Offline Handwriting Signature

This dataset consists of 12,600 images from 420 unique individuals, as well as 30 signatures for each participant. All signatures were provided by students and faculty from the University of Raparin in Ranya, Iraq, and Firat University in Elazig, Turkey. The methods for collecting our dataset are detailed below.
1. Set up a grid of five rows and three columns on an A4 sheet of paper so that each person can sign 15 signatures on each side.
2. Provide each signer with two pages of paper with a grid for the required thirty signatures.
3. The signatures of each individual must be of the same type.
4. Each signature on the grid paper must be written in blue and black ink.
We have a team dedicated to accumulating signatures and providing feedback prior to, following, and throughout the signing process. The accumulation of datasets occurred over the span of two months.
Each A4 page contained 15 signatures during the preprocessing phase; therefore, we automatically cropped and saved them using MATLAB's Crop function. Each signature was then preserved as a unique image and placed in the class corresponding to the corresponding individual. In the concluding phase, the images were resized to 500 x 600 pixels. We manually reviewed and cropped the images in search of these instances because, during automated cropping, it is conceivable that some signatures were removed from the boundary and other signature portions were obliterated.