Capture & Detect
Real cards are captured through the camera pipeline and classified in real-time with minimal latency.
PROJECT OVERVIEW
Why Machine Vision Is The Focus
Cardweaver is designed as a machine-vision-driven card game first. Real-time camera detection powered by a Raspberry Pi transforms physical playing cards into digital actions, creating a unique hybrid gaming experience.
The Core Vision
How detection powers gameplay
Reliable Recognition
Multi-attempt voting, confidence thresholds, and fallback handling ensure stable card detection during play.
Game Integration
Detected cards feed directly into game logic, handling state changes and validating gameplay actions in real-time.
Multiple Modes
Tabletop PvP card battles use the same vision pipeline, proving the system is flexible and reusable for real-player match flow.
System Architecture
Building blocks that power the experience
Detection Pipeline
Frame capture → Raspberry Pi YOLO model → Classification → Result resolution for gameplay.
Gameplay Layer
Game states, card validation, multiplayer synchronization, and effects resolution all built on the detection output.
Supporting Systems
Voice chat, networking, UI controls, and themed visuals create a polished, complete experience.
Why This Approach
Project philosophy and design decisions
Unified Focus
Rather than multiple separate projects, Cardweaver demonstrates how a single well-designed vision pipeline can drive multiple game experiences.
Complete Polish
Themes, settings, audio, logging, and debug tools show this is a finished project, not a proof of concept.
Technical Depth
Multiplayer systems, networked state management, and custom detection models show serious engineering beyond the initial prototype.