Creating the Cytoprint™
Atto has developed a multidimensional image analysis approach to extracting patterns from microscopy images. The technique involves applying a large set of phenotypic feature extraction algorithms to each cell in the image, including both biologically intuitive and non-biologically intuitive features. The results of the phenotypic feature extraction algorithms, and higher order combinations of the results, form a fingerprint of the image, called a Cytoprint™.
Processing Cytoprints™
Statistical criteria are used to automatically select those features in the Cytoprint that enable the discrimination of image changes. Atto has developed proprietary similarity metrics for measuring the distance between descriptor vectors resulting from multidimensional image analysis.
Copyright © 2004 Atto Bioscience All rights reserved