Quality-driven ultrasound-based visual servoing

Contact: Pierre Chatelain
Last update: May 2016

Context

This project aims at optimizing the quality of robotic ultrasound acquisitions. Ultrasound image quality is highly dependent on the position of the probe and the contact forces with the patient's body. During manual exams, sonographers constantly adjust the orientation of the ultrasound probe in order to find the optimal acoustic window for the anatomy of interest. Reproducing such behavior for a robotic system would allow a finer control of the image contents, and result in higher quality acquisistions.

Method overview

We use confidence maps to estimate the quality of the ultrasound signal within the images. Using this information, we design a confidence-based control law which aims at optimizing the orientation of the probe. Two different approaches are considered:

• A global confidence control: the image quality is optimized globally, without caring about the image contents. This approach is intended to be used in a teleoperation scenario, where the control is shared between the automatic controller and a human operator. In this situation, the human operator steers the probe towards the desired location, and the orientation of the probe is adapted automatically in order to optimize the image quality.
• A target-specific confidence control: an anatomic target (or region of interest) is tracked in the ultrasound stream, and the aim is to find the optimal acoustic window for this specific target.

Experimental results

References

1. P. Chatelain, A. Krupa, N. Navab. Optimization of ultrasound image quality via visual servoing. In IEEE Int. Conf. on Robotics and Automation, ICRA'15, Pages 5997-6002, Seattle, WA, May 2015.
2. P. Chatelain, A. Krupa, N. Navab. Confidence-driven control of an ultrasound probe: target-specific acoustic window optimization. In IEEE Int. Conf. on Robotics and Automation, ICRA'16, Stockholm, Sweden, May 2016.