Automatic detection of the foveal center in optical coherence tomography
| Autor: | Thomas Theelen, Vivian Schreur, Sascha Fauser, Bram van Ginneken, Carel B. Hoyng, Bart Liefers, Clara I. Sánchez, Freerk G. Venhuizen |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
genetic structures
Image quality Computer science Center (group theory) Convolutional neural network Article Sensory disorders Donders Center for Medical Neuroscience [Radboudumc 12] 030218 nuclear medicine & medical imaging Convolution 03 medical and health sciences 0302 clinical medicine Optical coherence tomography Foveal medicine Computer vision Pixel medicine.diagnostic_test business.industry eye diseases Atomic and Molecular Physics and Optics Data set 030221 ophthalmology & optometry sense organs Artificial intelligence business Rare cancers Radboud Institute for Health Sciences [Radboudumc 9] Biotechnology |
| Zdroj: | Biomedical Optics Express, 8, 11, pp. 5160-5178 Biomedical Optics Express, 8, 5160-5178 |
| ISSN: | 2156-7085 |
| Popis: | Contains fulltext : 181865.pdf (Publisher’s version ) (Open Access) We propose a method for automatic detection of the foveal center in optical coherence tomography (OCT). The method is based on a pixel-wise classification of all pixels in an OCT volume using a fully convolutional neural network (CNN) with dilated convolution filters. The CNN-architecture contains anisotropic dilated filters and a shortcut connection and has been trained using a dynamic training procedure where the network identifies its own relevant training samples. The performance of the proposed method is evaluated on a data set of 400 OCT scans of patients affected by age-related macular degeneration (AMD) at different severity levels. For 391 scans (97.75%) the method identified the foveal center with a distance to a human reference less than 750 mum, with a mean (+/- SD) distance of 71 mum +/- 107 mum. Two independent observers also annotated the foveal center, with a mean distance to the reference of 57 mum +/- 84 mum and 56 mum +/- 80 mum, respectively. Furthermore, we evaluate variations to the proposed network architecture and training procedure, providing insight in the characteristics that led to the demonstrated performance of the proposed method. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|