Signal normalization among multiple optical coherence tomography devices
AssigneeUniversity of Pittsburgh—Of the Commonwealth System of Higher Education
InventorsChieh-Li Chen, Hiroshi Ishikawa, Joel Steven Schuman, Chaim-Gadi Wollstein
Life of this patent
Estimating the term…
Abstract
The present disclosure is directed to systems and methods for normalizing an image of an eye among multiple optical coherence tomography (OCT) devices. The system can receive an OCT image and generate a normalized OCT image. The system is configured to normalize a sample density of the received image, normalize an amount of noise using a suitable noise reduction technique, normalize a distribution of signal amplitudes of the image, and then normalize a signal quality of the image. The resulting normalized image of the eye can then be used in the comparison among OCT images taken from multiple OCT devices.
Claims18
A method for normalizing an image of an eye among multiple optical coherence tomography (OCT) devices comprising, in the following specified sequence:
- (a) receiving a first pixmap comprising a first plurality of pixels, wherein the first pixmap represents an image of an eye and the first pixmap has a first sample density;
- (b) normalizing a sample density;
- (c) normalizing an amount of noise using a suitable noise reduction technique;
- (d) applying a histogram matching (HM) process to normalize a distribution of signal amplitudes by:
- (i) dividing the first pixmap into a first portion and a second portion along a partitioning line;
- (ii) calculating a first histogram from the first portion of the first pixmap;
- (iii) mapping the first histogram to a first reference histogram to generate a first mapped histogram;
- (iv) calculating a second histogram from the second portion of the first pixmap;
- (v) mapping the second histogram to a second reference histogram to generate a second mapped histogram;
- and
- (vi) generating a second pixmap based on a combination of the first mapped histogram and the second mapped histogram;
- and
- (e) normalizing a signal quality of the second pixmap;
- wherein the sequence of steps results in a normalized image of the eye among multiple optical coherence tomography devices.
A system for normalizing an image of an eye among multiple optical coherence tomography devices comprising one or more data processors and one or more storage devices storing instructions that when executed by the one or more data processors cause the one or more data processors to:
- (a) receive a first pixmap comprising a first plurality of pixels, the first pixmap representing an image of an eye and having a first sample density;
- (b) generate a second pixmap comprising a second plurality of pixels by resampling the first plurality of pixels at a predetermined sampling density;
- (c) filter the second pixmap to remove at least a portion of a noise signal in the second pixmap;
- (d) rescale a distribution of signal amplitudes of the filtered second pixmap such that the distribution of signal amplitudes spans a substantial portion of a dynamic range of the filtered second pixmap;
- (e) normalize the distribution of signal amplitudes of the rescaled, filtered second pixmap with a histogram matching (HM) process, wherein to normalize the distribution of signal amplitudes the instructions cause the one or more data processors to:
- (i) divide the second pixmap into a first portion and a second portion along a partitioning line;
- (ii) calculate a first histogram from the first portion of the second pixmap;
- (iii) map the first histogram to a first reference histogram to generate a first mapped histogram;
- (iv) calculate a second histogram from the second portion of the second pixmap;
- (v) map the second histogram to a second reference histogram to generate a second mapped histogram;
- and
- (vi) generate a third pixmap based on a combination of the first mapped histogram and the second mapped histogram;
- and
- (f) provide a normalized image of the eye based on the third pixmap.
Dependencies are inferred from the claim text.