Basic Information
Abstract Number: 60-2    
Author Name: Joseph Sgro Affiliation: FastVision, LLC
Session Title: Machine Vision Solutions for Life Sciences Applications
Event Type: Workshop
Event Title: Machine Vision Solutions for Life Sciences
Presider(s): Earle, Colin Start Time: 01:50 PM ( Slot # 3 )
Date: 03/12/2006 Location: 202A
Keywords: Bioanalytical, Materials Characterization, Surface Analysis, Trace Analysis

Abstract Content
Advances in the medical science, biotechnology, genetic engineering and new requirements for food monitoring has dramatically increased the demand for machine vision solutions in the life sciences The great majority of traditional machine vision tasks can be satisfied with measurements of objects geometrical shape, position and texture. Often the objects of interest are deterministic and are easy to define and scan. Thus the majority of machine vision tasks can be solved by monochrome imaging. Traditional machine vision methods are most useful in production lines where the material composition of objects is known or unimportant. Life science applications often involve objects whose shapes are irregular, moving, and are much harder to define algorithmically. Most life science applications also deal with small, even microscopic objects that are hard to distinguish in visual light without the use of specialized dyes. Machine vision applications in life sciences often require imaging in the light wavelengths (IR, UV, X-ray) that are sensitive to the material composition and the chemistry of the objects. Image analysis for life sciences moves beyond RGB colors and requires multi- or even hyper- spectral analysis. It even includes polarization analysis of light and the laser-induced fluorescence. In some cases, images captured in the separate modalities must be ‘fused’ together producing multi-spectral equivalent that is necessary for the right diagnosis or detection. Irregular and moving shapes, multi-spectral imaging and 3-D objects dramatically increase amount of data that must be collected and processed by the machine vision setup. New machine vision architectures are required to solve demanding requirements imposed by life science applications. We will present several difficult application areas and their complex solutions. These areas are: biotechnology development and drug discovery, food quality and safety, body fluids and tissue analysis.