Score indicating OCR engine certainty in recognized text.
Lower confidence can signal noise, skew, or poor scan quality.
Useful for QA thresholds in extraction workflows.
Use this concept when optimizing, converting, validating, or troubleshooting document workflows.
Optical Character Recognition that turns scan images into machine-readable text.
Cutoff score used to accept or reject OCR-detected characters/words.
Pixel-level decision method used in binarization to separate foreground text from background.