2025-07-29 11:52 AM - last edited on 2025-09-18 6:20 AM by Andrew Neil
I've implemented a multi-model face recognition system on the STM32N6 (originally based on the single model object detection demo) that efficiently runs detection and recognition models simultaneously on the NPU. Sharing the complete implementation since deploying and managing multiple AI models on the hardware has some non-obvious challenges.
Repository: https://github.com/PeleAB/STM32N6-FaceRecognition
Next steps include creating an abstraction layer for JPEG and H264 compression to handle video streams more efficiently.
Happy to discuss multi-model deployment strategies or NPU optimization approaches. The automated build system alone might save others significant setup time.
2025-07-30 12:02 AM
Hello @PeleAB,
Thank you very much for sharing your work!
Have a good day,
Julian
2025-09-18 6:06 AM
Hello Pele AB,
Thanks for sharing,
I have a few questions though,
Does St modelzoo services really provide a solution of face verification achitectures, like "edgeface, mobilefacenet, ...", Yes I know it has the required backbones but the required lose functions implementations, output shapes and a few other specifics for this scenarios are a bit different.
How exactly did you do this?, Did you have use the yaml file :thinking_face:, or Did you have to custome change the code alitle in the right places?
Am a bit curious.