Title:  Novel mathematical techniques for the analysis of bio-medical data.
<br><br>
Abstract: We introduce several new manifold learning and recovery techniques for the analysis of bio-medical data. These methods are based on implementation of data-dependent graph operators, such as Laplacian or Schroedinger operators with appropriately constructed barrier potentials as carriers of labeled information. We apply them to analyze two complex bio-medical datasets: multispectral retinal images and microarray gene expressions.