Development and validation of an affordable freehand scanning 3D ultrasound reconstruction system with visual positioning.

Abstract

OBJECTIVE: Three-dimensional (3D) ultrasound provides more intuitive, detailed, and comprehensive diagnostic information. The primary objective of this study was to develop affordable, lightweight, and wide-field 3D ultrasound probes which is a critical challenge for expanding the accessibility and utility of this technology. METHODS: We describe an affordable freehand scanning 3D ultrasound reconstruction system. The system utilizes a QR code film attached to the skin surface for positioning. During each ultrasound scan, a lightweight camera fixed on the probe records the QR codes within the scanning area. Based on QR code features, a skin map is constructed, and ultrasound images and film image information are preprocessed. The calibration relationship between the ultrasound probe and camera field of view is determined. Finally, comprehensive algorithms such as ORB-SLAM positioning are used for three-dimensional image reconstruction. RESULTS: Through hand-scanning tests, the results show that the scanning speed is 15 mm/s, and when the image acquisition frame rates are 10 Fps/s, 20 Fps/s, and 30 Fps/s, the maximum cumulative error of ORB-SLAM algorithm positioning does not exceed 2%, meeting the positioning requirements of the intelligent skin texture image reconstruction system for thyroid ultrasound. Additionally, the effectiveness of three-dimensional reconstruction is verified through model volume testing. CONCLUSION: The intelligent skin texture image reconstruction system for ultrasound can achieve 3D image reconstruction, displaying lesion features and the spatial positions of surrounding tissues in three-dimensional space, thereby assisting ultrasound doctors in making accurate and rapid diagnoses. Clinical trial number: not applicable.