自适应光学
计算机科学
泽尼克多项式
视网膜
光学
变形镜
中央凹
光学相干层析成像
成像体模
可视化
稳健性(进化)
计算机视觉
视野
视网膜
波前
人工智能
医学影像学
限制
临床前影像学
光学成像
图像质量
扫描激光检眼镜
图像处理
物理
分割
人眼
神经纤维层
光学像差
对比度(视觉)
视野
光纤
投影(关系代数)
图像分辨率
图像配准
图像分割
作者
Yao Cai,Olivier Martinache,Maxime Bertrand,Clémentine Callet,Olivier Thouvenin,Kate Grieve,Pedro Mecê
摘要
Adaptive optics (AO) enables cellular-resolution retinal imaging by correcting ocular aberrations, but its widespread clinical adoption remains limited by the narrow field of view (FOV) imposed by the isoplanatic patch of the eye. In this study, we present a deformable mirror (DM)–based sensorless AO time-domain full-field OCT (FFOCT) system that overcomes these limitations by leveraging the inherent robustness of FFOCT to ocular aberrations under spatially incoherent illumination. Using both phantom eye simulations and in vivo experiments, we demonstrate that correction of only three to five Zernike modes (defocus, astigmatism, and coma) is sufficient to significantly enhance SNR and resolve fine retinal structures. This includes reliable visualization of cone photoreceptors as close as 0.3 ∘ from the foveal center and depth-resolved imaging of inner retinal features such as nerve fiber bundles, vessel walls, capillaries, internal limiting membrane, macrophage-like cells, and Gunn’s dots, across a 5 ∘ ×5 ∘ FOV at 500 Hz. By simplifying AO implementation while achieving wide-field cellular resolution, this approach addresses key limitations of current AO ophthalmoscopes and offers a promising pathway toward a wider clinical deployment of high-resolution retinal imaging.
科研通智能强力驱动
Strongly Powered by AbleSci AI