[Editor's Note] On-device AI is expanding its application in industries, medicine, and smart homes due to its advantages of reducing cloud dependence and reducing delay, transmission costs, and power consumption. Driven by this expansion of applications, shipments of edge AI-equipped MCUs are expected to reach 1.8 billion units by 2030, and developers' choice of MCU is expected to become an important development point. Among them, STMicroelectronics' STM32N6 MCU is equipped with a proprietary Neural-ART Accelerator™ and enables real-time, high-resolution AI vision processing with a single MCU chip without a separate SoC. This can secure an advantage in the market by reducing component costs and simplifying design. Meanwhile, ST plans to present 'ST Edge AI Solution based on STM32N6' at the ' 2025 e4ds Tech Day' event to be held on September 9th. Accordingly, this magazine arranged an opportunity to hear about the 'STM32N6-based ST Edge AI solution' through an interview with ST Manager Moon Hyun-soo, who was in charge of presenting at this event.


■ A representative application example where you can actually experience the AI processing capabilities of STM32N6 is

The STM32N6 delivers outstanding performance in a variety of vision-based on-device AI applications.

For example, it can be applied to applications that detect human movement or, further, analyze human posture or movement using a Pose Estimation model.

Good examples include scenarios where a person's direction is determined in an indoor space, or where a healthcare device tracks a user's posture in real time to correct exercise posture.



For example, the AI Smart Mirror is a solution implemented by adopting this Pose Estimation model, which applies AI technology to the mirror to estimate a person's pose in real time and provide feedback on it.

The most effective way to perform pose estimation applications is to use convolutional neural networks (CNNs).

The AI smart mirror used one of the models available in the STM32 model zoo.

PartWe used Yolov8n_pose, provided by Ultralytics, which is a state-of-the-art model for pose estimation applications and provides accurate keypoints with high frame rates.

These models require high computational power, and the most suitable MCU option is the STM32N6, which features a Neural-ART accelerator, a neural network processing unit. To easily optimize and deploy this model on the STM32N6, we used scripts provided by the STM32 model zoo and the STM32Cube.AI code optimization tool.

More examples of ST Edge AI applications can be found here. https://bit.ly/4onLiSH

■ What are the strengths of the AI framework or development environment supported by STM32N6, and what are the key functions provided by ST to help developers implement AI models quickly and efficiently?

ST is supporting the STM32N6 AI Ecosystem (STM32N6-AI), a collection of ST tools and resources to support AI model development and deployment on high-performance STM32N6 microcontrollers.

The STM32N6-AI tools are designed to simplify the development process and help developers achieve optimal performance and efficiency.

These tools support both BYOD (bring your own data) and BYOM (bring your own model) approaches, tailored to your best development practices and preferences.

Tools such as the STM32 model zoo (github.com/STMicroelectronics/stm32ai-modelzoo), ST Edge AI Developer Cloud (STEDGEAI-DC), STM32Cube.AI (X-CUBE-AI), and ST Edge AI Core (STEdgeAI-Core) will further expand the possibilities of the STM32N6 for AI and computer vision (CV) applications.

Additionally, the STM32N6-AI provides a variety of software packages that serve as examples and starting points for users' AI projects.

For AI, there are various applications ranging from simple CV and audio applications such as person detection and image classification to pose estimation, instance segmentation, It includes more complex and optimized applications such as hand landmark detection and audio scene recognition.

For video, it provides video streaming applications via USB and Ethernet using an H.264 hardware encoder.

The STM32N6-AI, together with the STM32N6, provides access to a complete ecosystem of software and tools to support building next-generation edge machine learning applications.

The Neural-ART Accelerator NPU built into the STM32N6 efficiently handles AI inference tasks and provides outstanding acceleration for executing neural network models.

This integration makes edge AI on MCUs practical and widespread, providing powerful, efficient, and scalable solutions for a wide range of applications.

The Neural-ART Accelerator is fully supported on ST Edge AI Core, including STM32Cube.AI and ST Edge AI Developer Cloud.

These tools optimize neural network models and generate code to run on NPU hardware.

Analyze neural networks, prepare data, and map operators to appropriate hardware resources to maximize the capabilities of the NPU and achieve optimal AI acceleration.

Streamline your AI development pipeline by seamlessly supporting models from popular AI frameworks such as Keras, TensorFlow™, and ONNX.

All tools and software packages contributing to the STM32N6 AI ecosystem are part of ST's ST Edge AI Suite, a unified set of software tools that facilitate the development and deployment of embedded AI applications.

This comprehensive suite covers machine learning algorithms and implementations from data collection to final hardware deployment. It streamlines the workflow of experts in various fields by supporting the optimization and deployment of neural network models.

The ST Edge AI Suite supports a wide range of ST products, including STM32 microcontrollers and microprocessors, Neural-ART Accelerators, Stellar microcontrollers, and smart sensors.

The ST Edge AI Suite is a strategic move to popularize edge AI technology and is becoming a key resource for developers seeking to efficiently and effectively harness the power of AI in embedded systems.

■ From a developer’s perspective, what business opportunities or competitive advantages do you see when developing on-device AI products using the STM32N6?

The STM32N6 is a competitive product that can implement high-performance AI products with a single MCU, reducing component costs and simplifying design.

Furthermore, by processing AI directly at the edge without relying on cloud computing, data transmission costs can be reduced and user privacy can be strengthened, enabling the development of competitive products in diverse markets such as industrial, smart home, healthcare, and home appliances, where security and real-time performance are crucial.

In this way, the STM32N6 provides developers with both technical flexibility and commercial possibilities, and helps customers accelerate product development and secure competitiveness in the market through an environment where differentiated AI functions can be easily implemented.


Meanwhile, ST will participate in the ' 2025 e4ds Tech Day ' held at the ST Center on September 9th and present the 'STM32N6-based ST Edge AI Solution’ will be presented on the topic. Applications for ' 2025 e4ds Tech Day ' can be made on the official website ( https://www.e4ds.com/conference/techday/ ).