This technology standardizes OCT images by resampling to a fixed pixel density, reducing speckle noise with targeted smoothing, and adjusting signal amplitudes through histogram analysis. It employs techniques like linear interpolation, a 2D mean filter, and selective noise reduction to preserve important image details. The process further enhances image quality by rescaling pixel intensities between specific percentile thresholds, and offers either high dynamic range processing—dividing and linearly scaling different intensity channels—or histogram matching based on anatomical segmentation, thereby producing a consistent, normalized output that can be directly compared across different device sources.
Description
What sets this technology apart is its device-independent approach, eliminating the need for device-specific tuning or training, which enables uniform OCT data analysis and interpretation across diverse platforms. By methodically addressing variations in sampling density, noise, dynamic range, and overall image quality, it allows for accurate longitudinal and cross-device studies. This seamless integration of standardized normalization methods not only streamlines qualitative and quantitative assessments but also facilitates more reliable and consistent clinical evaluations and research studies in the field of OCT imaging.
Applications
- Automated OCT normalization
- Multi-device OCT standardization
- Direct OCT comparison
- Longitudinal OCT analysis
- Universal OCT browsing
Advantages
- Enables direct comparison of OCT images from different devices by standardizing image formats.
- Reduces variability introduced by differences in sampling density, speckle noise, and dynamic range.
- Improves image quality while preserving critical anatomical details through selective smoothing and adaptive scaling.
- Facilitates longitudinal and multi-device studies with a consistent, device-independent data format.
- Eliminates the need for device-specific tuning or conversion parameters, streamlining the analysis process.
IP Status
https://patents.google.com/patent/US9844315B2
