Purchase this article with an account.
Hiroshi Ishikawa, Gadi Wollstein, Larry Kagemann, Juan Xu, Joel S. Schuman; Effect of Speckle Noise Reduction on Spectral Domain Optical Coherence Tomography (SD-OCT) En Face Images. Invest. Ophthalmol. Vis. Sci. 2011;52(14):1303.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To investigate the effect of a new speckle noise reduction method on SD-OCT en face images.
Two-hundred-forty SD-OCT three-dimensional cube images (200x200 sampling points) with various pathologies were processed (Cirrus HD-OCT; Carl Zeiss Meditec, Inc., Dublin, CA, and RTVue; Optovue, Fremont, CA). At each sampling point, the true signal intensity profile was estimated based on two-dimensional mean filtered data. Only pixels with an intensity value higher than the modeled signal intensity profile were adjusted to the model intensity. This way, high frequency components that have low intensity can be preserved unlike many previously published algorithms that introduce a blurring artifact. OCT en face images were subjectively evaluated for the speckle noise reduction effect.
All images showed notable improvement in image/signal quality, regardless the scanning location (macula or optic nerve head), pathology, and signal strength. An example of speckle reduction effect is shown in the Figure. The speckle reduced en face image (B) shows noticeably clearer borders of retinal vessels and optic nerve head contour as well as choroidal vessels, compared to the raw image (A). In the raw cross-sectional image (C), speckle noise obscures the retinal layer structure, while the processed image (E) exhibits minimal blurring while preserving edges between different layers. It is difficult to distinguish true tissue borders in the raw A-scan profile (D, along the vertical white line on C), while the processed one (F) displays retinal structure signal more easily distinguishable.
Visulization of retinal structures within SD-OCT en face images can be improved by the proposed method while preserving the image details. This method may be a clinically useful OCT post-processing tool.
This PDF is available to Subscribers Only