Super-Resolution Imaging Combining The Design Of An Optical Coherence Microscope Objective With Liquid-Lens Based Dynamic Focusing Capability And Computational Methods

In this paper, we present the design of a 0.2 NA microscope objective operating across a 120nm broadband spectral range that requires only two doublets and an embedded liquid lens to achieve 3μm invariant lateral resolution throughout a large 8 cubic millimeter imaging sample. Achieving invariant lateral resolution comes with some sacrifice in imaging speed, yet in the approach proposed, high speed in vivo imaging is maintained up to a resolution of 3μm for a 2×2 mm sample size. Thus, in anticipation to ultimately aim for a resolution of 0.5 to 1μm, we are investigating the possibility to further gain in resolution using super-resolution methods so both hardware solutions and image processing methods together can provide the best trade-off in overall resolution and speed of imaging. As a starting point to investigate super-resolution methods, we evaluate in this paper three well-known super-resolution algorithms used to reconstruct a high resolution image from down-sampled low resolution images of an African frog tadpole acquired en face using our OCM set-up. To establish ground truth necessary for assessment of the methods, low resolution images were simulated from a high resolution OCM image. The specification and design performance of the custom designed microscope will be presented as well as our first results of super-resolution imaging. The performance of each algorithm was analyzed and all performances compared using two different metrics. Early results indicate that super-resolution may play a significant role in the optimization of high invariant resolution OCM systems. © 2008 SPIE