Purchase this article with an account.
Ernesto Blanco, Xiaolin Wang, Yuhua Zhang; Evaluate and Optimize the Performance of Adaptive Optics for Retinal Imaging. Invest. Ophthalmol. Vis. Sci. 2011;52(14):4059.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To develop accurate wavefront control strategy thereby approaching the "best achievable" wavefront compensation for an adaptive optics scanning laser ophthalmoscope (AOSLO).
We designed an effective iteration algorithm for adaptive optics (AO) wavefront correction in which the control matrix was derived from the normalized transpose of the influence matrix. We applied this strategy to an AOSLO that consists of two electromagnetic deformable mirrors (DM) (miraoTM 52-e from ImagineEyes, Hi-Speed dm97 from Alpao, France) and a custom Shack-Hartman wavefront sensor. We used the retinal image quality (brightness, contrast and resolution) that was obtained from a model-eye as the criterion to assess the performance of the new algorithm. We compared the images that were taken with the new technique, with a model-based Monte-Carlo control algorithm, and with the traditional inverse influence matrix based control that has been widely adopted by the retinal imaging community.
The new algorithm is unconditionally stable and convergent; it can drive the residual aberration to the minimum; thus, it was able to achieve the maximal brightness. For the two DMs, the r.m.s. residual aberration could be reduced to less than 0.03 micron over -6 ~ 5 D spherical aberration. Compared to the model-based Monte-Carlo control, the new technique achieved close image quality; whereas compared to the tradition inverse influence function based control, it always achieved better image brightness and contrast, i.e, better AO correction. It took about 150 loops for the controller to reduce the aberration to less than 0.1 micron. The converging speed is slower than that of the inverse influence function matrix based control; but it is much more efficient than that of the Monte-Carlo method, thus, allowing us to obtain a quick estimation of the "best-achievable" wavefront correction. With appropriate design of the controller, this method was able to work for imaging of the living human eye.
We demonstrated a method that can be used to evaluate the AO performance in high resolution retinal imaging. We were able to optimize the AO control algorithm and improve the accuracy of the wavefront compensation.
This PDF is available to Subscribers Only