Feature segmentation for non-uniform illumination and low-contrast images

The recipe lets you segment small features on images with non-uniform illumination and low contrast.


Segmentation of image captured under natural environment is often rendered ineffective due to the presence of non-uniform illumination (the average intensity fluctuates along z-axis) and low contrast. These may produce different intensity level and thus lead to inaccurate segmented small features. The recipes proposed an intensity/variance-based image correction integrated with adaptive thresholding that was able to partition small features correctly.

For recipes:

  • Unzip Xtra_Feature-segmentation.zip file;
  • Load the data and recipe in PROJECT and RECIPES view, respectively.
  • Click "Run Recipe"
  • Notes: The reference image in should be a "segmentable slice" from the original image stack

Data courtesy of Yang YU