Customize my helmet: A novel algorithmic approach based on 3D head prediction

Anthropometric knowledge of the 3D head forms is essential for the individual headgear customization. In 3D complete-head scanning, the face region is easily captured, but the scalp region is usually missed, because of the low hair transmissivity. Most existing methods use a tight cap to address this problem, which is inconvenient and uncomfortable for participants. Therefore, instead of the physical approaches, we developed an algorithmic approach to predict the 3D head region from incomplete scans for efficient individual-helmet customization. To achieve this, two powerful statistical shape models (SSM) of human faces and heads were constructed from a large-scale parameterized complete head database. Then, a face-to-head model regression matrix was computed to obtain head model coefficients from face model coefficients, thereby generating the head from an incomplete scan. The experimental results showed that the average (± standard deviation) helmet fitness indexes (HFI) of our individual customization and conventional mass customization were 76.84 (± 5.44) and 71.04 (± 7.93) respectively. The results suggest that our method can improve the HFI for the consumers efficiently and significantly.