Millimeter-wave imaging using neural networks for object recognition

An active-mode mm-wave (60 GHz) imaging system with Yagi antenna has been developed. The optics for the system was designed with the ray tracing method to reduce aberration. A signal processing using a neural network has been successfully introduced to recognize objects distorted with coherent mm-wave illumination. With 10/spl times/10 sampling points the recognition rate of 98% has been obtained for the objects of 10 alphabetical letters and the 5 teaching trials.