Methods and apparatus for reducing artifacts in OCT angiography using machine learning techniques

1. A method of reducing artifacts, comprising:

obtaining OCT/OCTA data from an OCT/OCTA imager;

preprocessing OCTA/OCT volume data;

extracting features from the preprocessed OCTA/OCT volume data;

classifying the OCTA/OCT volume data to provide a probability determination data, wherein classifying the OCTA/OCT data includes returning a probability determination data that represents the probability in each base unit that the base unit belongs to one of a set of classification categories;

determining a percentage data from the probability determination data; and

reducing artifacts in response to the percentage data.

2. The method of claim 1 , wherein extracting features includes extracting features in each base unit of the OCTA/OCT data.

3. The method of claim 2 , wherein the base unit can be a single voxel.

4. The method of claim 2 , wherein the base unit is a plurality of voxels.

5. The method of claim 1 , wherein the set of classification categories includes a purely true flow signal, a purely artifact signal, or a mixture of both true flow and artifact signals.

6. The method of claim 1 , wherein classifying the OCTA/OCT data includes using a trained classifier to determine the probability determination data.

7. The method of claim 6 , further including training the trained classifier, training the trained classifier comprising:

providing a training dataset;

preprocessing the training dataset;

extracting features in the training dataset;

classifying the training dataset to obtain probability determination data;

comparing the probability determination data with human labeled probability data;

refining the trained classifier such that the probability determination data matches the human labeled probability data.

8. A method of reducing artifacts, comprising:

obtaining OCT/OCTA data from an OCT/OCTA imager;

preprocessing OCTA/OCT volume data;

extracting features from the preprocessed OCTA/OCT volume data;

classifying the OCTA/OCT volume data to provide a probability determination data;

determining a percentage data from the probability determination data; and

reducing artifacts in response to the percentage data,

wherein preprocessing OCTA/OCT volume data, comprises:

detection of regions with the OCTA/OCT signals above background noise,

excluding regions in the OCTA/OCT signals that are not above background noise,

detecting landmarks along each OCTA/OCT A-line, and

flattening to align all of A-scans with a chosen landmark.

9. A method of reducing artifacts, comprising:

obtaining OCT/OCTA data from an OCT/OCTA imager;

preprocessing OCTA/OCT volume data;

extracting features from the preprocessed OCTA/OCT volume data;

classifying the OCTA/OCT volume data to provide a probability determination data;

determining a percentage data from the probability determination data; and

reducing artifacts in response to the percentage data,

wherein determining a percentage data from the probability determination data includes a transform equation or matrix to transform the probability of the base unit belonging to each categories to a single true flow signal percentage value in the base unit.