A classroom-focused AI Cell Recognition project uses the HUSKYLENS 2 AI vision sensor and UNIHIKER K10 board to enable real-time cell identification and classification under a microscope, showcasing a complete AI workflow from data collection to edge inference for STEAM learning.

The HUSKYLENS 2, powered by K230, can run both pre-trained and user-trained models with up to 6 TOPS AI performance and low latency, making it ideal for hands-on AI concepts in middle school biology.

The demonstration emphasizes a full AI workflow for education—data collection, model training, and edge inference—to help students grasp AI concepts in biology contexts.

Xia Qing, a Senior Engineer at DFRobot, explains that TinyML sensing and local AI models can transform sensor data into actionable insights with potential applications in coffee flavor analysis, fermentation monitoring, and food freshness detection.

The Electronic Nose project combines edge AI and embedded hardware, using four MEMS gas sensors with an ESP32 running a TinyML model for real-time odor analysis and on-device processing with no network dependency.

In a live demonstration, odor sampling took 20 to 30 seconds, illustrating the practical speed of edge AI for applications like coffee flavor, fermentation monitoring, and freshness detection.

The system provides end-to-end on-device AI for sensor data, highlighting potential uses in flavor analysis, fermentation monitoring, and freshness detection.

DFRobot is collaborating with DigiKey to promote open-source hardware and AI education, with plans to expand technical content, global marketing, and classroom-ready solutions to reduce AI adoption barriers.

DFRobot showcased open-hardware, AI-enabled projects at the DigiKey booth during the Robot Hokoten event in Akihabara, Tokyo, highlighting STEAM education and maker applications.

The initiative aims to lower barriers to AI and open-hardware adoption and accelerate the move from maker projects to practical STEAM classroom applications.