Abstract
Purpose :
Many commercial fundus cameras provide infrared video preview of the retina to aid patient alignment. For systems that illuminate a wide field-of-view, relatively high-power infrared sources are required, particularly if indocyanine-green angiography must be supported as a secondary function. Laser diodes can provide high radiant power with very low cost and a small form-factor, but the coherent light produces speckle patterns that significantly degrade image contrast. Pulse-shaping of injection current (Figure 1A) can suppress laser speckle, improving imaging contrast (Figure 1C). We compared the effectiveness of three electronic methods for reducing speckle in clinical images acquired with a widefield fundus camera.
Methods :
A prototype 90° widefield slit-scanning fundus camera with a 790 nm single-mode infrared laser diode was used for video preview imaging at 10 frames/second. In order to suppress speckle, the laser diode was driven with modulated current pulses. Three different pulse types were tested: i) overdriven, ii) sawtooth, iii) high-frequency modulated. Pulse-shaping induces thermal tuning of the laser wavelength, with the resulting superposition of speckle leading to reduced speckle contrast (SC) in the final image. We imaged N=9 human subjects, without pupil dilation. SC was calculated directly from the image data in each case (Figure 2).
Results :
Mean SC across subjects was 0.12±0.024 for overdriven pulses, 0.11±0.018 for sawtooth pulses, and 0.12±0.017 for high-frequency modulated pulses. Overall, the achieved SC was not significantly different among any of the three tested methods (p>0.05).
Conclusions :
Pulse-modulation of laser diode current can provide a simple and low-cost means of reducing the impact of speckle in coherent widefield fundus imaging. Though we found that three different pulsing schemes achieved similar levels of SC reduction in fundus images, there may be other important metrics (such as impact on laser lifetime) to consider when choosing the most suitable method for a given application.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.