Recording dynamic facial micro-expressions with a multi-focus camera array

Area of Science:

• Biomedical Imaging
• Optical Engineering

Background:

• Traditional single-camera systems face limitations in capturing high-resolution, dynamic, non-planar surfaces like the human face.
• Achieving both high resolution and extended depth-of-field (DOF) simultaneously presents a significant challenge in optical imaging.

Purpose of the Study:

• To introduce a novel multi-camera array system for capturing dynamic, high-resolution videos of the human face.
• To overcome the inherent resolution-DOF trade-off in conventional camera designs.

Main Methods:

• Utilized a 54-camera array, with each camera focusing on a distinct object plane (multi-focus strategy).
• Employed image stitching techniques to create high-resolution composite video frames (709 megapixels total).
• Quantified lateral resolution and composite depth-of-field across the entire facial surface.

Main Results:

• Achieved a lateral resolution of 26.14 ± 5.8 µm across a composite depth-of-field of approximately 43 mm, covering over 85 cm².
• Demonstrated a nearly 10-fold increase in effective depth-of-field compared to single-focus configurations.
• Successfully captured dynamic facial expressions at microscopic resolution.

Conclusions:

• The multi-camera, multi-focus array effectively captures high-resolution dynamic facial data, overcoming traditional imaging limitations.
• This technology holds significant potential for various biomedical applications requiring detailed facial analysis.
• The system offers a substantial advancement in imaging curved, dynamic biological surfaces.