Abstract:Handwritten signature verification poses a formidable challenge in biometrics and document authenticity. The objective is to ascertain the authenticity of a provided handwritten signature, distinguishing between genuine and forged ones. This issue has many applications in sectors such as finance, legal documentation, and security. Currently, the field of computer vision and machine learning has made significant progress in the domain of handwritten signature verification. The outcomes, however, may be enhanced depending on the acquired findings, the structure of the datasets, and the used models. Four stages make up our suggested strategy. First, we collected a large dataset of 12600 images from 420 distinct individuals, and each individual has 30 signatures of a certain kind (All authors signatures are genuine). In the subsequent stage, the best features from each image were extracted using a deep learning model named MobileNetV2. During the feature selection step, three selectors neighborhood component analysis (NCA), Chi2, and mutual info (MI) were used to pull out 200, 300, 400, and 500 features, giving a total of 12 feature vectors. Finally, 12 results have been obtained by applying machine learning techniques such as SVM with kernels (rbf, poly, and linear), KNN, DT, Linear Discriminant Analysis, and Naive Bayes. Without employing feature selection techniques, our suggested offline signature verification achieved a classification accuracy of 91.3%, whereas using the NCA feature selection approach with just 300 features it achieved a classification accuracy of 97.7%. High classification accuracy was achieved using the designed and suggested model, which also has the benefit of being a self-organized framework. Consequently, using the optimum minimally chosen features, the proposed method could identify the best model performance and result validation prediction vectors.

What problem does this paper attempt to address?

The paper attempts to address a key issue in handwritten signature verification, namely how to distinguish between genuine and forged signatures to improve the accuracy of signature verification. Specifically, the paper proposes an offline handwritten signature verification framework that combines transfer learning and feature selection methods, aiming to enhance classification accuracy by optimizing feature selection and using various machine learning algorithms. ### Main Research Questions: 1. **Improve the accuracy of signature verification**: Enhance the classification accuracy of signature verification by optimizing feature selection and using various machine learning algorithms. 2. **Reduce the time complexity of the model**: Reduce the training and testing time of the model by selecting the optimal minimal feature set. 3. **Construct a balanced dataset**: Create a balanced dataset containing 420 different individuals, each with 30 signatures, to ensure the generalization ability of the model. ### Solutions: 1. **Data Collection and Preprocessing**: Collected 12,600 images, 30 signatures per individual, and performed preprocessing operations such as cropping and scaling. 2. **Feature Extraction**: Used the MobileNetV2 model to extract 1280 features from each image. 3. **Feature Selection**: Used three feature selection methods (Neighborhood Component Analysis NCA, Chi-Square Test Chi2, Mutual Information MI) to select 200, 300, 400, and 500 features respectively. 4. **Classifier Application**: Used various machine learning algorithms (SVM, KNN, Decision Tree DT, Linear Discriminant Analysis LDA, Naïve Bayes) for signature verification. ### Experimental Results: - **Without using feature selection**: The classification accuracy was 91.3%. - **Using NCA feature selection method**: The classification accuracy increased to 97.7% using only 300 features. ### Main Contributions: 1. **New Signature Dataset**: Created a balanced dataset containing 420 different individuals, each with 30 signatures. 2. **Optimized Feature Selection**: Improved the classification accuracy of the model and reduced the time complexity by selecting the optimal minimal feature set. 3. **High Classification Accuracy**: Achieved a classification accuracy of 97.7% using the NCA feature selection method across various machine learning algorithms. In summary, the paper significantly improves the accuracy and efficiency of offline handwritten signature verification by combining transfer learning and feature selection methods.

Offline Handwriting Signature Verification: A Transfer Learning and Feature Selection Approach

A Survey of Offline Handwritten Signature Verification Based on Deep Learning

Offline signature verification using deep neural network with application to computer vision

Offline Handwritten Signature Identification based on Hybrid Features and Proposed Deep Model

Offline Handwritten Signature Recognition Using Deep Convolution Neural Network

Handwritten Signature Identification Based on Hybrid Features and Machine Learning Algorithms

Offline Signature Recognition via Convolutional Neural Network and Multiple Classifiers

A Comparative Study among Handwritten Signature Verification Methods Using Machine Learning Techniques

Digital image processing: Offline handwritten signature identification using local binary pattern and rotational invariance local binary pattern with learning vector quantization

A multi-dimensional review on handwritten signature verification: strengths and gaps

DeepHSV: User-Independent Offline Signature Verification Using Two-Channel CNN

Offline Handwritten Signature Verification - Literature Review

Offline Signature Verification Based on Feature Disentangling Aided Variational Autoencoder

Offline Signature Verification on Real-World Documents

Signature Detection, Restoration, and Verification: A Novel Chinese Document Signature Forgery Detection Benchmark

Active Transfer Learning for Persian Offline Signature Verification

Attention-based Multiple Siamese Networks with Primary Representation Guiding for Offline Signature Verification

Off-Line Signature Verification Based On Multitask Learning

Offline Handwritten Signature Verification Using Deep Neural Networks

Multi-dilation Convolutional Neural Network for Automatic Handwritten Signature Verification

Off Line Handwritten Signature Verification Based on Feature Fusion