Multi-phase filter

1. An apparatus comprising:

processing circuitry configured to:

acquire a plurality of sets of medical imaging data of a region of a subject, each set of data corresponding to a respective different measurement period; and

apply a filter to the plurality of medical imaging data sets to produce a plurality of filtered medical imaging data sets corresponding to the different measurement periods, the filter being applied to each medical imaging data set to produce a corresponding filtered medical imaging data set for each measurement period, wherein

applying of the filter is such that, for each of the medical imaging data sets, the filtering uses at least some information from the other medical imaging data sets acquired at the different measurement periods, and

applying of the filter comprises applying at least one constraint or condition and the constraint or condition comprises preserving at least one measure of intensity for each medical imaging data set for each respective measurement period.

2. The apparatus according to claim 1 , wherein the medical imaging data set comprises CT scan data and wherein the at least one measure of intensity comprises at least one measure of density.

3. The apparatus according to claim 1 , wherein the at least some information from the other medical imaging data sets used during the filtering comprises gradient information and/or edge information.

4. The apparatus according to claim 1 , wherein the filter comprises a four dimensional filter wherein the four dimensions comprise three spatial dimensions and a time dimension and/or wherein the filter comprises an anisotropic filter.

5. The apparatus according to claim 1 , wherein the filter comprises one or more trained multi-layer neural networks or other trained model.

6. The apparatus according to claim 1 , wherein the processing circuitry is further configured to perform a registration procedure to spatially align the plurality of sets of medical imaging data and wherein the filter is applied to the spatially aligned medical imaging data sets.

7. The apparatus according to claim 1 , wherein the processing circuitry is further configured to obtain combined data from at least some of the plurality of data sets and wherein applying the filter to each of the plurality of data sets comprises using the combined data and/or information obtained from the combined data.

8. The apparatus according to claim 7 , wherein obtaining the combined data comprises performing a combining process on the plurality of medical imaging data sets, wherein the combining process is characterised by one or more weighting parameters.

9. The apparatus according to claim 8 , wherein performing the combining process comprises performing at least one of: an averaging process; a summing process; at least one further filtering step; a projection from a higher dimensional representation to a lower dimensional representation; or applying at least one trained model and/or other function associated with a machine learning process.

10. The apparatus according to claim 7 , wherein the processing circuitry is further configured to adjust the filtered data sets using noise or edge information obtained from the combined data set after the applying of the filter.

11. The apparatus according to claim 1 , wherein the at least some information comprises or is comprised in a further function or mapping, wherein the further function or mapping is representative or at least indicative of at least one of: anisotropy and/or features and/or edges of the volume being scanned.

12. The apparatus according to claim 1 , wherein the filter is characterised by at least one parameter and wherein the at least one parameter and/or the at least some of the information used by the filter is determined as part of a machine learning process.

13. The apparatus according to claim 1 , wherein the applying of the filter comprises an iterative process comprising:

at least a first application to update each of the medical imaging data sets based on the at least some information from the plurality of medical imaging data sets acquired at the different measurement periods and a second application of the filter to the updated image data sets using at least some information from the other updated medical imaging data sets corresponding to the different measurement periods.

14. The apparatus according to claim 1 , wherein the at least one preserved measure of intensity for each set of medical imaging data comprises at least one of:

average intensity, total intensity, a measure of intensity determined using a moving window, or one or more measures of intensity at one or more scales.

15. The apparatus according to claim 1 , wherein the filter comprises a filter function and/or mapping, and wherein at least one of a), b), c), d) or e) occurs:

a) the constraint and/or condition is encoded in or added to the filter function and/or mapping;

b) the filter function and/or mapping is characterised by one or more filter parameters and the one or more filter parameters of the filter function and/or mapping are pre-determined, and wherein the one or more pre-determined filter parameters encode or characterise the at least one constraint or condition;

c) the filter function and/or mapping is characterised by one or more filter parameters and the one or more filter parameters and/or mapping are pre-determined from a training process performed on training data, wherein the training process comprises penalizing differences between a measure of intensity in an input imaging data and a measure of intensity in an output imaging data;

d) the filter function and/or mapping is characterised by one or more filter parameters and wherein the processing circuitry is further configured to obtain one or more filter parameters by applying a machine learning process to a plurality of training data sets comprising medical image data.

