Sensorless ultrasound probe 6DoF pose estimation through the use of CNNs on image data

Ultrasound probe pose estimation has many applications in medical practice and research. Currently, ultrasound probe pose estimation with respect to the human body requires the use of sensors attached to the ultrasound probe, and may get computationally costly. We explore the use of Convolutional Neural Networks (CNNs) to provide sensorless pose estimation. The Ultrasound CNN model proposed in this paper learns to regress the six degree of freedom (6-DoF) camera pose from a single ultrasound image in an end-to-end manner. Ultrasound images are easier to obtain than other forms of medical imaging, but suffer from poor quality, which will be a challenge for the Ultrasound CNN model. The most promising model from our experiments is a 23 layer deep CNN based off of GoogLeNet. In previous literature, CNNs have demonstrated that they can be used to solve complicated out of image plane regression problems. We show how the proposed method can regress the 6DoF pose within a certain degree of accuracy.

Department dc:contributor.department

Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science.

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2018

Author dc:creator

• Xue, Elise Yuan

Advisor dc:contributor.advisor

• Brian W. Anthony.

• Electrical Engineering and Computer Science.

Statement dc:rights

• MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission.

Licence dc:rights.uri

• http://dspace.mit.edu/handle/1721.1/7582

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/1721.1/119697

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/119697