AI Smart Drawing Board based on ESP32-S2

Client

An educational technology company creating a digital drawing tablet for children that provides real-time feedback on handwriting and drawing skills.

Challenge

Achieving low-latency handwriting recognition and shape classification on-device for children aged 3-8, supporting both free drawing and structured tracing exercises.

Product Concept & Architecture

The ESP32-S2 with USB OTG reads a Wacom EMR digitizer (4096 pressure levels) over SPI. A 10.5-inch monochrome LCD provides a paper-like drawing experience.

Engineering Solution

We implemented a real-time stroke processing pipeline: raw pen coordinates (200Hz) → Douglas-Peucker simplification → stroke classification (TinyML CNN, 30KB). In trace mode, the system compares the child's stroke to the template using Dynamic Time Warping and provides visual correction cues.

Result

<30ms pen-to-display latency. 94% shape recognition accuracy. Used in 200+ occupational therapy practices.