Extracting discontinuities in early vision with networks of resonant tunneling diodes

Discusses the application of resonant tunneling diodes (RTDs) to the extraction of spatial discontinuities in early vision systems. The current-voltage (I-V) characteristic of a RTD was used to approximate that of a resistive fuse element, which has been previously shown to be useful in spatial feature extraction networks. The resulting element, vertically integrated into a lithography plane for maximal density, is a true two-terminal device with no common-mode limitations and has transition times on the order of a picosecond. This technology offers the potential of extremely high pixel densities with rapid dynamic settling times. A prototype implementation and extensions to the single RTD fuse are considered