Top Banner
Vasilios Aris Morikis Dan DeLahunta Dr. Hyle Park, Ph.D.

Image processing in Spectral Domain Optical Coherence Tomography (SD-OCT)

Dec 31, 2015




Image processing in Spectral Domain Optical Coherence Tomography (SD-OCT). Vasilios Aris Morikis Dan DeLahunta Dr. Hyle Park, Ph.D. Overview. Optical Coherence Tomography An Overview of OCT System Setup Sample Arm Galvanometer Project Overview Methodology Results Conclusions. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
  • Vasilios Aris MorikisDan DeLahuntaDr. Hyle Park, Ph.D.

  • Optical Coherence TomographyAn Overview of OCTSystem SetupSample ArmGalvanometerProject OverviewMethodologyResultsConclusions

  • High resolution sub-surface imagingNon-invasiveNot harmful to subjectPotential in many fieldsOphthalmology (RNFL thickness, AMD)Dermatology (photoaging, BCC detection)Cardiology (assessment of vulnerable plaques)Gastroenterology (Barretts esophagus)

  • Time delay between reflected light is measured to determine depth of the reflecting structureDue to the short time delays between signals OCT must use an interferometer to detect the reflected light.Interference fringes are formed when the sample and reference arms are within a small range.A depth profile is formed by the detection of the interference pattern between the reference and sample arm as the reference arm is scanned.

  • The intensity of the depth profile is encoded on a logarithmic scale.A 2D cross section or even a 3D volume can made by scanning the beam across the sample.

  • Helped to construct the Sample Arm.

  • Built the box to power and control the GalvoVideo of the Galvo moving

  • Develop analysis/processing code in MATLABObjective: Mathematically focus raw data obtained from the 1310 nanometer system.Adjust the incident angle, focal length, and the wavelength.Increase the signal to noise ratio (SNR) to produce high resolution image.

  • Read the ImageFlip Matrix (if necessary)Zero PaddingFFTDisplay ImageInterpolate

  • Raw data obtained when the reference and sample arm are 600 microns apart.Image taken of the mirror.Pixel NumberIntensity

  • Completely unprocessed data.To create accurate image point spread function should be narrow and high (ignore all the noise in the middle).Creates a blurred black line when the actual image is formed.

  • Splits the matrix and adds many 0s in Fourier space.Doubles the size of the original graph.Used to increase the point density to interpolate more accurately.IntensityIntensityPixel NumberPixel Number

  • Used to find remap the data linearly in wave number (k) to improve the results of a subsequent FFTTakes the Intensity vs. Pixel number graph and Intensity vs. k.

  • Fourier transform switches one complex valued function into another.Transforming k (wave number) into actual space.

  • Side CameraStraight CameraIncident Angle49 degrees51 degreesGrating Spacing1.0e-3/11451.0e-3/1145

    Focal Length92 millimeters95 millimetersWavelength1350 nanometers1350 nanometersPixel Width25 micrometers25 micrometers

  • Now that the parameters are correct a much more focused image is created.Dark line at the top is thin and not blurry

  • Image obtained form the 1310 nanometer system before processing (left) and after processing (right).Image width:100 micronsImage height: 500 microns

    Very first image acquired with either system.

  • I would like to thank NSF and the UC Riverside BRITE program for funding, as well as the University of California, Riverside and NIH (R00 EB007241), and Dr. Hyle Park and the rest of the Park Research Group for their guidance.Mr. Jun Wang for organizing and Dr. Victor Rodgers for directing the program.

  • **********************