The client—a subsidiary of one of the  largest global medical device,  pharmaceutical, and consumer goods  manufacturers—needed an automated  method for identifying defects in the  contact lenses it manufactures. It also  needed to be able to classify those  defects in real time (i.e., on the  assembly line) so that appropriate,  immediate corrective measures could be  taken.

The client sells a wide range of lenses,  which are produced in masses. Prior to  the solution delivered by Intuceo’s  Platform, lens defects had to be checked  manually—an approach that was costly,  tedious, and prone to human error. The  impact on the business was obvious:  needless additional manufacturing and  logistical costs (e.g., due to product  returns) as well as lower customer  satisfaction.

Lens defects involve edge tears and  surface distoritions. The client provided  images (i.e., unstructured data) of non-  defective and defective lenses to help  the Intuceo team identify, compare, and categorize defect types. Intuceo created a  means of reading the image data and  measuring the deviation between non-  defective and defective lenses of various  types. With the implementation of a  custom algorithm to classify defects,  (taking into account the differences of  lens radii) the client was able to plot  their outcomes (see figure below) and  easily classify the defective lens in real  time.

Intuceo developed an algorithm that  enabled users to categorize lens defects  into various classes which employs  heuristics (i.e., if–then rules) to classify  defective lenses with top notch  accuracy. Overall, the client was able to  identify and classify defective lenses in  real time, resulting in manufacturing and  logistical cost savings, better product  quality, and moreover, a substantial  competitive edge.

DOWNLOAD NOW