Handwritten Signature Verification Using CNN with Data Augmentation

Rabbi, Md. Tariqulhasan Fazle; Rahman, Shafizur; Biswash, Prokash; Kim, Jinsul; Sheikh, A. Q.; Saha, Aloke Kumar; Uddin, Mohammad Shorif

doi:10.9728/jcc.2019.12.1.1.25

Cited by 3 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rabbi et al [9] proposed a study on offline handwritten signature verification using CNN. They used data augmentation with the CNN model and also presented a comparative study with Multilayer Perceptron (MLP) and Single-Layer Perceptron (SLP).…”

Section: Related Workmentioning

confidence: 99%

Evaluation of Classification Performance of New Layered Convolutional Neural Network Architecture on Offline Handwritten Signature Images

Ozkan,

Erdogmus

2024

Symmetry

View full text Add to dashboard Cite

While there are many verification studies on signature images using deep learning algorithms in the literature, there is a lack of studies on the classification of signature images. Signatures are used as a means of identification for banking, security controls, symmetry, certificates, and contracts. In this study, the aim was to design network architectures that work very fast in areas that require only signature images. For this purpose, a new Si-CNN network architecture with existing layers was designed. Afterwards, a new loss function and layer (Si-CL), a novel architecture using Si-CL as classification layer in Si-CNN to increase the performance of this architecture, was designed. This architecture was called Si-CNN+NC (New Classification). Si-CNN and Si-CNN+NC were trained with two datasets. The first dataset which was used for training is the “C-Signatures” (Classification Signatures) dataset, which was created to test these networks. The second dataset is the “Cedar” dataset, which is a benchmark dataset. The number of classes and sample numbers in the two datasets are symmetrical with each other. To compare the performance of the trained networks, four of the most well-known pre-trained networks, GoogleNet, DenseNet201, Inceptionv3, and ResNet50, were also trained with the two datasets with transfer learning. The findings of the study showed that the proposed network models can learn features from two different handwritten signature images and achieve higher accuracy than other benchmark models. The test success of the trained networks showed that the Si-CNN+NC network outperforms the others, in terms of both accuracy and speed. Finally, Si-CNN and Si-CNN+NC networks were trained with the gold standard dataset MNIST and showed superior performance. Due to its superior performance, Si-CNN and Si-CNN+NC can be used by signature experts as an aid in a variety of applications, including criminal detection and forgery.

show abstract

Section: Related Workmentioning

confidence: 99%

Evaluation of Classification Performance of New Layered Convolutional Neural Network Architecture on Offline Handwritten Signature Images

Ozkan,

Erdogmus

2024

Symmetry

View full text Add to dashboard Cite

show abstract

“…We use the data augmentation process during the training stage to add sample signature variants to the network during the training process. Data augmentation is modifying image data so that the computer recognizes it as a different image [6], [26], [38]. Otherwise, it will be perceived as the same image object.…”

Section: B Data Pre-processing and Augmentationmentioning

confidence: 99%

“…The results show that the classification performance after using deep learning-based CNN increases by 20%, so CNN is very effective in obtaining the image information needed in data classification. Rabbi et al [26] compared several deep learning signature identification methods, including CNN, CNN with data augmentation, MLP (Multilayer Perceptron), and Single Layer Perceptron (SLP). The dataset is derived from a private dataset of 30 individuals, each with 24 authentic and 24 forged signatures.…”

Section: Introductionmentioning

confidence: 99%

CNN with Batch Normalization Adjustment for Offline Hand-written Signature Genuine Verification

Fatihia

Fariza

Karlita

2023

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

Signature genuine verifications of offline hand-written signatures are critical for preventing forgery and fraud. With the growth of protecting personal identity and preventing fraud, the demand for an automatic system for signature verification is high. The signature verification system is then studied by many researchers using various methods, especially deep learning-based methods. Hence, deep learning has a problem. Deep learning requires much training time for the data to obtain the best model accuracy result. Therefore, this paper proposed a CNN Batch Normalization, the CNN architectural adaptation model with a normalization batch number added, to obtain a CNN model optimization with high accuracy and less training time for offline hand-written signature verification. We compare CNN with our proposed model in the experiments. The research method in this study is data collection, pre-processing, and testing using our private signature dataset (collected by capturing signature images using a smartphone), which becomes the difficulties of our study because of the different lighting, media, and pen used to sign. Experiment results show that our model ranks first, with a training accuracy of 88.89%, an accuracy validation of 75.93%, and a testing accuracy of 84.84%—also, the result of 2638.63 s for the training time consumed with CPU usage. The model evaluation results show that our model has a smaller EER value; 2.583, with FAR = 0.333 and FRR = 4.833. Although the results of our proposed model are better than basic CNN, it is still low and overfitted. It has to be enhanced by better pre-processing steps using another augmentation method required to improve dataset quality.

show abstract

Node Pruning-based Compact Deep Convolutional Neural Network Model for Object Classification in Images

Oh¹,

Oh²,

Kim³

et al. 2022

dcs

View full text Add to dashboard Cite

Handwritten Signature Verification Using CNN with Data Augmentation

Cited by 3 publications

References 0 publications

Evaluation of Classification Performance of New Layered Convolutional Neural Network Architecture on Offline Handwritten Signature Images

Evaluation of Classification Performance of New Layered Convolutional Neural Network Architecture on Offline Handwritten Signature Images

CNN with Batch Normalization Adjustment for Offline Hand-written Signature Genuine Verification

Node Pruning-based Compact Deep Convolutional Neural Network Model for Object Classification in Images

Contact Info

Product

Resources

About