3D Scanning
The document discusses “CHECKER,” a real-time scan quality checking system designed to enhance the user experience in mobile 3D body scanning. Developed by Size Stream, the system uses machine learning to identify and correct scan errors during the scanning process including obscured body parts, improper attire, and pose errors.
“Checker” aids in the delivery of improved pose guidance to scan subjects, enabling more error-free scans, greater clarity on how to perform a successful scan, and the acquisition of better data and a superior scan experience for users.
Key takeaway
Scan accuracy in real-world, unsupervised settings depends not just on the underlying technology, but on the user completing the scan correctly, and Size Stream’s CHECKER system addresses this directly by using machine learning to detect common errors like baggy clothing and improper positioning in real time, guiding users to correct them before the scan is complete. Trained on over 10,000 real-world scans and achieving precision rates of up to 99%, CHECKER effectively makes the at-home scanning experience self-correcting. This is a critical step in scaling mobile 3D body measurement beyond supervised clinical environments, lowering the barrier to accurate data collection for apparel sizing, fitness tracking, and health monitoring alike.