Lattice AI stack supports edge facial biometrics with 10X performance boost

May 21, 2019, 12:04 pm EDT | Chris Burt

Low-power programmable chip maker Lattice Semiconductor has enhanced its Lattice sensAI hardware and software stack with a 10-times performance boost over its previous version to bring always-on AI functionality to low-power edge devices.

Latency, bandwidth, and privacy considerations are driving OEMs to design always-on edge devices with minimal need to send data to the cloud, and IHS forecasts there will be 40 billion devices operating at the edge by 2025, according to the announcement. Those OEMs can use Lattice sensAI to add low power AI inferencing optimized for their application environments to existing systems.

“At Pixcellence, our expertise is in developing image processing and computer vision solutions with industry-first features like color night vision. Interest in the IoT is fueling demand for smart cameras that support AI applications like presence detection or facial recognition,” says Pixcellence, Inc. CEO Seton P. Kasmir. “However, smart cameras have strict power consumption and cost requirements that make it a challenge to use off-the-shelf ASSPs. The Lattice sensAI solutions stack lets us easily add low power, flexible AI inference support to our existing and new camera designs, and get our value-added products to market faster.”

The solution’s enhanced performance is driven by an updated CNN IP and neural network compiler with 8-bit activation quantization, smart layer merging, and a dual-DSP engine. Lattice sensAI also features expanding neural network and machine learning frameworks support, including for Keras, support for quantization and fraction setting schemes for neural network training to eliminate iterative post-processing, and simple neural network debugging through USB. Customizable reference designs can accelerate time to market for object counting, presence detection and other popular use cases.

“Our sensAI stack was developed to address the growing need for Edge devices with AI support, and we’re encouraged by the market’s response to our solution,” comments Lattice Semiconductor Senior Director of Segment and Solutions Marketing Deepak Boppana. “In addition to the multiple industry awards received by sensAI, it is exciting to see the broader partner ecosystem and customers developing low power sensAI-based solutions for always-on IoT devices such as smart doorbells and security cameras.”

Lanner launched a vision analytics solution for low-power edge processing in April, and bringing biometrics and speech recognition capabilities to the edge was a focus at this year’s MWC Barcelona 2019.