eduzhai > Applied Sciences > Engineering >

Needle tip force estimation by deep learning from raw spectral OCT data

  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: Purpose. Needle placement is a challenging problem for applications such asbiopsy or brachytherapy. Tip force sensing can provide valuable feedback forneedle navigation inside the tissue. For this purpose, fiber-optical sensorscan be directly integrated into the needle tip. Optical coherence tomography(OCT) can be used to image tissue. Here, we study how to calibrate OCT to senseforces, e.g. during robotic needle placement.Methods. We investigate whether using raw spectral OCT data without a typicalimage reconstruction can improve a deep learning-based calibration betweenoptical signal and forces. For this purpose, we consider three differentneedles with a new, more robust design which are calibrated using convolutionalneural networks (CNNs). We compare training the CNNs with the raw OCT signaland the reconstructed depth profiles.Results. We find that using raw data as an input for the largest CNN modeloutperforms the use of reconstructed data with a mean absolute error of 5.81 mNcompared to 8.04 mN.Conclusions. We find that deep learning with raw spectral OCT data canimprove learning for the task of force estimation. Our needle design andcalibration approach constitute a very accurate fiber-optical sensor formeasuring forces at the needle tip.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×