The brain mimicking Visual Attention Engine: An 80×60 digital Cellular Neural Network for rapid global feature extraction

Abstract

The Visual Attention Engine(VAE), an 80x60 digital Cellular Neural Network, rapidly extracts global features used as attentional cues to streamline detailed object recognition. A peak performance of 24GOPS is achieved by 120 processing elements (PE) shared by the cells. 2D Shift register based data transactions enable 93% PE utilization. Integrated within an object recognition SoC, the 4.5mm2 VAE running at 200MHz improves object recognition frame rate by 83% while consuming just 84mW.