16. The apparatus according to claim 1 , wherein at least one of a), b), c), d) or e) occurs:

a) the region of the subject comprises an anatomical feature of interest;

b) the medical imaging data sets comprise at least one contrast medical imaging data set and at least one non-contrast medical imaging data set;

c) the medical imaging data sets comprise at least one data set acquired prior to a medical treatment and at least one data set acquired after said medical treatment;

d) the different measurement periods comprise different time points or different, non-overlapping periods of time;

e) the filtered data sets have reduced noise and/or improved edge definition compared to the data sets before applying of the filter.

17. The apparatus according to claim 1 , wherein

the medical imaging data set comprises CT scan data,

the at least one measure of intensity comprises at least one measure of density,

the at least some information from the other medical imaging data sets used during the filtering comprises gradient information, and

the filter comprises one or more trained multi-layer neural networks or other trained model.

18. The apparatus according to claim 1 , wherein

the processing circuitry is further configured to

obtain combined data from at least some of the plurality of data sets, and

extract gradient information or edge information from the combined data, and

applying of the filter comprises using the extracted gradient information or edge information from the combined data.

19. A medical imaging method comprising:

acquiring a plurality of sets of medical imaging data of a region of a subject, each set of data corresponding to a respective different measurement period;

applying a filter to the plurality of medical imaging data sets to produce a plurality of filtered medical imaging data sets corresponding to the different measurement periods, the filter being applied to each medical imaging data set to produce a corresponding filtered medical imaging data set for each measurement period, wherein

the applying of the filter is such that, for each of the medical imaging data sets, the filtering uses at least some information from the other medical imaging data sets acquired at the different measurement periods, and

the applying of the filter comprises applying at least one constraint or condition and the constraint or condition comprises preserving at least one measure of intensity for each medical imaging data set for each respective measurement period.

20. A system comprising:

processing circuitry configured to train a filter for filtering medical images, wherein the filter is characterized by one or more filter parameters, by:

obtaining a plurality of sets of medical image training data of a region of a subject; and

performing a machine learning process on the plurality of sets of medical image training data to determine values for the one or more filter parameters thereby to obtain a trained filter,

wherein the trained filter is such that filtering a plurality of sets of medical imaging data of a region of a subject, each set of data corresponding to a respective different measurement period and the trained filter being applied to each medical imaging data set to produce a corresponding filtered medical imaging data set for each measurement period, using the trained filter, uses at least some information from the other medical imaging data sets acquired at the different measurement periods and comprises applying at least one constraint or condition and the constraint or condition comprises preserving at least one measure of intensity for each medical imaging data set for each respective measurement period.

21. A method for training a filter for filtering medical images, wherein the filter is characterized by one or more filter parameters, the method comprising:

obtaining a plurality of sets of medical image training data of a region of a subject;

performing a machine learning process on the plurality of sets of medical image training data to determine values for the one or more filter parameters thereby to obtain a trained filter,

wherein the trained filter is such that filtering a plurality of sets of medical imaging data of a region of a subject, each set of data corresponding to a respective different measurement period and the trained filter being applied to each medical imaging data set to produce a corresponding filtered medical imaging data set for each measurement period, using the trained filter, uses at least some information from the other medical imaging data sets acquired at the different measurement periods and comprises applying at least one constraint or condition and the constraint or condition comprises preserving at least one measure of intensity for each medical imaging data set for each respective measurement period.

22. A non-transitory computer-readable storage medium comprising computer-readable instructions that are executable to:

acquire a plurality of sets of medical imaging data of a region of a subject, each set of data corresponding to a respective different measurement period;

apply a filter to the plurality of medical imaging data sets to produce a plurality of filtered medical imaging data sets corresponding to the different measurement periods, the filter being applied to each medical imaging data set to produce a corresponding filtered medical imaging data set for each measurement period, wherein

the applying of the filter is such that, for each of the medical imaging data sets, the filtering uses at least some information from the other medical imaging data sets acquired at the different measurement periods, and

the applying of the filter comprises applying at least one constraint or condition and the constraint or condition comprises preserving at least one measure of intensity for each medical imaging data set for each respective measurement period.