This program focuses on the design and development of an automated vision inspection station for the classification of blood specimens in the medical laboratory. The proposed system will classify specimens according to tube type, cap color, volume of contents, bar code, and problems with sample integrity such as hemolysis, icterus, or lipemia. The system will replace the subjective decisions made by clerical laboratory staff with quantitative measures that will become a permanent part of the medical specimen record. The current manual procedures are slow, labor intensive, and a major source of life threatening errors. The proposed system may be used either standalone or as part of a fully automated clinical sample processing system. Application areas include both the hospital clinical laboratory and in commercial blood banks for evaluating blood sample quality. Phase l tasks include analysis of application requirements, breadboard prototyping of imaging concepts and image processing/analysis options, configuration of a preliminary system design, and limited prototype evaluation. Phase l prototyping will be performed on state-of- the-art image processing/analysis systems implemented on standard computer platforms. Based on this work future R&D needs will be identified and the Phase Il effort planned.Proposed commercial application:The proposed system will result in improved consistency, greater efficiency, lower costs, and reduced downtime in clinical laboratories. In addition the core technology to be developed will be applicable in other potentially high volume applications such as sample evaluation in blood banking, turbidity measurement for pharmaceutical compound screening, blood clotting and typing tests, and evaluating the solubility of soil samples in environmental testing applications.National Institute of Heart, Lung, and Blood Institute (NHLBI)
