PHOTO SCORING — MODEL CREDITS

DeepDarts D2 (CVSports 2021)
William McNally, Kanav Narasimhan, Srinivasu Guttikonda, Alexander Yampolsky,
John McPhee and Alexander Wong.
DeepDarts: Modeling Keypoints as Objects for Automatic Scorekeeping in Darts
using a Single Camera.
Paper: https://arxiv.org/abs/2105.09880
Original project: https://github.com/wmcnally/deep-darts
Published model and dataset: https://ieee-dataport.org/open-access/deepdarts-dataset
License: Creative Commons Attribution 4.0 International (CC BY 4.0)
https://creativecommons.org/licenses/by/4.0/

TFLite export supplied by iambhabha/Dart-Vision:
https://github.com/iambhabha/Dart-Vision/tree/ec4f4daf264083a8fd4dda404d2bf17a7cfd4fb6
Source file: deepdarts_d2/model.tflite

Changes for Dart Night: converted the floating-point TFLite graph to ONNX for
browser inference. No retraining. Independent JavaScript post-processing,
perspective mapping, review controls and scoring integration.
Recognition may be wrong; every visit requires player review and confirmation.

Runtime: Microsoft ONNX Runtime Web 1.23.2 (MIT).
Runtime license and third-party notices are in ../photo-runtime/.