Fig. 4: Cellular-level indocyanine green (ICG) retinal pigment epithelial (RPE) cell images from diseased eyes generated using stratified cycleGAN. | Communications Medicine

Fig. 4: Cellular-level indocyanine green (ICG) retinal pigment epithelial (RPE) cell images from diseased eyes generated using stratified cycleGAN.

From: Artificial intelligence assisted clinical fluorescence imaging achieves in vivo cellular resolution comparable to adaptive optics ophthalmoscopy

Fig. 4

Late phase ICG images from four different diseases (before and after artificial intelligence (AI) enhancement) are shown. Images shown were acquired from retinal locations ~0.5 mm away from the fovea for age-related macular degeneration (AMD), retinitis pigmentosa, and choroideremia carrier, and 3 mm for vitelliform macular dystrophy in an area with no apparent outer retinal lesions. Conventional ICG and ground truth adaptive optics (AO)-ICG images were acquired from identical locations. Stratified cycleGAN trained on healthy data only was applied to the conventional ICG image to generate the AI assisted (AI-ICG) images. Examples of individual RPE cells are shown by magenta arrows. Scale bar: 100 µm.

Back to article page