Monocular 3D Probe Tracking for Generating Sub-Surface Optical Property Maps From Diffuse Optical Spectroscopic Imaging.

OBJECTIVE: Diffuse optical spectroscopic imaging (DOSI) is a promising biophotonic technology for clinical tissue assessment, but is currently hampered by difficult wide area assessment. A co-integrative optical imaging system is proposed for dense sub-surface optical property spatial assessment. METHODS: The proposed system fuses a co-aligned set of camera frames and diffuse optical spectroscopy measurements to generate spatial sub-surface optical property maps. A 3D rigid body motion estimation model was developed by fitting automatically detected target features to an a priori geometric model using a single overhead camera. Point-wise optical properties were measured across the tissue using frequency domain photon migration DOSI. The 3D probe trajectory and temporal optical property data were fused to generate 2D spatial optical property maps, which were projected onto the tissue image using pre-calibrated camera parameters. RESULTS: The system demonstrated sub-millimeter positional accuracy (error 0.24 ± 0.35 mm) across different probe speeds (1.0-3.8 cm/s), and displacement accuracy in overhead ([Formula: see text] mm) and tilted (0.51 ± 0.51 mm) camera orientations. Unstructured scans on a tumor inclusion phantom showed strong contrast under different probe paths, and significant ( ) changes in optical properties in an in vivo leg cuff occlusion protocol with spatial anatomy localization. CONCLUSION: The proposed co-integrative optical imaging system generated dense sub-surface optical property distributions across wide tissue areas with sub-millimeter accuracy at different probe speeds and trajectories, and does not require pre-planned probe route for tissue assessment. SIGNIFICANCE: This system provides a valuable tool for real-time non-invasive tissue health and cancer screening, and enables longitudinal disease progression assessment through unstructured probe-based optical tissue assessment.