Document AI & OCR
Image Binarization
Converting an image into two pixel classes, often dark text and light background, using a thresholding rule. A global threshold applies one cutoff; adaptive methods use local information. The output can simplify document processing but can also erase faint strokes or merge nearby marks.
Also called: Binarization
Reviewed
Sources
Member lesson
The definition and sources are public. The complete practical lesson is for members.