A dataset of 11k+ handwritten Arabic words from Egyptian writers, extracted and tightly cropped from scanned paper forms. This dataset offers diverse handwriting samples ranging from children to elderly contributors, making it ideal for training robust Arabic handwriting recognition models.
Each form contains 6 unique words, resulting in 24 handwritten word images per form.
Each word is written four times by the same writer to capture… See the full description on the dataset page:
https://huggingface.co/datasets/OmarMDiab/Egyptian-Handwriting-Dataset.
