rwitz/pinball-score-detection-yolov11S

YOLOv11x Model for Pinball Score Detection

GUIDE

Remember to use 1280 px for image resolution and 0.45 for IOU :)

Overview

This repository contains a YOLOv11x model trained to detect pinball scores in images. The model aims to facilitate the segmentation of pinball scores for Optical Character Recognition (OCR) tasks.

Dataset

• Images: 1,100 hand-annotated images.
• Annotations: Bounding boxes around pinball scores.

Training Details

The model was trained over 100 epochs using an A100 GPU with 40GB VRAM.

Performance Metrics at Final Epoch

Validation Metrics

                 Class     Images  Instances      P          R      mAP50   mAP50-95
100%|██████████| 7/7 [00:01<00:00,  4.28it/s]
                   all        152        187    0.977      0.989      0.99      0.737

• Precision (P): 0.977
• Recall (R): 0.989
• Mean Average Precision at IoU=0.50 (mAP50): 0.99
• Mean Average Precision at IoU=0.50:0.95 (mAP50-95): 0.737

Purpose

The primary goal of this model is to detect pinball scores in any image to enable segmentation for OCR processing.