Real-time poker table recognition system

Desktop C# app for real-time poker table recognition (multi-game compatible) using YOLO, PaddlePaddle, RapidOCR, WMO, and OpenCV.
The system detects and recognizes both graphic (via YOLO) and text (via PP, RapidOCR, and WMO) in real time from the captured screen, achieving high performance with minimal resource use.
Implemented on-the-fly data analysis to reconstruct the table structure and track dynamic game states in memory, generating situation-aware analytical outputs, serializing and transmitting via socket to external analytical software for strategy computation and visualization.