Block diagram of retinal imaging system.
CCD imager attached to fundus camera. A—CCD camera, B—viewing eyepiece, C—support structure for camera.
Images from subject B—from left to right and top to bottom, the wavelengths are 504, 525, 557, 582, 600, and 620 nm. The images shown are an average of three exposures and have been scaled for display purposes to the same minimum and maximum intensity and gamma corrected with a coefficient of 0.7. The reduction in visible features as the wavelength increases can be clearly seen. This is caused by the lower absorption coefficient of both hemoglobins and melanin at the longer wavelengths.
Image acquisition parameters used for retinal imaging: central wavelength of the filtered LED, the source intensity, and the exposure duration.
Linear displacement between two image frames taken at different wavelengths of the static artificial eye. Only the results of the sequential registration are shown as the results of registration to the image at 525 nm were very similar. Values are shown in pixels (pixel size is approximately ).
Linear displacement showing the magnitude of translation required to align two images taken at five different wavelengths in the same sequence. is a translation vector returned by the registration algorithm (Ref. 17) (see Sec. III B). Values are shown in pixels (pixel size is approximately ).
SNR (decibel) measurements from subject B. A region was used from each of the respective areas. The SNR was computed using .
Article metrics loading...
Full text loading...