Abstract: A concealed object detection system for detecting objects concealed on a person includes radar transponders that are each configured and positioned to direct a radar signal at a person and to detect a portion of the radar signal reflected by the person. A processor connected to the radar transponders processes the portions of the radar signals detected by the radar transponders to determine whether the person is carrying a concealed object. The system may produce a real-time alert, such as an audible alert, when a concealed object is detected.

Abstract: A handheld radar frequency scanner system for detecting objects concealed on a person, the system includes a radar transmitter coupled to a transmit antenna that is configured and positioned to direct a radar signal at a person, a radar receiver coupled to a receive antenna that is configured to detect a portion of the radar signal reflected by the person, and a processor connected to the radar receiver and operable to process the portions of the radar signals detected by the radar receiver to determine whether the person is carrying a concealed object. The system may produce a real-time alert, such as an audible alert, when a concealed object is detected.

Abstract: A concealed object detection system for detecting objects concealed on a person includes radar transponders that are each configured and positioned to direct a radar signal at a person and to detect a portion of the radar signal reflected by the person. A processor connected to the radar transponders processes the portions of the radar signals detected by the radar transponders to determine whether the person is carrying a concealed object. The system may produce a real-time alert, such as an audible alert, when a concealed object is detected.

Abstract: A real-time learning (RTL) neural network is capable of indicating when an input feature vector is novel with respect to feature vectors contained within its training data set, and is capable of learning to generate a correct response to a new data vector while maintaining correct responses to previously learned data vectors without requiring that the neural network be retrained on the previously learned data. The neural network has a sensor for inputting a feature vector, a first layer and a second layer. The feature vector is supplied to the first layer which may have one or more declared and unused nodes. During training, the input feature vector is clustered to a declared node only if it lies within a hypervolume defined by the declared node's automatically selectable reject radius, else the input feature vector is clustered to an unused node. Clustering in overlapping hypervolumes is determined by a decision surface.

Abstract: A combination of image magnification and/or image zoom, superresolution processing techniques, and a parallel (SIMD) machine implementation achieve image enhancement and resolution that are better than diffraction-limited image enhancement and hence suitable for machine processing pattern recognition systems. One application is the separation of (normally) unresolved point source objects from background clutter and adjacent point objects that would, without superresolution processing, be indistinguishable. Furthermore, the real-time processing of the present invention allows the shapes of objects to be discerned at ranges two to four times greater than would normally be achievable.