A video-based method for supporting automated facial mask customization

Mass customization is the concept of integrating advanced designing and fabrication approaches to produce personalized products while ensuring efficiency and low cost over the entire customization cycle. The escalation of COVID-19 in the past few years has stimulated massive demand for personal protective equipment (PPE), including surgery masks, oxygen respirators, and other wearable equipment that could prevent the virus from spreading. Both performance and comfort are affected by the fit of facial masks; thus, implementing mass customization on PPE could improve user comfort and enhance air impermeability while maintaining low costs during design and fabrication. Current mass customization methods for surgery masks, oxygen respirators, or other general purpose medical masks mostly use a 3D scanner or integrated smartphone applications to acquire the patient’s facial data profile. However, these methods are manual and either require a high-precision scanner or cameras with the depth mapping function. To fill these gaps, this paper proposed a video-based method for supporting automated facial mask design customization. More specifically, the proposed method records a short facial video as the primary input, where facial data is extracted from each video frame and forms a cross-validated dataset with facial variations. The dataset further supports the generative design of customized facial masks based on defined regions of interest. Finally, design considerations of additive manufacturing are integrated into the fabrication of customized masks. Preliminary user feedback has shown promising results for the proposed method.