Shape Reconstruction and Attitude Estimation of Bevel-Tip Needle via CT-Guidance

The bevel-tip flexible needle is an improvement of the traditional rigid needle and has the potential to reduce traumas, improve the controllability and reach multi-targets in one insertion. The needle steering robot system has the ability to deal with complex puncture environment and to improve the puncture precision, but it requires a precise feedback of needle tip's position and attitude. Computerized tomography (CT) is a widely available imaging system that provides precise and clear images. Unfortunately, CT is not a real-time system and it is impossible to fetch the attitude of bevel-tip needle from CT image. A new method is proposed to estimate the attitude, in which the unscented kalman filter is employed. We extract the positions of the needle in CT images, reconstruct the needle shape, calculate the control quantity according to the insert distance and then estimate the attitude of the needle tip. Simulations and experiments demonstrate that the method proposed is effective. Implementation of shape reconstruction and attitude estimation of the bevel-tip needle in biological tissue is a significant step toward the precise control of bevel-tip needles.