This talk presents statistical concepts that allow to extract latent information from high-resolution biomedical images such as Computed Tomography scans and Confocal Microscopy images. We first discuss the discrimination of plaque tissue components by analysis of textural consistencies in Micro-CT scans of excised human coronary arteries. The second part of the talk concerns the use of confocal spectrophotometry for the purpose of reliably tracking stem cells labeled with Green Fluorescent Protein on the background of autofluorescence in infarcted murine myocardium. This talk contains many illustrations and is intended for a general audience.