Color Images

Quantitative analysis of 2D color microscopy images is supported in XamFlow. For example this is useful for histopathology, histomorphometry of histological sections, or for petrographic thin sections. The goal might be to accurately count cells or grains, to measure tissue area and density, and analyze particle distributions with respect to their size, shape, and orientation.

Use the standard Lucid.Util.ImportImageFromForeignFile. Ensure that the “convert RGB to grayscale” option is set to “No” to preserve the color information, and confirm that “convert 2D stack to 3D” is set to “No” when processing individual 2D images.

Example.skimage.Gauss supports color images. The Gaussian Filter allows denoising images by low-pass filtering.

XF.Interactive.DefineSegmentationRGB is the main interactive task type for thresholding color images. A red overlay indicates the objects identified based on the defined color range. Different color spaces can be used. In HSV space, the Hue, Saturation and Value range sliders are used to select the features of interest. It is recommended to start with a full range for each slider, and reduce each separately as needed. The checkbox can invert the range, which is mostly useful to define a Hue range of red hues from both ends of the color spectrum, but excluding green or blue hues.

Some Machine Learning models also support using multiple color channels.