Research & Development

Automatic TDP Classification

Problem / Purpose
  • DOD noticed that metal parts manufactured by suppliers using casting and forging processes had a higher failed inspection rate
  • DOD also noted that metal parts that were cast and forged also had a higher unfilled order rate
  • The meta-data associated with a given NSN in JEDMICS does not have an identifier for procurement personnel to know if an NSN requires casting and forging processes, because this information, along with other valuable process requirement information, is currently located in the drawing
Solution

VizSeek’s shape search and OCR technologies index the raster drawings and capture all text on the drawing. This text is then intelligently matched against manufacturing process specific information during indexing to determine the probable process. The related process requirement information is also captured and added to the JEDMICS meta-data field.

Customers / Stakeholders
  • Defense Logistics Agency
  • Jacobs
Benefits / Milestones
  • Achieved 92% extraction accuracy for legacy drawings
  • Enabled procurement personnel to easily identify cast and forged parts prior to generating a solicitation to add process requirement information to help suppliers
  • Enabled the CAST-IT and FORG-IT teams to now rely on the extracted text information prior to diagnosing problem parts
Copyright © 2011 Imaginestics, LLC Imaginestics is a U.S. Small Business Administration 8(a) certified small disadvantaged business.