shenshuming / OCT

GUI for segmenting the lumen of intravascular OCT pullbacks

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

OCT segmentation

This repository is a MatLab GUI that contains a series of functions that segments the lumen of OCT images. Users of this repository should cite the following papers

  1. Evaluation of a framework for the co-registration of intravascular ultrasound and optical coherence tomography coronary artery pullbacks. Molony DS, Timmins LH, Rasoul-Arzrumly E, Samady H, Giddens DP. 2016. J Biomech, Vol 49(16)
  2. Volmetric quantification of fibrous caps using intravascular optical coherence tomography. Wang Z, Chamie D, Bezerra HG, Yamamoto H, Kanovsky J, Wilson DL, Costa MA, Rollins AM. 2010. Biomed Optics Express, Vol 3(6)

Algorithm

The code uses a dynamic programming algorithm to segment the lumen. A cost function based on the image gradient is first calculated in the polar domain. The cumulative cost of this cost function is then calculated. The lumen border is found by backtracking a path through the minimum cost path of the cumulative cost image. Refer to Wang et al. for more details

GUI

The user launches the GUI by typing the following in the MatLab command line

OCT_GUI

On the launch screen the user then selects to load the OCT pullback. Currently OCT data can be loaded in either it's raw format (must be converted to .tif, this can be done with ImageJ) or in dicom format

Alt Text

After loading the data the OCT images are displayed in both the cartesian and polar domains.

Alt Text

Next, the guidewire is removed from the image by selecting Segment guidewire. If the default values do not result in adequate removal of the guidewire these values can be changed by editing the Catheter and Guidewire values. If successful the removed guidewire can be clearly seen in the polar domain view.

Alt Text

The final pre-processing step is the removal of the sheath by selecting Segment sheath. This brings up a pop-up window where the user selects two points representing the diameter of the outer sheath and presses return. Next, the user selects two points representing the diameter of the inner sheath and presses return

Alt Text

The sheath segmentation begins and a progress bar tracks the completion of this task. This takes approximately 30 minutes on a laptop.

Alt Text

The success of the sheath segmentation can be judged in the polar domain by checking whether the contour has correctly found the sheath outer boundary

Alt Text

There are 2 choices for lumen segmentation. These options can be selected by the Segment lumen or Threshold lumen button. The Segment lumen method should take less than 30 minutes on a laptop. The success of the lumen segmentation can be judged in both the polar and cartesian domains. The second method of segmenting the lumen is through thresholding with the Threshold lumen button. This can be applied at the individual slice level or at all levels by checking the Threshold all box.

Alt Text

If some of the lumen contours are not satifactory these can be edited by pressing the Edit lumen button. This generates a spline where the points can be moved from until satisfied. When satisfied with the position of the points press the Recompute button.

Alt Text

The Copy lumen button will copy the contour from the previous image to the current image Both the sheath and the lumen contours can be saved and loaded by the Save and Load lumen buttons respectively.

Requirements

The code has been tested on Matlab R2016b. It requires the signal processing and image processing toolboxes. This code also adapts the interactive spline code splineroi from here for manual editing.

About

GUI for segmenting the lumen of intravascular OCT pullbacks


Languages

Language:MATLAB 100.0%