Self-calibrating eye tracker using image saliency. / Självkalibrerande ögonspårare med hjälp av image saliency.

No / To examine foveal structure in amblyopia using spectral-domain optical coherence tomography (SD-OCT). DESIGN: Prospective, cross-sectional study. PARTICIPANTS AND CONTROLS: Two subject groups were recruited to the study: 85 amblyopes (34 adults, 51 children) and 110 visually normal controls (44 adults, 66 children). METHODS: A detailed eye examination, including an SD-OCT scan, was performed in all participants. A total of 390 eyes of 195 subjects were imaged using a 3-dimensional (3D) macula scan covering a nominal 6 x 6-mm area with a resolution of 256 x 256 (65,536 axial scans). Data from the B-scans bisecting the fovea both horizontally and vertically were fitted with a mathematical model of the fovea to determine a range of foveal parameters. MAIN OUTCOME MEASURES: Foveal thickness, foveal pit depth, and foveal pit slope. RESULTS: Bilateral differences between the eyes of amblyopes compared with visually normal controls were found. The difference between foveal structure in amblyopic participants relative to structure in subjects with normal vision persisted even when variables such as age, ethnicity, axial length, and sex were taken into account. Amblyopes showed increased foveal thickness (+8.31 mum; P = 0.006) and a reduction in pit depth in the horizontal meridian (-10.06 mum; P = 0.005) but not in the vertical meridian (P = 0.082) when compared with subjects with normal vision. Foveal pit slopes were found to be approximately 1 degree flatter in the nasal (P = 0.033) and temporal (P = 0.014) meridians in amblyopes, but differences between amblyopes and controls in the superior (P = 0.061) and inferior (P = 0.087) meridians did not reach statistical significance. No statistically significant interocular differences were found in the foveal structure between amblyopic and fellow eyes. CONCLUSIONS: Differences were found in the foveal structure in both eyes of amblyopes compared with subjects with normal vision. These differences consisted of increased foveal thickness, reduced pit depth when measured along the horizontal meridian, and flattening of the nasal and temporal sides of the foveal pit.