Signature Matching Dataset

What’s Inside the Dataset?

1. Signature Data

• Total Users: 124 individuals with a mix of genuine and forged signatures.
• Genuine Signatures: Each person has approximately 10 authentic signatures they personally created.
• Forged Signatures: Each user also has around 10 forged signatures imitated by another person.

2. Structured Directory System

• The dataset is well-organized by user ID.
• Genuine signatures are stored under the user’s unique number.
• Forged signatures are labeled using the user’s number followed by “_forg” for easy classification.

3. Data Source

• Extracted from ICDAR 2011 Signature Dataset and CEDAR Signature Verification Dataset for high accuracy.
• Carefully organized for user-friendly accessibility and research.

Advantages of the Signature Matching Dataset

Ideal for AI & ML Training – Train signature verification models with real and forged data.
Fraud Detection Applications – Improve authentication systems for banking and document security.
Deep Learning Research – Use CNNs and RNNs for signature analysis.
High-Quality Data – Sourced from trusted datasets for superior accuracy.
Structured & Labeled – Organized format ensures easy usability in AI projects.

Applications

Banking Security: Detect forged signatures in financial transactions.
Forensic Analysis: Assist forensic experts in fraud investigations.
Identity Verification: Improve biometric authentication systems.
AI Model Training: Enhance machine learning datasets for fraud prevention.