Hybrid Electrical-Optical Resonator for Image Classification

Optical correlators are known to be fastest engine for template based image recognition. There are, however, applications that require speeds even faster than the best optical correlators have accomplished. This effort will demonstrate the Hybrid Electrical-Optical Resonator for Image Classification (HEORIC) and correlation plane discrimination enhancement. The HEORIC has been demonstrated in simulations to provide extrememly high speed image recognition and classification. The HEORIC is a system which performs up to one thousand full frame image correlations in parallel and converges on the best match in order one microsecond. This is one million times faster than any competitor's system. We will demonstrate a scaled down version of the systems in the laboratory as Phase I proof of the concept and develop correlation plane discrimination enhancement further proving the feasibility of the system. In addition to the application of this correlator to advanced missile guidance schemes, this system will find use in applications from high speed fingerprint recognition and counterfeit currency detection for law enforcement to the Post Office for reading mailing addresses