Handwritten Signature Identification
收藏www.kaggle.com2023-04-27 更新2025-01-21 收录
下载链接：
https://www.kaggle.com/omgits0mar/signature-verification-and-identification
下载链接
链接失效反馈官方服务：
资源简介：
Simplified version of Handwritten Signature datasets where it has 5 classes (PersonA...PersonE) splited into training/ testing folders. and a `CSV files` folder containing whether its real or forged for each class in training / test.
The goal here is to train 2 models (1 for identifying the person sig. and 2 for validating that it is a real one).
This dataset is part of the Computer Vision course by FCIS-ASU divided into 2 main tasks :
> 1. Signature Identification : Building a classical vision or DL techniques to identify the person of this signature.
> 2. Signature Verification : Build a 5 different models to identify for each class if it is a real or fake OR build a 1 siamese network that verifies the signature of the person whether it is a real one or fake one.
本数据集为简化版手写签名数据集，分为5个类别（PersonA至PersonE），并划分为训练集/测试集文件夹。此外，还有一个包含各类别真伪标识的`CSV文件`文件夹，分别对应训练集和测试集。本研究的目的是训练2个模型（1个用于识别签名者身份，另1个用于验证签名的真实性）。该数据集隶属于FCIS-ASU计算机视觉课程，分为两大主要任务：
> 1. 签名识别：构建经典视觉或深度学习技术以识别签名者。
> 2. 签名验证：构建5个不同的模型，用于识别每个类别的签名是真还是伪，或者构建一个用于验证个人签名真实性的单一Siamese网络。
提供机构：
Kaggle
