Method and system for image analysis
Method and system for image registration or image segmentation. The method includes receiving an image which is to be processed by a first machine-learning model to perform, for example, image registration or segmentation, and using a second machine-learning model to determine if the received image is of a quality suitable for the first machine-learning model to act upon.
1. A method for processing an image, the method comprising:
receiving an image to be processed by a first machine-learning model, wherein the image comprises metadata;
estimating, using a second machine-learning model, numerical physical information from the received image;
extracting, using the second machine-learning model, numerical physical information from the metadata of the received image;
determining whether differences between a value of the estimated numerical physical information and a value of the extracted numerical physical information are below a threshold; and
applying, when the differences between the value of the estimated numerical physical information and the value of the extracted numerical physical information are below the threshold, the first machine-learning model to the received image.
2. The method as claimed in claim 1 , further comprising:
obtaining, using the first machine-learning model, image features from the received image,
wherein the step of estimating numerical physical information from the received image comprises using the obtained image features to estimate the numerical physical information.
3. The method as claimed in claim 1 , wherein the first machine-learning model and the second machine-learning model have been trained on the same training data.
4. The method as claimed in claim 1 , wherein the threshold depends on an imaging technique used to acquire the image.
5. The method as claimed in claim 1 , wherein when the differences between the value of the estimated numerical physical information and the value of the extracted numerical physical information exceed the threshold, the method further comprises:
outputting data indicating that the received image is not suitable for the first machine-learning model to analyze.
6. The method as claimed in claim 5 , wherein the step of outputting data comprises outputting a modified image which is a version of the received image that is modified to indicate at least one abnormality in the received image which caused the differences to exceed the threshold.
7. The method as claimed in claim 6 , wherein the modified image comprises a heatmap to indicate a probability of the at least one abnormality in the received image causing the differences to exceed the threshold.
8. The method as claimed in claim 6 , wherein the at least one abnormality is any one or more of: an image artefact, a pathology, a guide wire, and a pace maker.
9. The method as claimed in claim 1 , wherein the estimated numerical physical information and extracted numerical physical information comprises any one or more of: an operating parameter associated with the imaging device, an angle of the imaging device relative to an object being imaged, and a distance of the imaging device from an object being imaged.
10. The method as claimed in claim 1 , wherein the first machine-learning model is applied to the received image to perform image registration.
11. The method as claimed in claim 1 , wherein the first machine-learning model is applied to the received image to perform image segmentation.
12. The method as claimed in claim 1 , wherein the received image is any one of: an X-ray image, a magnetic resonance image, an ultrasound image, and a computerized tomography scan.
13. The method as claimed in claim 1 , wherein the first machine-learning model and/or the second machine-learning model comprises a convolutional neural network.
14. A non-transitory computer readable medium having processor control code which, when implemented in a system, causes the system to carry out the method of claim 1 .
15. An image processing system, comprising:
an image capture device which is configured to capture an image;
an image processor which is configured to receive an image from the image capture device and carry out a method for processing the image, the method comprising:
receiving the image by a first machine-learning model, wherein the image comprises metadata;
estimating, using a second machine-learning model, numerical physical information from the received image;
extracting, using the second machine-learning model, numerical physical information from the metadata of the received image;
determining whether differences between a value of the estimated numerical physical information and a value of the extracted numerical physical information are below a threshold; and
applying, when the differences between the value of the estimated numerical physical information and the value of the extracted numerical physical information are below the threshold, the first machine-learning model to the received image; and
a user interface which is configured to display an output result generated by the image processor.