DigitConfuse-23k: A Synthetic Dataset of Digit Confusion Patterns
收藏IEEE2026-04-17 收录
下载链接：
https://ieee-dataport.org/documents/digitconfuse-23k-synthetic-dataset-digit-confusion-patterns
下载链接
链接失效反馈官方服务：
资源简介：
\ud83d\udcca DigitConfuse-23k: A Synthetic Dataset of Digit Confusion PatternsDigitConfuse-23k is a synthetic dataset containing 23,000 images of digit pairs designed to capture visual anomalies and confusion cases commonly encountered in OCR, CAPTCHA recognition, optical illusions and human digit interpretation tasks.Each image contains two-digit numbers generated using the Humor-Sans font (font_size=32, cell_w=60, cell_h=40). For each confusion category, ~1000 images are included.The dataset is also made available on other trusted public dataset repositories for better transparency and wider usage. \ud83d\udd22 Categories of Digit Anomalies\ud83d\udd38 Digit shape confusion (similar glyphs) \u2192 11 \u2194 17, 21 \u2194 27, 71 \u2194 77\ud83d\udd04 Mirror \/ rotation confusion \u2192 69 \u2194 96, 68 \u2194 86, 89\u219498, 26 \u2194 62\ud83c\udfaf One-pixel stroke differences \u2192 33 \u2194 38, 35 \u2194 36, 53 \u2194 58, 39\u219489\ud83c\udf00 Closed vs. open loop confusion \u2192 38 \u2194 88, 98 \u2194 99, 18 \u2194 19, 56\u219458, 28\u219488\u27bf Nearly identical when repeated \u2192 88 \u2194 89, 11 \u2194 12, 55 \u2194 56\ud83d\udc40 Human OCR-like errors (CAPTCHA\/OCR cases) \u2192 47 \u2194 17, 57 \u2194 37, 12 \u2194 72, 14 \u2194 74\ud83c\udfaf Applications\ud83e\uddea Benchmarking OCR systems\ud83d\udee1 Studying digit recognition robustness\ud83d\udd11 Training models for noisy \/ CAPTCHA-like digits\ud83d\udea8 Anomaly detection in digit datasets
提供机构：
Bhuma Chandra Mohan
