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.
Handwritten signature verification is a complex problem with many practical applications in industries such as finance, legal documentation, and security. Due to advancements in computer vision and machine learning, significant progress has been made in this field. However, there is still room for improvement in terms of accuracy and efficiency.
This article presents a suggested strategy for handwritten signature verification, which consists of four stages. The first stage involves collecting a large dataset of 12,600 images from 420 different individuals, with each individual providing 30 signatures of a certain kind. It is worth noting that all authors’ signatures used in the dataset are genuine, which ensures the reliability of the collected data.
In the second stage, a deep learning model called MobileNetV2 is employed to extract the best features from each image. This step is crucial as it helps to reduce the dimensionality of the data and extract meaningful and discriminative features for signature verification. Three feature selection techniques, namely neighborhood component analysis (NCA), Chi2, and mutual info (MI), are utilized to select the most relevant features.
The third stage involves applying various machine learning techniques such as SVM with different kernels (rbf, poly, and linear), KNN, DT, Linear Discriminant Analysis, and Naive Bayes. By utilizing these techniques, 12 different results are obtained, each corresponding to a specific combination of features and machine learning model.
Notably, the article highlights the importance of feature selection techniques in improving the classification accuracy of handwritten signature verification. Without employing any feature selection techniques, the proposed offline signature verification achieves a respectable classification accuracy of 91.3%. However, by using the NCA feature selection approach with just 300 features, an impressive accuracy of 97.7% is achieved.
This demonstrates that careful selection of features can significantly enhance the performance of the handwritten signature verification system. Furthermore, the article emphasizes that the suggested method is not only accurate but also self-organized, offering a flexible framework for validation prediction vectors.
In conclusion, this article provides valuable insights into the multi-disciplinary nature of handwritten signature verification, involving computer vision, deep learning, feature selection techniques, and various machine learning algorithms. The proposed strategy shows promising results in terms of accuracy, and it opens up possibilities for further research and advancements in the field. Future work could focus on exploring additional feature selection techniques and experimenting with different machine learning models to further enhance the performance of handwritten signature verification systems.
Read the original article