Renesas and Witekio: Embedded Artificial Intelligence / Machine Learning

Online, On Demand

Embedded Artificial Intelligence / Machine Learning

Live presentation date: June 16^th, 2020

Technical Seminar: e-AI / ML Considerations and Design examples for Embedded Developers

Live presentation date: June 16, 2020

Presented by Renesas, Witekio & EBV Elektronik

Presented by EBV, Witekio and Renesas AI/ML Technology Specialists this technical workshop will cover Convolutional Neural Networks Theory, practical embedded design considerations, Renesas e-AI RZ/A2M MPU details, e-AI design flow example and Vehicle Number Plate recognition demo implementation running on the RZ evaluation board.

EBV Elektonik is one of the leading specialists in European semiconductor distribution.
EBV are adding value at all stages of our customers' supply chains: Starting with in-depth design expertise and application know-how through value-added services right up to full logistics solutions
ww.ebv.com

Witekio, embedded systems and IoT expert
Specializing in software for the embedded system and IoT objects, Witekio are the co-pilots of your innovation projects
www.witekio.com

Renesas
The RZ/A2M MPU is designed for e-AI based imaging in smart appliances, networked cameras, service robots, scanner products, and industrial machinery that require high-speed image processing. It features a unique hybrid approach to image recognition and machine vision by combining proprietary DRP technology for fast pre-processing of image data and feature extraction, closely coupled to an Arm^® Cortex^®-A9 CPU.
www.renesas.com 

Agenda

• Brief history of Artificial Neural Networks
• Machine Learning, data and tools
• Linear Classification
• Neurons & Activation Functions
• Feedforward, Backpropagation
• Convolutions & Kernels
• CNN Building Blocks
• CNN: LeNet, AlexNet and VGGNet
• CNN: InceptionNet and ResNet
• CNN: classification & object detection
• Books, Docs and Web resources
   

• RZ/A roadmap
• RZ/A2M block diagram
  eAI example flow for RZ/A2M using DRP technology
• data (image) pre-processing flow
• concept of DRP
• supported DRP libraries for image processing
  implementation of a neural network
• training data preparation
• setup of a Convolutional Neural Network
• e-AI translator
• Demo: Number plate detection (RZ/A2M)
   

• ECC to reduce soft errors
• Verified Linux Package
• RZ/G AI Board Support Package (BSP)
• Demo: RZ-G Foss AI object detection demo
   

Outlook for DRP-AI

• Renesas eAI Roadmap
• Demo: POC1.1 with DRP-AI

Register now