Artifact removal from multimodality OCT images

A non-transitory computer-readable medium configured to store instructions that when executed by one or more processors of a computing device, cause the computing device to perform a method of removing artifacts from multimodality image data of a biologic lumen, the method comprising:

• acquiring a multimodality image of a biological lumen through a catheter, wherein the multimodality image of the biological lumen includes optical coherence tomography (OCT) data and fluorescence data;
• receiving a plurality of B-scan frames of the OCT and fluorescence data of the multimodality image acquired by the catheter;
• comparing fluorescence data values of each B-scan frame to a threshold value;
• for each B-scan frame detected to have a fluorescence data value equal to or higher than the threshold value, setting the B-scan frame as a potential error frame;
• acquiring a fluorescence data value for a preceding B-scan frame and a fluorescence data value for a subsequent B-scan frame relative to the B-scan frame set as the potential error frame;
• inputting the fluorescence data value of the potential error frame, the fluorescence data value of the preceding B-scan frame, and the fluorescence data value of the subsequent B-scan frame into a two-class classifier;
• and
• determining, using the two-class classifier, whether the fluorescence data value of the potential error frame is a fluorescence artifact or not,
• wherein the two-class classifier includes a machine learning classification algorithm trained to determine whether the fluorescence data value of the potential error frame corresponds to one or more known fluorescence artifacts with which the algorithm has been trained, and
• wherein the machine learning classification algorithm is a supervised algorithm selected from the group consisting of a neural network algorithm, a support vector machine (SVM) algorithm, a Nearest Neighbor (NN) algorithm, a k-NN algorithm, and a random forest (RF) algorithm.

A method, comprising:

• acquiring a multimodality image of a biological lumen during a pullback of a catheter inserted into the biological lumen, the multimodality image including optical coherence tomography (OCT) data co-registered with fluorescence data, the OCT and fluorescence data respectively collected by scanning the lumen with light of first and second wavelengths transmitted through the catheter inserted into the lumen;
• analyzing a plurality of B-scan frames of the OCT and fluorescence data of the multimodality image acquired by the catheter, wherein each B-scan frame includes a plurality of A-lines of that B-scan frame;
• for the plurality of B-scan frames, calculating a parameter value of the fluorescence data for each B-scan frame;
• sorting the parameter values of the plurality of B-scan frames from lowest to highest and arranging the sorted parameter values along a parameter curve;
• tracing a straight line from the lowest value to the highest value of the parameter curve;
• calculating a maximum perpendicular distance between the straight line and the parameter curve;
• setting the value along the parameter curve corresponding to the maximum perpendicular distance as a threshold value;
• comparing the fluorescence data values of each B-scan frame to the threshold value;
• and
• for each B-scan frame detected to have a fluorescence data value equal to or higher than the threshold value, setting that B-scan frame as a potential error frame.