Fast Localization and Segmentation of Tissue Abnormalities by Autonomous Robotic Palpation

Robot-assisted minimally invasive surgery (RMIS) has become increasingly popular in the resection of cancers. However, the lack of tactile feedback in clinical RMIS limits the surgeon's haptic understanding of tissue mechanics, making it hard to detect tissue abnormalities (e.g., tumor) efficiently. In this letter, we propose an approach that can simultaneously localize and segment the hard inclusions (artificial tumor) in artificial tissue via autonomous robotic palpation with a tactile sensor. By using Bayesian optimization guided probing, the tumor can be quickly localized within 30 iterations of the algorithm. And by continuously sliding the sensor over the tissue surface, the boundary of the tumor can be precisely segmented from the surrounding soft tissue with a high sensitivity up to 0.999 and specificity up to 0.973. Moreover, the tumor depth can be estimated with Gaussian Process (GP) regression with the root mean squared error (RMSE) being only around 0.1 mm. Our method is proven to be robust and efficient in both simulation and experiments, which provides new insight into fast tissue abnormalities detection during RMIS and could be beneficial to relevant surgical tasks like tumor removal.