mazy964 / oct_preprocess Goto Github PK

Preprocess of Optical coherence tomography (OCT) includes following steps:

1. octSpectrailisReader convert OCT image into python processable nd-array and retrieve useful information in the image OCT image in the first layer shown by matlab.pyplot.imshow:

2. retinaDetect find the boundaries of inner limiting membrane(ILM), inner segment(IS), outer segment (OS) and Bruch’s membrane (BM) Three lines on the image shown are the ILM, ISOS(combinatino of IS and OS), and BM boundaries detected by the code:

3. normalizeOCT normalize and reduce noise of the OCT image after normalizing the image, grayscale image looks like:

4. retinaFlatten calculate the shitfs based on return value in retinaDetect and flatten the image using BM as baseline. The final image in both grayscale and RGB: