Improving ocr accuracy using specialized character verification models
Improving OCR Accuracy Through Multi-Classifier Verification of Visually Similar Characters
Computer Vision and Pattern RecognitionInformation Retrieval
Summary
OCR systems sometimes confuse characters that look very alike, like the letter O and the number 0, causing mistakes when copying text from images. The authors created a method that checks these tricky characters one by one after OCR has done its job, using special classifiers trained to tell apart such pairs. This method works without relying on dictionaries or word context, so it can fix errors in codes, serial numbers, or license plates. Tests showed the approach helped multiple OCR engines improve their accuracy by up to nearly 15 percentage points with few wrong fixes.
What this means in practice
- •For document digitization teams: Enhance OCR output accuracy for documents containing ambiguous alphanumeric characters without relying on dictionary checks.
- •For license plate recognition providers: Improve recognition of visually similar characters in license plates to reduce misreads and enhance identification reliability.
Authors
Nuzhat Khan, Ab Al-Hadi Ab Rahman, Mohamad Anuar Kamaruddin, Muhammad Paend Bakht, Sana Fatima, Farhad Khodadadi, Ibrahim Yousef Alshareef
Abstract
Distinguishing visually similar characters is a persistent challenge for optical character recognition (OCR) systems, causing errors in critical applications of text digitization. While preprocessing techniques improve image quality and word-level post-processing corrects spelling errors, neither addresses character-level visual ambiguity, an intrinsic property of certain character pairs that limits even state-of-the-art OCR models. We present the Character Verification Model (CVM), a character-level post-processing framework that verifies and corrects ambiguous characters in OCR outputs using specialized binary classifiers trained on visually confusable pairs such as 'O'/'0', 'Z'/'2', and 'S'/'5'. CVM operates independently of dictionary validation, making it suitable for non-lexical data including serial numbers, alphanumeric codes, and license plates. Experiments on three OCR engines (Tesseract, EasyOCR, TrOCR) and ten fonts with low to high complexity demonstrate accuracy improvements of 3.95-14.81 percentage points, with an 87.3% correction rate for visually similar character errors and a false-correction rate of only 2.3%. The modular architecture integrates with existing OCR pipelines without modification to underlying recognition models.