Registration method and apparatus
An apparatus comprises processing circuitry configured to receive first image data; receive second medical image data; and apply a transformation regressor to perform a registration process to obtain a predicted displacement that is representative of a transformation between the first image data and the second image data; wherein the transformation regressor is trained in combination with a discriminator in an adversarial fashion by repeatedly alternating a transformation regressor training process in which the transformation regressor is trained to predict displacements, and a discriminator training process in which the discriminator is trained to distinguish between predetermined displacements and displacements predicted by the transformation regressor.
1. An apparatus comprising processing circuitry configured to:
receive first image data;
receive second image data; and
apply a transformation regressor to perform a registration process to obtain a predicted displacement that is representative of a transformation between the first image data and the second image data;
wherein the transformation regressor is trained in combination with a discriminator in an adversarial fashion by repeatedly alternating a transformation regressor training process in which the transformation regressor is trained to predict displacements, and a discriminator training process in which the discriminator is trained to distinguish between predetermined displacements and displacements predicted by the transformation regressor.
2. An apparatus according to claim 1 , wherein the predicted displacement comprises a predicted displacement field, and the predetermined displacements comprise predetermined displacement fields.
3. An apparatus according to claim 2 , wherein the predicted displacement field comprises a non-rigid displacement field.
4. An apparatus according to claim 1 , wherein at least one of the transformation regressor and the discriminator comprises a deep neural network.
5. An apparatus according to claim 1 , wherein the processing circuitry is further configured to use the predicted displacement in a further process, the further process comprising at least one of a further registration, a subtraction, a segmentation, an atlas-based process, an image fusion, an anatomy detection, a pathology detection.
6. An apparatus according to Claim 1 , wherein the transformation is a non-parametric transformation.
7. An apparatus according to Claim 1 , wherein the transformation is a parametric transformation.
8. A method comprising:
receiving first image data;
receiving second image data; and
applying a transformation regressor to perform a registration process to obtain a predicted displacement that is representative of a transformation between the first medical image data and the second medical image data;
wherein the transformation regressor is trained in combination with a discriminator in an adversarial fashion by repeatedly alternating a transformation regressor training process in which the transformation regressor is trained to predict displacements, and a discriminator training process in which the discriminator is trained to distinguish between predetermined displacements and displacements predicted by the transformation regressor.