Automated combination of optical coherence tomography images and fundus images

We discuss approaches to combining multimodal multidimensional images, namely, three-dimensional optical coherence tomography (OCT) data and two-dimensional color images of the fundus. Registration these two modalities can help adjust position obtained OCT on retina. Some existing matching fundus are based finding key points that considered invariant affine transformations common images. However, errors in identification such lead registration errors. There also methods for iterative adjustment conversion parameters, but they some manual settings. In this paper, we propose a method full or partial search possible combinations image transformation find best approximation true transformation. The is determined using measure comparison preprocessed pixels. Further, compared with available assess quality algorithm. structure work includes: pre-processing extraction blood vessels, random grid over parameters (shift, rotation scaling), evaluation