Learn about event-based vision technology and its applications in embedded edge AI systems through this technical talk from Prophesee's CTO. Explore the fundamentals of event-based vision sensing, which offers efficient sparse data acquisition and robust performance in varying lighting conditions. Discover how this emerging paradigm is revolutionizing visual perception systems across industrial, surveillance, IoT, AR/VR, and automotive applications. Delve into Prophesee's latest generation of event sensors, specifically engineered for improved integration and usability in embedded edge vision systems. Examine key features including event data pre-processing, interface compatibility, low-latency connectivity to various processing platforms, and ultra-low power operation modes. Gain insights into the sensor's advanced capabilities such as global contrast detection, noise filtering, anti-flicker functionality, and embedded RISC-V CPU core. Master the technical aspects of power management, data interfaces, and streaming modes while understanding how these innovations address the challenges of implementing event-based vision in resource-constrained edge environments.
Event Sensors for Embedded Edge AI Vision Applications - Design and Implementation
EDGE AI FOUNDATION via YouTube
Overview
Syllabus
Intro
Outline
Image Sensing and Computer Vision
Introduction Event-based Vision
Introduction Event Sensor Acquisition Principle
Event Sensors for Edge-Al Applications
Design Targets
Sensor Chip Overview
Sensor Features Global Contrast Detector
Processing features Noise Filter (NFL)
Processing features Anti-Flicker Filter (AFK)
Processing features AFK Measurements
Processing features Embedded CPU (RISC-V)
Event Sensor → Compute Platform
Output Data Interfaces
Power Modes - Overview
Power Modes and Wake-Up Features
Power Modes Streaming Modes: MIPI vs CPI
Taught by
EDGE AI FOUNDATION
