Apparatus and method for determining motion of an ultrasound probe

1. A method of determining a three-dimensional motion of a movable ultrasound probe during acquisition of an ultrasound image of a volume portion by the ultrasound probe, the method comprising:

Receiving a stream of ultrasound image data from the ultrasound probe while the ultrasound probe is moved along the volume portion;

Inputting at least a sub-set of the ultrasound image data representing a plurality of ultrasound image frames into a machine-learning module, wherein the machine learning module has been trained to determine the relative three-dimensional motion between ultrasound image frames; and

Inputting further sensor data into the machine-learning module, wherein the further sensor data is synchronized with the ultrasound image data, and wherein the further sensor data includes at least one of position data, obtained by a tracking system tracking a position of the ultrasound probe, acceleration data representing the acceleration corresponding to the at least two ultrasound image frames, the acceleration being detected by an acceleration sensor attached to the ultrasound probe and gyroscope data; and

Determining, by the machine-learning module, a three-dimensional motion indicator indicating the relative three-dimensional motion between the ultrasound image frames.

2. The method according to claim 1 , further comprising pre-processing the ultrasound image data, the pre-processing including at least one of an image filtering, image resampling and image segmentation.

3. The method according to claim 1 , wherein the machine learning module comprises a neural network, preferably a convolutional neural network.

4. The method according to claim 1 , wherein

the step of inputting the at least sub-set of the ultrasound image data includes inputting local image data corresponding to a pair of ultrasound image frames to the machine learning module, and wherein

the three-dimensional motion indicator indicates the relative three-dimensional motion between the pair of ultrasound image frames, and wherein

the inputting and determining steps are repeated for consecutive pairs or subsets of image frames.

5. The method according to claim 1 , wherein

the step of inputting the at least sub-set of the ultrasound image data includes inputting a global set of image data substantially spanning the whole set of ultrasound image frames to the machine learning module, and wherein

the three-dimensional motion indicator indicates the relative three-dimensional motion for determining the relative three-dimensional motion of each of the ultrasound image frames with respect to a first one of the ultrasound image frames.

6. The method according to claim 1 , wherein the ultrasound image data includes at least one of A-Mode data, B-Mode data, continuous harmonic imaging data, Doppler data, plain wave imaging data, and raw radio frequency data.

7. The method according to claim 1 , further comprising determining, from the three-dimensional motion indicator, a probe position and orientation of the ultrasound probe for each image frame.

8. The method according to claim 7 , further comprising tracking the position of the movable ultrasound probe by a further tracking system thereby generating a tracked position information, detecting whether the tracking system fails, and if the tracking system is determined to fail, substituting the tracked position information by the probe position and orientation determined from the three-dimensional motion indicator.

9. The method according to claim 7 , further comprising reconstructing a three-dimensional ultrasound image using the stream of ultrasound image data and the probe position and orientation determined from the three-dimensional motion indicator.

10. The method according to claim 1 , wherein the method comprises directly predicting the ultrasound probe motion from the stream of ultrasound images using the three-dimensional motion indicator, without using a further tracking system.

11. The method according to claim 1 , further comprising detecting an inconsistency between the determined three-dimensional motion indicator and the sensor data.

12. The method according to claim 1 , wherein the further sensor data is obtained from an IMU sensor.

13. An apparatus for determining a three-dimensional motion of a movable ultrasound probe during acquisition of an ultrasound image of a volume portion by the ultrasound probe, the apparatus comprising:

a probe input interface for receiving a stream of ultrasound image data from the ultrasound probe while the ultrasound probe is moved along the volume portion; and

a machine-learning module having

(a) an input section adapted for receiving, as an input, at least a sub-set of the ultrasound image data representing a plurality of ultrasound image frames, wherein the input section is characterized in that the input section is adapted for further receiving, as an input, sensor data, wherein the sensor data is synchronized with the ultrasound image data and wherein the sensor data includes at least one of position data, obtained by a tracking system tracking a position of the ultrasound probe, acceleration data representing the acceleration corresponding to the at least two ultrasound image frames, the acceleration being detected by an acceleration sensor attached to the ultrasound probe and gyroscope data,

(b) a training memory section containing a training memory having been trained to determine the relative three-dimensional motion between ultrasound image frames, wherein

the machine-learning module is adapted for determining, from the input and using the training memory, a three-dimensional motion indicator indicating the relative three-dimensional motion between the ultrasound image frames.