A quadratic programming approach for the mosaicing of virtual slides that incorporates the positioning accuracy of the microscope stage

We describe a novel approach for creating virtual slides that incorporates the positioning accuracy of the microscope stage in the optimization step. To capture a complete slide in microscopy, a large number of fields of view have to be acquired by moving the microscope stage in a controlled way. These fields of view are aligned in such a way that a globally consistent virtual slide is formed. However, depending on the positioning repeatability of the stage and the accuracy of the stage calibration, this results in alignment errors. These errors are usually resolved by applying a mosaicing algorithm. Our algorithm extends known mosaicing approaches by analyzing the positioning accuracy of the stage and incorporating this knowledge to make the mosaicing process more robust.