Ball 3D Localization From A Single Calibrated Image

TL;DR

This paper introduces a novel method for 3D ball localization in team sports using only a single calibrated image, leveraging neural networks to estimate ball diameter and confidence, enabling accurate 3D localization without multi-view setups.

Contribution

It presents the first model for 3D ball localization from a single image, along with a new annotation method and a high-quality evaluation dataset.

Findings

Model achieves accurate 3D localization in basketball datasets.

Confidence output improves detection rate by filtering false positives.

Method works in various game situations with partly visible balls.

Abstract

Ball 3D localization in team sports has various applications including automatic offside detection in soccer, or shot release localization in basketball. Today, this task is either resolved by using expensive multi-views setups, or by restricting the analysis to ballistic trajectories. In this work, we propose to address the task on a single image from a calibrated monocular camera by estimating ball diameter in pixels and use the knowledge of real ball diameter in meters. This approach is suitable for any game situation where the ball is (even partly) visible. To achieve this, we use a small neural network trained on image patches around candidates generated by a conventional ball detector. Besides predicting ball diameter, our network outputs the confidence of having a ball in the image patch. Validations on 3 basketball datasets reveals that our model gives remarkable predictions on ball 3D localization. In addition, through its confidence output, our model improves the detection rate by filtering the candidates produced by the detector. The contributions of this work are (i) the first model to address 3D ball localization on a single image, (ii) an effective method for ball 3D annotation from single calibrated images, (iii) a high quality 3D ball evaluation dataset annotated from a single viewpoint. In addition, the code to reproduce this research is be made freely available at https://github.com/gabriel-vanzandycke/deepsport.