Related Experiment Video
Updated: Apr 25, 2026
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
4.5K
Online handwritten signature verification using neural network classifier based on principal component analysis.
Vahab Iranmanesh1, Sharifah Mumtazah Syed Ahmad1, Wan Azizun Wan Adnan1
1Department of Computer and Communication Systems Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia.
Thescientificworldjournal
|August 19, 2014
Summary
This study introduces a novel online signature verification (OSV) method using principal component analysis (PCA) features with a multilayer perceptron (MLP). The approach enhances accuracy by utilizing discarded PCA data, reducing both false acceptance and rejection rates.
Area of Science:
- Computer Science
- Biometrics
- Pattern Recognition
Background:
- Online signature verification (OSV) faces challenges due to inherent variability in genuine signatures and sophisticated forgeries.
- Effective feature extraction is crucial for distinguishing authentic signatures from fakes in OSV systems.
Purpose of the Study:
- To propose a systematic approach for online signature verification using multilayer perceptron (MLP) on principal component analysis (PCA) features.
- To investigate a feature selection technique that leverages typically discarded PCA information to improve OSV accuracy.
Main Methods:
- Utilized a multilayer perceptron (MLP) classifier for signature verification.
- Applied principal component analysis (PCA) for feature extraction and developed a method to utilize a subset of these features, including previously discarded information.
- Evaluated the proposed method on 4000 signature samples from the SIGMA database.
Main Results:
- Achieved a False Acceptance Rate (FAR) of 7.4%.
- Achieved a False Rejection Rate (FRR) of 6.4%.
- Demonstrated the effectiveness of incorporating previously discarded PCA features in reducing error rates.
Conclusions:
- The proposed systematic approach using MLP on a subset of PCA features, including discarded information, is effective for online signature verification.
- The feature selection technique significantly contributes to reducing error rates in OSV systems.
- The method shows promise for improving the reliability and accuracy of biometric security systems.
