Retinal atrophy in eyes with resolved papilledema detected by optical coherence tomography.

BACKGROUND: To apply automated spectral domain optical coherence tomography (SD-OCT) segmentation to eyes with resolving papilledema. METHODS: Ninety-four patients with idiopathic intracranial hypertension seen at the Duke Eye Center neuro-ophthalmology clinic between November 2010 and October 2011 were reviewed. Excluded were eyes with papilledema with Frisén grade >2, other optic neuropathies or retinopathies, and those that did not have SD-OCT imaging. The remaining 43 patients were split into 2 groups: non-atrophic papilledema and atrophic papilledema. Automated SD-OCT segmentation was performed on patients with non-atrophic papilledema and age-matched controls for each of the 9 regions of the Early Treatment Diabetic Retinopathy Study map. Bonferroni correction was used for multiple comparisons. All SD-OCT scans were reviewed for retinal structural abnormalities. RESULTS: Total macular thickness was significantly thinner within the fovea and inner macular ring in non-atrophic papilledema vs control eyes (266 vs 276 µm, P = 0.04; 333 vs 344 µm P < 0.01, n = 26 non-atrophic papilledema, 30 controls). SD-OCT demonstrated thinning within the fovea, inner macular ring, and outer macular ring of the outer plexiform layer plus nuclear layer in non-atrophic papilledema vs control (124 vs 131 µm, P < 0.01; 112 vs 118 µm, P = 0.03; 95 vs 100 µm, P = 0.03). Retinal structural changes were seen in 21/33 eyes with atrophic papilledema vs none of the eyes with non-atrophic papilledema or controls. CONCLUSIONS: SD-OCT shows qualitative and quantitative changes in the macula of eyes with resolved papilledema.

Baking Neural Radiance Fields for Real-Time View Synthesis.

Neural volumetric representations such as Neural Radiance Fields (NeRF) have emerged as a compelling technique for learning to represent 3D scenes from images with the goal of rendering photorealistic images of the scene from unobserved viewpoints. However, NeRF's computational requirements are prohibitive for real-time applications: rendering views from a trained NeRF requires querying a multilayer perceptron (MLP) hundreds of times per ray. We present a method to train a NeRF, then precompute and store (i.e. "bake") it as a novel representation called a Sparse Neural Radiance Grid (SNeRG) that enables real-time rendering on commodity hardware. To achieve this, we introduce 1) a reformulation of NeRF's architecture, and 2) a sparse voxel grid representation with learned feature vectors. The resulting scene representation retains NeRF's ability to render fine geometric details and view-dependent appearance, is compact (averaging less than 90 MB per scene), and can be rendered in real-time (higher than 30 frames per second on a laptop GPU). Actual screen captures are shown in our video.

Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields.

Neural Radiance Fields (NeRF) is a popular view synthesis technique that represents a scene as a continuous volumetric function, parameterized by multilayer perceptrons that provide the volume density and view-dependent emitted radiance at each location. While NeRF-based techniques excel at representing fine geometric structures with smoothly varying view-dependent appearance, they often fail to accurately capture and reproduce the appearance of glossy surfaces. We address this limitation by introducing Ref-NeRF, which replaces NeRF's parameterization of view-dependent outgoing radiance with a representation of reflected radiance and structures this function using a collection of spatially-varying scene properties. We show that together with a regularizer on normal vectors, our model significantly improves the realism and accuracy of specular reflections. Furthermore, we show that our model's internal representation of outgoing radiance is interpretable and useful for scene editing.

Shape Estimation from Shading, Defocus, and Correspondence Using Light-Field Angular Coherence.

