260 likes | 265 Views
Jeremy Leakakos George Hodgson Elizabeth Leib Douglas Krofcheck David McClelland. Team Haplocyonopsis. Pro Quo Books Book Cover Scanning System. Synopsis. Who is Pro Quo Books? What is the problem we need to solve?
E N D
Jeremy Leakakos George Hodgson Elizabeth Leib Douglas Krofcheck David McClelland Team Haplocyonopsis Pro Quo Books Book Cover Scanning System
Synopsis • Who is Pro Quo Books? • What is the problem we need to solve? • You'll have a list of 30 million books. You'll be given an image of a book, and you need to find out which of the 30 million it is... in 9 seconds. • What are we doing about it? • Why is this a difficult problem?
Deliverables • Book cover preprocessing library • Book cover identification library • Test application
Requirements Elicitation • Initial sponsor meeting • Conference calls • Application prototype • Performance requirements • 9 second time limit to identify a book • 20% identification rate with 90% confidence
Technologies • C# / Visual Studio • MySQL • AForge.NET • Open source image processing library
Technical Process • Evolutionary delivery • Incremental development • 2-3 week increments • Test app increment • Prototype increment • Product increment
Metrics • Effort: Time tracking • Team and individual • Progress: Average time taken to identify a book • Progress: Number of true positives vs. number of false positives
Testing Strategy • Test application • Deliverable • Testing platform • Simulate production environment • Input rate • Time threshold • Database size
Risks • Unfamiliar domain • Image processing • Inadequate testing material • Test database not large enough • Miscommunication • Misunderstanding of terminology
Design • Pipe and filter architecture • Start with filters to whittle down the dataset • Finish with identification algorithms
Sequence - Preprocessing Fly Fishing
Sequence – Imaging 0.093 chance of matching
Progress so far… • Test application • Multiple functional filters and identification algorithms • Preprocessing algorithms • Solid plan for spring quarter
Reflections • The good stuff • Communication • Test application • Working filters and identification algorithms • Well-defined scope
Reflections • The not so good stuff • Time estimation • Production problems
Spring Quarter • Prototype increments • Products increments • Finish development