Enabling Visual AI at the Edge: From Surveillance Cameras to People Counters

New AI-at-the-edge processors with improved efficiencies and flexibility are unleashing a huge opportunity to democratize computer vision broadly across all markets, enabling edge AI devices with small, low-cost, low-power cameras. Synaptics has embarked on a roadmap of edge-AI DNN processors targeted at a range of real-time computer vision and multimedia applications. These span from enhancing the image quality of a high-resolution camera’s output using our VS680 multi-TOPS processor to performing computer vision in battery-powered devices at lower resolution using our Katana Edge-AI SoC. In this talk, we will show how these edge AI SoCs can be used to: Achieve exceptional color video in very low light conditions; De-noise and distortion-correct both 2D and 3D imagery from a time-of-flight depth camera that images through a smartphone OLED display; Perform super-resolution enhancement of high-resolution video imagery; and Recognize objects using lower-resolution sensors under battery power.