“Deploy Your Embedded Vision Solution on Any Processor Using Edge Impulse,” A Presentation from Edge Impulse
0 likes30 views
The document discusses Edge Impulse's embedded machine learning platform, designed for developing AI applications across various devices and processors. It emphasizes the complexity of building such applications and highlights Edge Impulse's ability to streamline development with built-in MLOps infrastructure. Additionally, it showcases the integration of machine learning capabilities in various semiconductor products to enhance IoT applications by improving efficiency and smart decision-making.
“Deploy Your Embedded Vision Solution on Any Processor Using Edge Impulse,” A Presentation from Edge Impulse
- 1. Deploy Your Embedded Vision Solution on Any Processor Using Edge Impulse Amir Sherman Senior Director Global Business Development Semiconductors & Eco Partners Edge Impulse
- 2. From cloud AI to edge AI/ML to endpoint AI 2 © 2023 Edge Impulse
- 3. Cameras are everywhere 3 © 2023 Edge Impulse
- 4. What is the “best/right” technology ? 4 © 2023 Edge Impulse
- 5. But it is all so complex! 5 © 2023 Edge Impulse
- 6. The leading embedded ML platform for any technology – Edge Impulse 6 © 2023 Edge Impulse
- 7. Typical development of EdgeML applications 7 © 2023 Edge Impulse Requires 20+ man years, expertise in ML and embedded to build the infrastructure and integrate dozens of different tools.
- 8. Develop EdgeML applications with Edge Impulse 8 © 2023 Edge Impulse An end-to-end platform for projects using any data or device, built for developers with MLOps infrastructure built-in.
- 9. Some of the semiconductors & IP’s we support 9 © 2023 Edge Impulse
- 10. How it looks ? 10
- 11. BYOM-Bring Your Own Model 11 © 2023 Edge Impulse
- 12. 12 © 2023 Edge Impulse Technology Examples
- 13. 13 © 2023 Edge Impulse
- 14. 14 © 2023 Edge Impulse AI/ML capabilities of the EFR32MG24 with MVP Bringing machine learning (ML) to IoT applications reduces bandwidth requirements, saves power, and increases a device’s ability to make smarter decisions. Silicon Labs supports machine learning in all Series 1 and Series 2 wireless SoCs including newly released BG24 and MG24 products with built-in AI/ML hardware accelerator. The MVP accelerator is a co-processor designed to perform matrix and vector operations. Using hardware accelerated kernel implementations will reduce neural network inference time, as well as off-load the main processor to allow it to perform other tasks or go to sleep.
- 15. 14 © 2023 Edge Impulse AI/ML capabilities of the EFR32MG24 with MVP Bringing machine learning (ML) to IoT applications reduces bandwidth requirements, saves power, and increases a device’s ability to make smarter decisions. Silicon Labs supports machine learning in all Series 1 and Series 2 wireless SoCs including newly released BG24 and MG24 products with built-in AI/ML hardware accelerator. The MVP accelerator is a co-processor designed to perform matrix and vector operations. Using hardware accelerated kernel implementations will reduce neural network inference time, as well as off-load the main processor to allow it to perform other tasks or go to sleep.
- 16. 14 © 2023 Edge Impulse AI/ML capabilities of the EFR32MG24 with MVP For this project, we attached an Arducam mini 2MP plus to the xG24 Dev Kit in order to capture low-res images of people flow from a real environment.
- 17. 17 © 2023 Edge Impulse AI/ML capabilities of the EFR32MG24 with MVP For this project, we attached an Arducam mini 2MP plus to the xG24 Dev Kit in order to capture low-res images of people flow from a real environment.
- 18. 18 © 2023 Edge Impulse
- 19. ARM’s latest CortexM55 & microNPU Ethos-U55 16 © 2023 Edge Impulse
- 20. ARM’s latest CortexM55 & microNPU Ethos-U55 16 © 2023 Edge Impulse
- 21. ARM’s latest CortexM55 & microNPU Ethos-U55 16 © 2023 Edge Impulse
- 22. Official support for the Alif’s Ensemble family 22 © 2023 Edge Impulse
- 23. Astounding AI/ML performance benchmark 23 © 2023 Edge Impulse
- 24. 24 © 2023 Edge Impulse
- 25. Offical support for Renesas RZ/V2L 2 x CortexA55 and DRP-AI ML accelerator 20 © 2023 Edge Impulse
- 26. Offical support for Renesas RZ/V2L 2 x CortexA55 and DRP-AI ML accelerator 20 © 2023 Edge Impulse
- 27. 27 © 2023 Edge Impulse
- 28. TDA4VM multi-core embedded vision processor 28 © 2023 Edge Impulse
- 29. Industrial edge compute based on NVIDIA 29 © 2023 Edge Impulse
- 30. And so many other cores are supported 24 © 2023 Edge Impulse
- 31. And so many other cores are supported 24 © 2023 Edge Impulse
- 32. And so many other cores are supported 25 © 2023 Edge Impulse
- 33. And so many other cores are supported 25 © 2023 Edge Impulse
- 34. And so many other cores are supported 25 © 2023 Edge Impulse
- 35. Our business model © 2023 Edge Impulse 35
- 36. A platform that goes from data to algorithms © 2023 Edge Impulse 36
- 37. 37 © 2023 Edge Impulse Amir Sherman Senior Global Director, Semiconductor @ Ecosystem Business Development +972-52-2240811 +49-173-3232288 amir@edgeimpulse.com www.edgeimpulse.com