Purpose Recovering tissue deformation during robotic-assisted minimally invasive
surgery (MIS) procedures is important for providing intra-operative guidance,
enabling in vivo imaging modalities and enhanced robotic control. The tissue motion
can also be used to apply motion stabilization and to prescribe dynamic
constraints for avoiding critical anatomical structures.
Methods Image-based methods based independently on salient features or on
image intensity have limitations when dealing with homogeneous soft-tissues or
complex reflectance. In this paper, we use a triangular geometric mesh model in
order to combine the advantages of both feature and intensity information and
track the tissue surface reliably and robustly.
Results Synthetic and in vivo experiments are performed to provide quantitative
analysis of the tracking accuracy of our method, we also show exemplar results for
registering multispectral images where there is only a weak image signal.
Conclusions Compared to traditional methods, our hybrid tracking method is
more robust and has improved convergence in the presence of larger displacements,
tissue dynamics and illumination changes.