This section of my web page will be devoted to Neural Networks and their uses. Given the restrictive nature of intellectual property agreements today (yes I do work for a corporation that believes in the suck principle of IP - all in and none out) there will be limits on what I can say, but you may find some interesting studies of neural networks here soon.
Ventek is now back in private ownership and I can now discuss a few things. Creating the GS2000 Veneer Grader and some other "skunk works" projects, and keeping them upgraded against the changing Win-Tel software and hardware bugs, has been a very demanding job during the last few years.
Skunk Works Revealed:
The first North American Robotic Veneer Patching system. The folks in Finland may have invented this first, but I helped perfect it for our market.
First, some references to the work:
A very good book for learning the MLP (MultiLayer Perceptron): Neural Smithing - Supervised Learning in Feedforward Artificial Neural Networks by Reed & Marks II., MIT Press.
A good machine vision book with a short reference to the GS2000: Smart Inspection Systems - Techniques and Applications of Intelligent Vision, by D. T. Pham and R. J. Alcock, Academic Press.
Articles recently published:
Introduction to Machine Vision and MLPs (ScanPro 2000 Paper, 2.2MB PDF)
Matrox Newsletter (Oct. 2003, http://www.matrox.com/imaging/news_events/feature/lumbersupport.cfm)
Basler Camera Newsletter (near future)
Other References (the Fins created some of the first machine vision wood grading systems):
The Mecano Paper presented at ICANN'95, Paris: Wood Surface Inspection System Based on Generic Visual Features by Lampinen, Smolander and Korhonen.
Research Thesis covering details: DEVELOPMENT OF A COLOR MACHINE VISION METHOD FOR WOOD SURFACE INSPECTION by HANNU KAUPPINEN, Department of Electrical Engineering and Infotech Oulu, 1999
Download at: http://herkules.oulu.fi/isbn9514254244/
Interesting Sales Promotion Claim (I let this item speak for itself): a 100% accurate machine vision system!
The Folding-Hyperspace animation explained in Pictures: