CEDAR Signature

CEDAR签名数据集适用于笔迹识别和签名认证。数据集共收录了55位志愿者的笔迹结果，包括24位真实签名和24位伪造签名，共2,640张图像。以300 dpi的灰度扫描每个签名，并使用灰度直方图二值化。椒盐噪声去除和倾斜归一化是图像预处理涉及的两个步骤。

The CEDAR Signature Dataset is designed for handwriting recognition and signature verification. It includes handwriting samples from 55 volunteers, with 24 genuine signatures and 24 forged signatures per volunteer, totaling 2,640 images. Each signature was scanned in grayscale at 300 dpi, and binarized using the grayscale histogram. Two steps are involved in the image preprocessing: salt-and-pepper noise removal and skew normalization.

搜集汇总

背景与挑战

背景概述

CEDAR签名数据集是一个用于笔迹识别和签名认证的数据集，包含55位志愿者的2,640张签名图像，其中真实和伪造签名各24位。所有签名图像均以300 dpi灰度扫描，并经过二值化、噪声去除和倾斜归一化等预处理步骤。