Object Detection via SSD (Single-Shot Detection) on the RZ/G1 MPU

Added on 2019-06-20

OpenCV DNN inference running a Mobilenet SSD model demonstrate an end-point AI solution delivered through high-performance CPU cores in the RZ/G1 general purpose MPU. This is one of six demos Renesas Electronics presented at the Embedded Vision Summit 2019 in Santa Clara, CA, highlighting the capabilities of RZ/A and RZ/G MPUs for low-power, high-performance machine vision systems.

Key Features

  • Lowest power consumption and price for a CPU with the resources and performance required for this type of algorithm
  • Developed with free open-source frameworks and models
  • Approximately 70MB of weights in DDR memory
  • Ability to detect 90 different classes of objects
  • Get started with the RZ/G Linux Platform