Color Palettes Overview After Thresholding Process with Default Methods of ImageJ or FIJI∗

Abstract

Coloring techniques in life sciences continue to progress, which have now used deep learning in the staining stage to create virtual staining. Usually, manual staining produces an image with a more vigorous color intensity than virtual staining. This study aims to obtain an overview of the color palette (310 colors) after processing with the default threshold method in ImageJ v1.53s or FIJI (ImageJ2) version 2.3.0/1.53q with several variations of the color threshold and color space, which helps determine algorithms in image processing. The variants of the color threshold were (red, white(W), black(B), and B+W). The color space variants were hue-saturation-brightness (HSB), red-green-blue (RGB), Lab, and YUV. The background variant used in the image processing in this study is dark. Later, after the image is binarized, the object in the image to be analyzed is colored black. The separation results of 16 descriptions of threshold color and color space variants showed that the combination with HSB color space is the most sensitive to color differences, reducing red color the most. In contrast, the RGB color space is the least susceptible to color differences. Lab and YUV color spaces have almost the same segmentation effect. The results of this study not only can be applied in analyzing all types of manual or digital images in all branches of science and art.