UCL Discovery

Abstract

Recovering the three dimensional (3D) surface shape of tissues in minimally invasive surgery (MIS) is important for developing advanced image-guidance and navigation systems. Passive techniques for 3D reconstruction based on computational stereo are limited by the saliency of tissue texture and the view-dependent reflectance characteristics of the scene. Structured lighting provides a viable alternative by projecting known features onto the tissue surface. However, the correspondence problem (distinguishing individual projected features computationally) becomes difficult in tissue due to the presence of occlusions. Furthermore, miniaturisation of a light projection system for use in MIS, while maintaining the required light intensity, is a significant challenge.In this paper, a fibre-based probe is described that projects a spectrally-encoded pattern onto the target surface from its distal end. A dispersed broadband light source is used to project features of varying spectral content. The dominant wavelengths of imaged spots may be deduced from the RGB values of a standard colour camera using an algorithm that locates each colour on a chromaticity diagram. The results show that individual spots of a specified wavelength may be segmented and their centres of mass calculated, despite varying background colour. The probe has also been demonstrated on ex vivo tissue.