Related Experiment Video
Updated: May 21, 2025
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
1.7K
Learning features for offline handwritten signature verification using spatial transformer network
Wanghui Xiao1,2, Hao Wu3
1School of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing, China. xiaowanghui007@126.com.
Scientific Reports
|March 20, 2025
Summary
This study introduces a novel two-stage Siamese network for offline handwritten signature verification, improving accuracy in identifying genuine and forged signatures. The model uses a spatial transformer network and Focal loss for enhanced feature focus and imbalanced data handling.
Area of Science:
- Biometrics
- Document Forensics
- Computer Vision
Background:
- Offline handwritten signatures are crucial for identity verification in various applications.
- Skillful forgeries make accurate offline signature verification a significant challenge.
Purpose of the Study:
- To develop an advanced model for robust offline handwritten signature verification.
- To enhance the discrimination between genuine and forged signatures.
Main Methods:
- A two-stage Siamese network architecture was proposed.
- Integration of a spatial transformer network for feature reconstruction and focus.
- Utilized Focal loss to address class imbalance in signature datasets.
Main Results:
- The proposed model demonstrated superior performance on four diverse handwritten signature datasets.
- Achieved higher verification accuracy compared to existing state-of-the-art methods.
Conclusions:
- The novel Siamese network with spatial transformer and Focal loss effectively improves offline signature verification.
- The model shows promise for real-world applications requiring high-accuracy identity verification.
