Attribute Data Treatment of Automated Inspection Vision System For Product Mix-Up Detection

Abstract

The pharmaceutical industry continues to grow at an exponential pace, with number of medications being mixed, molded, stamped, and packaged every day in the millions. Regardless of the amount, the margin for error never changes. To prevent any mistakes, the human factor has been reduced and replaced by high speed vision inspection systems. This paper presents a statistical evaluation of product mix-up detection confidence levels, confidence intervals and sample size considerations for a filler vision system (i.e., Optel Vision Inspection) in bottle packaging line using executed engineering studies of “XYZ” Pharmaceutical Industry located at Puerto Rico. An Attribute Agreement Analysis will be used to investigate whether this system can be used for detect different tablet’s defects as broken, different shapes or colors presentations to assess the consistency of responses of appraisers vs. standard reference for the inspection system. Key Terms: Appraiser, Attribute Agreement Analysis, Confidence Intervals, Confidence Level.