Correlation-Based OCTA Processing to Reduce Projection Artefacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
OCTA enface projections often contain projection artefacts that mimic blood vessels, particularly at high-reflectance tissue layers like the Retina Pigment Epithelium, leading to misdiagnosis risks in clinical applications.
Innovation Solution
A method and apparatus process OCTA data by calculating correlation values between sequences of data elements in B-scans to correct false signals in the axial direction, reducing projection artefacts before generating enface projections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OCTA imaging is performed at deeper high-reflectance layers, then structural information is obtained, but projection artefacts from superficial vasculature increase
Solution Approach 1:
The patent applies preliminary action by performing projection artefact removal processing on OCTA data before generating enface projections. The system calculates correlation values between adjacent depth segments and suppresses projection artefacts in advance, ensuring that the structural information in deeper layers is not contaminated by artefacts from superficial vasculature.
Solution Approach 2:
The patent extracts and removes the harmful projection artefact component from the OCTA data. By calculating correlation values between adjacent depth segments and identifying pixels with high correlation (indicating projection artefacts), the system selectively removes these artefacts while preserving genuine blood flow signals, thereby separating the harmful component from the useful structural information.
2Object-generated harmful factors
If correlation-based correction is applied to reduce projection artefacts, then artefact reduction is achieved, but processing complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the OCTA data into multiple adjacent depth segments along the axial direction. By processing each segment separately and calculating correlation values between corresponding pixels in adjacent segments, the system reduces the computational complexity compared to processing the entire depth range at once, while still effectively identifying and removing projection artefacts.
Solution Approach 2:
The patent applies partial action by focusing the correlation-based correction only on pixels that exhibit high correlation values between adjacent depth segments. Rather than applying complex processing to all pixels uniformly, the system selectively processes only those pixels likely to contain projection artefacts, thereby reducing overall processing complexity while maintaining effectiveness.
Data Source
AI summary
A method of processing optical coherence tomography angiography, OCTA, data comprising an array of columns of data elements to generate corrected OCTA data showing reduced projection artefacts in enface projection, the OCTA data having been generated from repeat B-scans, the method comprising: (i) calculating a correlation value indicative of a correlation between a first sequence of the data elements, whose respective locations within respective B-scans correspond to a location of a first data element within the array, and a second sequence of the data elements, whose respective locations within respective B-scans correspond to a location of a second data element within the array; (ii) calculating a correction for the second data element based on the first data element and the calculated correlation value; and (iii) applying the calculated correction to the second data element, wherein processes (i) to (iii) are performed multiple times to generate the corrected OCTA data.