Lattice Semiconductor Corporation announced major performance and design flow enhancements for its award-winning Lattice sensAI solutions stack. The Lattice sensAI stack provides a comprehensive hardware and software solution for implementing low power (1mW-1W), always-on artificial intelligence (AI) functionality in smart devices operating at the edge. Additionally, enhancements to the stack’s memory schemes address a key bottleneck for AI designs.

IHS forecasts 40 billion devices will be operating at the network Edge by 2025. For reasons including latency, network bandwidth limitations, and data privacy, OEMs designing always-on Edge devices want to minimize sending data to the Cloud for analytics. Lattice sensAI enables such OEMs to seamlessly update their existing designs with low power AI inferencing optimized for their application requirements. Incorporating such local intelligence also lowers expenses related to Cloud-based analytics by only sending relevant data for further processing.

“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,” said Seton P. Kasmir, CEO, Pixcellence, Inc. “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.”

“There is a desire to enable AI-functionality in always-on applications involving presence detection and object counting in smart home control, industrial, and retail store cameras, for instance, that are also moving AI to the edge” says Boppana, Sr. Director of Segment & Solutions Marketing, at Lattice Semiconductor, “and because we have always focused on low power as a company, our FPGAs are also a good fit for these applications where every milliwatt counts, small form-factors are a must, and users require the flexibility to easily adapt to the evolving AI space.”

New enhancements to the Lattice sensAI solution stack include:

• 10x performance boost over previous version – driven by an updated CNN IP and neural network compiler with features like 8-bit activation quantization, smart layer merging and a dual-DSP engine
• Seamless user experience – accelerates the design cycle with several new features including:
  • Expanding neural network and ML frameworks support including Keras
  • Support for quantization and fraction setting schemes for neural network training eliminates iterative post-processing
  • Simple neural network debugging via USB
  • New customizable reference designs accelerate time to market for popular use cases like object counting and presence detection
• A growing design service partner ecosystem including full product design capability from partners such as Pixcellence makes it easier for customers to get to market faster.

Customers with prior experience in machine learning and AI can access the customizable reference designs that allow them to change the training data to optimize their applications without worrying about which neural network to choose.