No abstract

We introduce a novel CNN-based feature point detector -Greedily Learned Accurate Match Points (GLAMpoints)learned in a semi-supervised manner. Our detector extracts repeatable, stable interest points with a dense coverage, specifically designed to maximize the correct matching in a specific domain, which is in contrast to conventional techniques that optimize indirect metrics. In this paper, we apply our method on challenging retinal slitlamp images, for which classical detectors yield unsatisfactory results due to low image quality and insufficient amount of low-level features. We show that GLAMpoints significantly outperforms classical detectors as well as state-of-the-art CNN-based methods in matching and registration quality for retinal images.

The automatic segmentation of retinal layer structures enables clinically-relevant quantification and monitoring of eye disorders over time in OCT imaging. Eyes with late-stage diseases are particularly challenging to segment, as their shape is highly warped due to pathological biomarkers. In this context, we propose a novel fully-Convolutional Neural Network (CNN) architecture which combines dilated residual blocks in an asymmetric U-shape configuration, and can segment multiple layers of highly pathological eyes in one shot. We validate our approach on a dataset of late-stage AMD patients and demonstrate lower computational costs and higher performance compared to other state-of-the-art methods.

in this work we evaluated a postprocessing, customized automatic retinal oct B-scan enhancement software for noise reduction, contrast enhancement and improved depth quality applicable to Heidelberg engineering Spectralis oct devices. A trained deep neural network was used to process images from an oct dataset with ground truth biomarker gradings. performance was assessed by the evaluation of two expert graders who evaluated image quality for B-scan with a clear preference for enhanced over original images. objective measures such as SnR and noise estimation showed a significant improvement in quality. Presence grading of seven biomarkers IRF, SRF, ERM, Drusen, RPD, GA and iRoRA resulted in similar intergrader agreement. intergrader agreement was also compared with improvement in iRf and RpD, and disagreement in high variance biomarkers such as GA and iRoRA. OCT is a non-invasive, micrometer-resolution imaging technique that has found wide application in the diagnosis of corneal and retinal pathologies. Thanks to advances in electronics, precision optics and signal processing, OCT technology has steadily improved in image quality, speed and resolution. However, speckle noise and signal loss in deeper tissue remains a major limitation. Speckle noise is caused by a complex combination of thermal, electrical, multiple-scattering effects, as well as digital processing algorithms. Indeed, in retinal imaging, it is common to consider up to 75% of the pixel values as noise 1,2. A common approach to improving OCT image quality is to acquire and average multiple scans of the same location. Assuming that noise is uncorrelated between the acquired images, the average of N images will improve the signal-to-noise by a factor of N while correlated noise will reduce the improvement in practice. Consequently, the approach requires a longer acquisition time, by a factor of N, during which the patient is required to fixate motionless on a fixation target. While this approach helps to improve images of patients with clear media, it results in rather unsatisfactory results in patients with media opacities e.g. cataracts. To mitigate this, commercial OCT devices often include a separate optical eye tracking system to support the process, with corresponding increases in cost and device complexity. Imperfections in patient fixation and the eye tracking system lead to blurriness in the averaged scans. Combined, the above create a practical ceiling to the image quality improvement that can be extracted from image averaging. See Fig. 1 for denoising and averaging examples. Traditionally, digital noise removal attempts to post-process acquired images to reduce the amount of speckle noise without harming the structural information presence in the images sample. We identify two main areas in which OCT denoising has been evaluated, the first one considers spatial denoising methods, where image enhancement happens either via local image filtering such as median 3 or mean Gaussian filters 4 , or at global OCT volume scale. The latter includes B...

Purpose To develop a reliable algorithm for the automated identification, localization, and volume measurement of exudative manifestations in neovascular age-related macular degeneration (nAMD), including intraretinal (IRF), subretinal fluid (SRF), and pigment epithelium detachment (PED), using a deep-learning approach. Methods One hundred seven spectral domain optical coherence tomography (OCT) cube volumes were extracted from nAMD eyes. Manual annotation of IRF, SRF, and PED was performed. Ninety-two OCT volumes served as training and validation set, and 15 OCT volumes from different patients as test set. The performance of our fluid segmentation method was quantified by means of pixel-wise metrics and volume correlations and compared to other methods. Repeatability was tested on 42 other eyes with five OCT volume scans acquired on the same day. Results The fully automated algorithm achieved good performance for the detection of IRF, SRF, and PED. The area under the curve for detection, sensitivity, and specificity was 0.97, 0.95, and 0.99, respectively. The correlation coefficients for the fluid volumes were 0.99, 0.99, and 0.91, respectively. The Dice score was 0.73, 0.67, and 0.82, respectively. For the largest volume quartiles the Dice scores were >0.90. Including retinal layer segmentation contributed positively to the performance. The repeatability of volume prediction showed a standard deviations of 4.0 nL, 3.5 nL, and 20.0 nL for IRF, SRF, and PED, respectively. Conclusions The deep-learning algorithm can simultaneously acquire a high level of performance for the identification and volume measurements of IRF, SRF, and PED in nAMD, providing accurate and repeatable predictions. Including layer segmentation during training and squeeze-excite block in the network architecture were shown to boost the performance. Translational Relevance Potential applications include measurements of specific fluid compartments with high reproducibility, assistance in treatment decisions, and the diagnostic or scientific evaluation of relevant subgroups.