Evolving Poker Odds Calculations with Hand Analysis (EPOCH)

During my exchange semester I took the course “Deep Neural Networks”, which focused on the foundations of neural networks and deep learning. The course covered the basics of feed-forward-, convolutional-, recurrent-, transformer-, generative adversarial networks and autoencoders.

For the final project of the course two class mates and I evaluated multiple model architectures with the goal of being able to identify poker cards in a given image. Three architectures were analyzed for this purpose: You Only Look Once (YOLO), Faster Region-based Convolutional Neural Network (Faster R-CNN), and Detection Transformer (DETR). My focus was primarly on implementing, training and evaluating the Faster R-CNN model.

Read the full report here