CMOL-Based Cellular Neural Networks and Parallel Processor for Future Image Processing

Hybrid CMOS/molecular (CMOL) circuits are promising for future high-performance VLSIs. Recently, digital and mixed-signal CMOL-based image-processing circuits were proposed. Although these circuits have ultra-high performances, several problems exist. In this paper, CMOL-based analog cellular neural network (CNN) and digital parallel image processor is proposed. The CMOL-based CNN has high speed and good fabrication tolerance. The parallel processor has high peak performance with easy configurability.