A general purpose, low power, analog integrated image edge detector, based on a current-mode Gaussian function circuit

A low power analog integrated image edge detector is proposed consisting of Gaussian function and threshold circuits. The operating principle of this architecture is based on a hardware-friendly approximation of the Robert's Cross operator. The system level implementation can be easily modified to account for various image resolutions. Therefore, the proposed architecture can be used as a building block in larger smart sensor systems. The edge detector was evaluated using 3 medium resolution images, achieving average Peak Signal-to-Noise-Ratio of 27.7 dB and Structural Similarity Index Metric of 0.81, while consuming only 33 nW per pixel. To demonstrate the performance and accuracy of the proposed architecture, Monte-carlo simulation results are provided. Both schematic and post-layout simulations are carried out in 90 nm CMOS process, using the Cadence IC suite.