Renesas
AI Hub
1.00
The best solution for
starting your AI applications.
Provided by Renesas Electronics Corporation
Vision
Overview
Overview
Renesas’ Vision AI delivers high‑performance, low‑power computer vision through its DRP‑AI accelerator of MPUs and scalable lineup of MCUs, enabling efficient processing for applications ranging from building automations, industrial inspection, and to robotics. The extensive library of pre‑trained models and developer‑friendly workflows allows all skill levels to rapidly prototype and deploy vision solutions at the edge.
Applications
DetectNet_v2 delivers fast, efficient object detection for embedded AI, combining a ResNet‑18 backbone with ONNX‑ready DRP‑AI deployment and streamlined pruning and QAT support through the Renesas AI Model Deployer (RAIMD) —enabling high‑performance inference without reliance on TensorRT‑specific layers.
MobileNet_v2, a lightweight image‑classification model optimized for real‑time embedded applications. Using a MobileNet_v2 backbone exported to TFLite for tflite‑micro and accelerated by the Cortex‑M85 Helium MVE on RA8D1, it supports pruning and configurable hyperparameters through the Renesas AI Model Deployer (RAIMD).
Real‑time face detection using the RUHMI Framework, capturing camera images and identifying faces with a lightweight neural network. Results appear on an LCD with bounding boxes.
A MobileNet_v1 image‑classification application runs on an MCU with Arm Ethos‑U55 acceleration, using the RUHMI Framework for efficient deployment and delivering real‑time, low‑power inference from live camera input with top‑5 results shown on an LCD.
