Human Presence Detection

Lattice sensAI Reference Design

Add Human Presence detection (or any specific object of interest) to any Device – The human presence detection reference design enables systems to always search for the presence of a human via a CMOS image sensor. With an updated training model using the deep learning frameworks Caffe or Tensorflow, your AI systems can detect and locate any object of interest.

Optimized Power and Performance – The reference design enables a scalable systems design consuming between 1 mW and 0.85 W of power consumption depending on the chosen FPGA.

Production-priced, Space-efficient Design – The scalable inference engine fits inside of the iCE40 UltraPlus-5K FPGA and can be scaled up to fit in several devices of the ECP5 FPGA with increased performance.


  • Accelerated, low-power human presence detection at the network edge using neural network model
  • iCE40 UltraPlus
    • VGG8 like 16-bit CNN
    • 64*64*3 input
    • 6 zone searching
    • Up to 8 frames per second
    • 1 mW of power consumption
  • ECP5 85
    • VGG8 like 16-bit CNN
    • 128*128*3 input
    • 6 zone searching
    • 15 frames per second depending on network selection
    • 0.85 W of power consumption
  • Adjustable Frame-rate
  • Can be optimized between power and response time depending on system needs
Lattice sensAI

Jump to

Block Diagram


Technical Resources
Human Presence Detection Using CNN Accelerator IP - Documentation
FPGA-RD-02042 1.0 9/26/2018 PDF 1.2 MB
Human Presence Detection Using CNN Accelerator IP - Project Files
1.0 9/26/2018 ZIP 59.9 MB
Human Presence Detection Using Compact CNN Accelerator IP - Documentation
FPGA-RD-02045 1.1 10/11/2018 PDF 545.5 KB
Human Presence Detection Using Compact CNN Accelerator IP Detect Low Power - Project Files
1.1 10/11/2018 ZIP 1.1 MB

Like most websites, we use cookies and similar technologies to enhance your user experience. We also allow third parties to place cookies on our website. By continuing to use this website you consent to the use of cookies as described in our Cookie Policy.