Opportunity
BMK (Atea Group), an established IT solutions provider based in Vilnius, Lithuania, sought to enhance an innovative educational board game designed to teach children letters, numbers, and vocabulary interactively. The game featured physical symbol pieces tracked by a camera and visuals projected directly onto the board. To achieve a seamless and captivating experience, BMK partnered with 1D.works to implement advanced AI-driven computer vision.
Challenge
- Real-Time Recognition: Identifying and locating a broad set of game symbols quickly enough for interactive play.
- Resource Efficiency: Achieving robust recognition without significant computational overhead.
- Scalable Adaptability: Ensuring easy integration of new symbols, languages, and game variations without extensive re-training or labeling.
Solution
1D.works developed an AI-driven symbol recognition algorithm:
- Reference-Based Recognition: Designed to identify symbols from a single reference image per symbol, reducing the initial labelling requirement.
- Resource-Light Execution: Optimized for fast, resource-efficient performance suitable for real-time gameplay.
- Flexible Expansion: Easily adaptable to future symbol sets, languages, and diverse educational contexts.
Delivered Capabilities
- Interactive Recognition: Connected camera-based symbol detection to the projected game experience.
- Lower Labelling Requirement: Used reference-based recognition instead of a large labelled training set for each symbol.
- Expandable Symbol Set: Provided a technical path for adding symbols, languages, and game variants.
- Resource-Conscious Design: Optimised the recognition workflow for the hardware used by the game.
The project delivered the computer-vision component for an interactive educational game. Claims about learning outcomes, market performance, or quantified development savings require separate evidence.
Explore our Rapid Prototyping & Validation service. Read: Applying AI in the Real-World.