Learning layer-specific edges for segmenting retinal layers with large deformations

Abstract

Abstract:We present an algorithm for layer-specific edge detection in retinal optical coherence tomography images through a structured learning algorithm to reinforce traditional graph-based retinal layer segmentation. The proposed algorithm simultaneously identifies individual layers and their corresponding edges, resulting in the computation of layer-specific edges in 1 second. These edges augment classical dynamic programming based segmentation under layer deformation, shadow artifacts noise, and without heuristics or prior knowledge. We considered Duke's online data set containing 110 B-scans of 10 diabetic macular edema subjects with 8 retinal layers annotated by two experts for experimentation, and achieved a mean distance error of 1.38 pixels whereas that of the state-of-the-art was 1.68 pixels.

Transfer learning based classification of optical coherence tomography images with diabetic macular edema and dry age-related macular degeneration

“…3(b). The RPE is estimated as proposed in [42], the estimated RPE is smoothened, retinal flattening is performed, and the RPE lower contour is shifted to a fixed position (70% of the height) as shown in Fig. 3(c).…”

Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search

“…Mathematic model based methods construct a fixed or adaptive model based on prior assumptions for the structure of the input images, and include A-scan [16,17], active contour [18][19][20][21], sparse high order potentials [22], and 2D/3D graph [23][24][25][26][27][28][29][30] based methods. Machine learning based methods formulate layer segmentation as a classification problem, where features are extracted from each layer or its boundaries and used to train a classifier (e.g.…”

Detection and Identification of Animals in Wild Life Sanctuaries using Convolutional Neural Network

“…The concept of deep learning [13] enables automated extraction of additional abstraction layers on computer systems. Based on the inspiration obtained from the visual cortex of mammals [14], every layer of the CNN extracts data from the parts of images obtained from the outcome of previous layer [12]. A soft-max function comprises the CNN last layer specially designed for the purpose of classification.…”

“…(30). Hand held characteristics are harnessed on existing techniques for animal classification includes (8), where attempts are made to differentiate webcam recording that contain no animals (25) as well as the usage of support vector machine (14) for classification of images. They achieve an accuracy of 82%.…”