1. Image preprocessing

Last modified by rbakker on 2023/07/17 12:35

Image preprocessing

To ensure consistent performance of our tools on a variety of images, we apply a standard preprocessing strategy that consists of the following steps:

  • Invert so that background is black
  • Remove background gradient, by fitting a linear model through points sampled just outside the tissue
  • Remove background with OneCut (optional)
  • Enhance contrast with CLAHE (optional)
  • Resize images to common size
  • Center images

The steps are implemented as part of the registration pipeline collab:

https://drive.ebrains.eu/d/67ca7ef8c35b45ec99b9/