Digital super-resolution microscopy using example-based algorithm

We propose a super-resolution microscopy with a confocal optical setup and an example-based algorithm. The example-based super-resolution algorithm was performed by an example database which is constructed by learning a lot of sets of a high-resolution patch and a low-resolution patch. The high-resolution patch is a part of the high-resolution image of an object model expressed in a computer, and the low-resolution patch is calculated from the high-resolution patch in consideration with a spatial property of an optical microscope. In the reconstruction process, a low-resolution image observed by the confocal optical setup with an image sensor is converted to the super-resolved high-resolution image selected by a pattern matching method from the example database. We demonstrate the adequate selection of the patch size and the weighting superposition method performs the super resolution with a low signal-to noise ratio.