Statistical Imaging for Modeling and Identification of Bacterial Types

Sigal Trattner, Hayit Greenspan, Gabi Tepper, Shimon Abboud

Abstract: An automatic tool is developed to identify microbiological data types using computer-vision and statistical modeling techniques. In bacteriophage (phage) typing, representative profiles of bacterial types are extracted. Currently, systems rely on the subjective reading of the profiles by a human expert. This process is time-consuming and prone to errors. The statistical methodology presented in this work, provides for an automated, objective and robust analysis of the visual data, along with the ability to cope with increasing data volumes. Validation is performed by a comparison to an expert manual segmentation and labeling of the phage profiles.