Microscopic Image Recognition

At some fundamental level, every science is reduced to counting spots. Whether it be counting the number of stars in a particular quadrant or even the number of peas in a pod, at some point you must face the dots. It is inescapable. Unlike astronomers and particle physicists who have developed sophisticated software for tabulating dots, however, biologists had yet to develop an effective time-saving system that is comparable to what their colleagues have achieved in the physical sciences--until

Mar 15, 1999
Brent Johnson

At some fundamental level, every science is reduced to counting spots. Whether it be counting the number of stars in a particular quadrant or even the number of peas in a pod, at some point you must face the dots. It is inescapable. Unlike astronomers and particle physicists who have developed sophisticated software for tabulating dots, however, biologists had yet to develop an effective time-saving system that is comparable to what their colleagues have achieved in the physical sciences--until now.

BioSeparations of Tucson, Ariz., has developed Image Scan for life sciences and clinical researchers. Image Scan is designed to automatically scan the entire circle area of a microscope slide while segmenting, enumerating, and measuring cells and other objects specified by the user. It accomplishes this feat by dividing the focal area into numerous minifields. As the stage of the microscope moves--producing a scanning graphic--each field is analyzed for labeled spots such as chromosomes. The graphic indicates fields containing the subject of interest by appropriately shading them--either yellow or green--and the information is documented for quick reference. The computer uses autofocus and a compilation of z-stack images to automatically locate the labeled chromosomes. User-defined parameters can also be set to allow the selection of different object sizes, shapes, and colors.

BioSeparations plans to use ImageScan in its clinical trials to enumerate chromosomes in fetal cells separated from maternal blood. In Down syndrome, for instance, the presence of more than two copies of the 21st chromosome is usually associated with the abnormality. The ImageScan software is programmed to search for rare events--in this case, Down syndrome--and the selected fields that contain the chromosome are marked red. ImageScan software can also be easily adapted to a variety of other applications.

Betty Strietelmeier, an environmental scientist at Los Alamos National Laboratory, recently completed a training seminar on the new instrument and was very impressed with what it had to offer.

"The capabilities are tremendous. I just marvel at what BioSeparations has done," commented Strietelmeier. In her research, Strietelmeier studies microbes for the purpose of removing environmental pollutants. By investigating organisms that thrive on man-made carbon compounds, she attempts to engineer systems using native organisms that will metabolize harmful wastes. The ability of ImageScan to automatically sort out microorganisms and help isolate pure cultures makes it ideal for the kind of work that she does. In fact, she sees a number of possible applications for the ImageScan system.

Strietelmeier is using the software for her project that concerns the long-term isolation of low-level radioactive waste. Using bacteria inoculated into synthetic brine, researchers at the laboratory are attempting to simulate a release scenario in the event that a containment facility is breached. The ImageScan system will allow them to more effectively determine contamination that might occur by measuring the numbers of these bacteria that survive over time in order to help predict the effects of actinide speciation.

The author, Brent Johnson, can be contacted at bjohnson@the-scientist.com. For more information please contact (520) 622-5882, or visit the company website at www.bioseparations.com.