Multi-Camera Self-Calibration in Sports Motion Capture: Leveraging Human and Stick Poses

TL;DR

This paper presents a novel, efficient multi-camera self-calibration method for sports involving sticks, using synchronized videos of humans and sticks to accurately determine camera positions without calibration tools.

Contribution

It introduces a three-stage optimization pipeline leveraging human and stick cues, and provides the first dataset for benchmarking such calibration in stick-based sports.

Findings

Achieves state-of-the-art calibration accuracy with low rotation and translation errors.

Effectively refines camera extrinsics using joint human and stick trajectory reconstruction.

Demonstrates robustness across synthetic sequences in multiple sports categories.

Abstract

Multi-camera systems are widely employed in sports to capture the 3D motion of athletes and equipment, yet calibrating their extrinsic parameters remains costly and labor-intensive. We introduce an efficient, tool-free method for multi-camera extrinsic calibration tailored to sports involving stick-like implements (e.g., golf clubs, bats, hockey sticks). Our approach jointly exploits two complementary cues from synchronized multi-camera videos: (i) human body keypoints with unknown metric scale and (ii) a rigid stick-like implement of known length. We formulate a three-stage optimization pipeline that refines camera extrinsics, reconstructs human and stick trajectories, and resolves global scale via the stick-length constraint. Our method achieves accurate extrinsic calibration without dedicated calibration tools. To benchmark this task, we present the first dataset for multi-camera self-calibration in stick-based sports, consisting of synthetic sequences across four sports categories with 3 to 10 cameras. Comprehensive experiments demonstrate that our method delivers SOTA performance, achieving low rotation and translation errors. Our project page: https://fandulu.github.io/sport_stick_multi_cam_calib/.