Light-field cameras are quickly becoming commodity items, with consumer and industrial applications. They capture many nearby views simultaneously using a single image with a micro-lens array, thereby providing a wealth of cues for depth recovery: defocus, correspondence, and shading. In particular, apart from conventional image shading, one can refocus images after acquisition, and shift one's viewpoint within the sub-apertures of the main lens, effectively obtaining multiple views. We present a principled algorithm for dense depth estimation that combines defocus and correspondence metrics. We then extend our analysis to the additional cue of shading, using it to refine fine details in the shape. By exploiting an all-in-focus image, in which pixels are expected to exhibit angular coherence, we define an optimization framework that integrates photo consistency, depth consistency, and shading consistency. We show that combining all three sources of information: defocus, correspondence, and shading, outperforms state-of-the-art light-field depth estimation algorithms in multiple scenarios.

Automatic segmentation of up to ten layer boundaries in SD-OCT images of the mouse retina with and without missing layers due to pathology.

Accurate quantification of retinal layer thicknesses in mice as seen on optical coherence tomography (OCT) is crucial for the study of numerous ocular and neurological diseases. However, manual segmentation is time-consuming and subjective. Previous attempts to automate this process were limited to high-quality scans from mice with no missing layers or visible pathology. This paper presents an automatic approach for segmenting retinal layers in spectral domain OCT images using sparsity based denoising, support vector machines, graph theory, and dynamic programming (S-GTDP). Results show that this method accurately segments all present retinal layer boundaries, which can range from seven to ten, in wild-type and rhodopsin knockout mice as compared to manual segmentation and has a more accurate performance as compared to the commercial automated Diver segmentation software.

Fully automated detection of diabetic macular edema and dry age-related macular degeneration from optical coherence tomography images.

We present a novel fully automated algorithm for the detection of retinal diseases via optical coherence tomography (OCT) imaging. Our algorithm utilizes multiscale histograms of oriented gradient descriptors as feature vectors of a support vector machine based classifier. The spectral domain OCT data sets used for cross-validation consisted of volumetric scans acquired from 45 subjects: 15 normal subjects, 15 patients with dry age-related macular degeneration (AMD), and 15 patients with diabetic macular edema (DME). Our classifier correctly identified 100% of cases with AMD, 100% cases with DME, and 86.67% cases of normal subjects. This algorithm is a potentially impactful tool for the remote diagnosis of ophthalmic diseases.

Fully automatic software for retinal thickness in eyes with diabetic macular edema from images acquired by cirrus and spectralis systems.

PURPOSE: To determine whether a novel automatic segmentation program, the Duke Optical Coherence Tomography Retinal Analysis Program (DOCTRAP), can be applied to spectral-domain optical coherence tomography (SD-OCT) images obtained from different commercially available SD-OCT in eyes with diabetic macular edema (DME). METHODS: A novel segmentation framework was used to segment the retina, inner retinal pigment epithelium, and Bruch's membrane on images from eyes with DME acquired by one of two SD-OCT systems, Spectralis or Cirrus high definition (HD)-OCT. Thickness data obtained by the DOCTRAP software were compared with those produced by Spectralis and Cirrus. Measurement agreement and its dependence were assessed using intraclass correlation (ICC). RESULTS: A total of 40 SD-OCT scans from 20 subjects for each machine were included in the analysis. Spectralis: the mean thickness in the 1-mm central area determined by DOCTRAP and Spectralis was 463.8 ± 107.5 µm and 467.0 ± 108.1 µm, respectively (ICC, 0.999). There was also a high level agreement in surrounding areas (out to 3 mm). Cirrus: the mean thickness in the 1-mm central area was 440.8 ± 183.4 µm and 442.7 ± 182.4 µm by DOCTRAP and Cirrus, respectively (ICC, 0.999). The thickness agreement in surrounding areas (out to 3 mm) was more variable due to Cirrus segmentation errors in one subject (ICC, 0.734-0.999). After manual correction of the errors, there was a high level of thickness agreement in surrounding areas (ICC, 0.997-1.000). CONCLUSIONS: The DOCTRAP may be useful to compare retinal thicknesses in eyes with DME across OCT platforms.