How ST makes computer vision more accessible for IoT
Computer vision has always posed a significant challenge for IoT developers. Managing and processing vast amounts of images and data is both daunting and expensive. The STM32N6, a new type of microcontroller (MCU) launched in December 2024, is opening up a new frontier for advanced computer vision on IoT devices. The STM32N6 is not just the most efficient general-purpose STM32 MCU. It enables developers to seamlessly integrate AI into camera-based applications. All this without breaking their power budget. See it in action at IoT Solutions WC 2025, May 13-15, ST booth, Fira Hall 8 # D23.
What makes this new microcontroller so unique?
Two brand-new specialized units inside. This powerhouse integrates a neural processing unit (NPU), also called Neural-ART Accelerator, and an image signal processor (ISP). The ISP pre-processes the images coming from the camera, while the NPU handles neural network inferences with an optimal TOPS/W. What does this mean? In practical terms, you don’t need to send data to the Cloud to sort images. The STM32N6 can store and classify millions of images within its extensive memory. This reduces latency and allows battery-operated devices to perform classification tasks that were previously impossible. Additionally, the ISP can be easily programmed using the ISP IQTune free tool provided by ST. Developers can easily fine-tune image processing parameters.
Quickly bring AI to your microcontroller
Users can focus on their application development without worrying about the complexities of AI model integration and optimization. The STM32N6’s development ecosystem includes a suite of tools that facilitate the generation of optimized codes from your AI models. This ecosystem guarantees that your AI applications run efficiently on the hardware while consuming minimal power.
Computer vision: STM32N6 with Ultralytics models
Come see the STM32N6 in action at IoT Solutions WC 2025 using Ultralytics pre-trained AI models. Ultralytics, an ST Authorized Partner, brings state-of-the-art YOLO models to the STM32N6 MCUs. For those unfamiliar with the power of YOLO models, they enhance data processing, reduce manual effort, and ensure consistent results in computer vision projects. With Ultralytics, anyone can benefit from powerful AI models to enable professional-level solutions.
Discuss your next project. Get answers from ST experts.
See the live demo at our booth, Fira Hall 8 # D23. Join our speech on Tuesday 13th at IoT Solutions WC. Get inspired, read our customer success story.
Article By: STMicroelectronics