Hand Gesture Detection

Lattice sensAI Reference Design

Use Hand Gesture detection to enable smart IoT devices – The hand gesture detection reference design enables systems to detect open hand / closed hand gestures using compact Convolutional Neural Network (CNN) running on the iCE40 UltraPlus FPGA. Designers can retrain the same network with an updated training model using the deep learning framework, Caffe to detect any hand gesture of interest.

Ultra Low Power Consumption enables new applications – Systems can now enable artificial intelligence with an always-on image sensor, while consuming less than 3.3 mW of average active power.

Production-priced in a tiny package – The complete inference engine is implemented in the 5.5 mm2 iCE40 UltraPlus-5K FPGA and works with various popular image sensors. The reference design’s functionality has been validated with Himax’s HM01B0 low cost / power image sensor.

  • Detect hand gestures 6-12 inch away from the image sensor
  • 3.3 mW power consumption at 5 frames per second, frame-rate can be adjusted trading off with power consumption
  • Hardware platform available to demonstrate this reference design using UPduino and Himax’s HM01B0 image sensor
  • The model can be retained to recognize other hand gestures
Lattice sensAI

Jump to

Block Diagram


Technical Resources
Hand Gesture Detection Using Compact CNN Accelerator IP - Documentation
FPGA-RD-02046 1.0 9/26/2018 PDF 813.1 KB
Hand Gesture Detection Using Compact CNN Accelerator IP - Project Files
1.0 9/11/2018 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.