Ink and Identity - A Handwriting Image Dataset for Graphological Applications

The Ink and Identity dataset presents a comprehensive collection of 374 handwritten characters consisting of all lowercase and 12 uppercase alphabets, including a few digits, gathered from participants spanning diverse age groups. Utilized A4-size unruled paper for writing. As of now, the datasets available are on ruled paper containing only a single line or single word, which limits analyzing data in graphology, and all the datasets do not contain all alphabets, like numbers and characters. These limitations are covered in this dataset.

The dataset was collected using two methods: an in-person (300 images) and a web form (150 images). Total dataset: 450 images. The dataset comprises two folders, in-person and web form, and the image was captured in a 1:1 ratio on an iPhone 14 Plus 12MP. Web form images were instructed to be captured in a 1:1 ratio, and each one has been pre-processed with a unique ID like “001” and stored in .jpg format.

Additionally, researchers can employ this dataset as a reference standard for age-based handwriting analysis, personality trait prediction, AI/ML model training, cognitive and neurological research, and health and mental well-being.