convolution blocks, and attention-gating. 163 pairs of intensity and speckle variance (SV) OCT B-scans acquired from 32 living mouse eyes were used for training, validation, and testing of this CNN model for segmentation of the SC lumen. The proposed model achieved a mean Dice Similarity Coefficient (DSC) of 0.694 ± 0.256 and median DSC of 0.791, while manual segmentation performed by a second expert grader achieved a mean and median DSC of 0.713 ± 0.209 and 0.763, respectively. This work presents the first automatic method for segmentation of the SC lumen in OCT images of living mouse eyes. The performance of the proposed model is comparable to the performance of a second human grader. Open-source automatic software for segmentation of the SC lumen is expected to accelerate experiments for studying treatment efficacy of new drugs affecting intraocular pressure and related diseases such as glaucoma, which present as changes in the SC area.

Full circumferential morphological analysis of Schlemm's canal in human eyes using megahertz swept source OCT

We performed full circumferential imaging of the Schlemm's canal (SC) of two human eyes using a Fourier domain mode-lock laser (FDML) based 1.66-MHz SS-OCT prototype at 1060 nm. Eight volumes with overlapping margins were acquired around the limbal area with customized raster scanning patterns designed to fully cover the SC while minimizing motion artifacts. The SC was segmented from the volumes using a semi-automated active contour segmentation algorithm, whose mean dice similarity coefficient was 0.76 compared to the manual segmentation results. We also reconstructed three-dimensional (3D) renderings of the 360° SC by stitching the segmented SCs from the volumetric datasets. Quantitative metrics of the full circumferential SC were provided, including the mean and standard deviation (SD) of the cross-sectional area (CSA), the maximum CSA, the minimum and maximum SC opening width, and the number of collector channels (CC) stemming from the SC.