Vehicle detection using a hardware-implemented neural net
S. Mantri, Darcy Michael Bullock, James H. Garrett · IEEE Expert · 1997
The authors describe how they developed a vehicle-detection model based on a radial basis function network and implemented it using the NI1000 recognition accelerator, which can classify up to 32,000 patterns per second-several hundred times faster than the software approach to neural net processing. A detection system using this implementation had a success rate greater than 90%.