scispace - formally typeset
Search or ask a question

Showing papers on "Signature recognition published in 2020"


Journal ArticleDOI
TL;DR: A method for the pre-processing of signatures to make verification simple is proposed and a novel method for signature recognition and signature forgery detection with verification is proposed using Convolution Neural Network, Crest-Trough method and SURF algorithm & Harris corner detection algorithm.

54 citations


Proceedings ArticleDOI
01 Jun 2020
TL;DR: This approach presents a new technique for signature verification and recognition, using a tow dataset for training the model by a siamese network, and describes the ability of the suggested system in specifying the genuine signatures from the forgeries.
Abstract: Signatures are popularly used as a method of personal identification and confirmation Many certificates such as bank checks and legal activities need signature verification Verifying the signature of a large number of documents is a very difficult and time-consuming task As a result, explosive growth has been observed in biometric personal verification and authentication systems that relate to unique quantifiable physical properties (fingerprints, hand, and face, ear, iris, or DNA scan) or behavioral characteristics (gait, sound, etc) Several methods are used to describe the ability of the suggested system in specifying the genuine signatures from the forgeries This approach presents a new technique for signature verification and recognition, using a tow dataset for training the model by a siamese network

12 citations


Journal ArticleDOI
TL;DR: A novel procedure for online signature verification and recognition based on Dual-Tree Complex Wavelet Packet Transform (DT-CWPT) is presented, and favorable experimental results confirm the effectiveness of the presented method in both online signature verify and recognition objects.
Abstract: Background: With the increasing advancement of technology, it is necessary to develop more accurate, convenient, and cost-effective security systems. Handwriting signature, as one of the most popular and applicable biometrics, is widely used to register ownership in banking systems, including checks, as well as in administrative and financial applications in everyday life, all over the world. Automatic signature verification and recognition systems, especially in the case of online signatures, are potentially the most powerful and publicly accepted means for personal authentication. Methods: In this article, a novel procedure for online signature verification and recognition has been presented based on Dual-Tree Complex Wavelet Packet Transform (DT-CWPT). Results: In the presented method, three-level decomposition of DT-CWPT has been computed for three time signals of dynamic information including horizontal and vertical positions in addition to the pressure signal. Then, in order to make feature vector corresponding to each signature, log energy entropy measures have been computed for each subband of DT-CWPT decomposition. Finally, to classify the query signature, three classifiers including k-nearest neighbor, support vector machine, and Kolmogorov–Smirnov test have been examined. Experiments have been conducted using three benchmark datasets: SVC2004, MCYT-100, as two Latin online signature datasets, and NDSD as a Persian signature dataset. Conclusion: Obtained favorable experimental results, in comparison with literature, confirm the effectiveness of the presented method in both online signature verification and recognition objects.

10 citations


Proceedings ArticleDOI
18 Feb 2020
TL;DR: Experimental results verify the effectiveness of the models: VGG16 and SigNet for signature verification and the superiority of V GG16 in signature recognition task.
Abstract: Recently, deep convolutional neural networks have been successfully applied in different fields of computer vision and pattern recognition. Offline handwritten signature is one of the most important biometrics applied in banking systems, administrative and financial applications, which is a challenging task and still hard. The aim of this study is to review of the presented signature verification/recognition methods based on the convolutional neural networks and also evaluate the performance of some prominent available deep convolutional neural networks in offline handwritten signature verification/recognition as feature extractor using transfer learning. This is done using four pretrained models as the most used general models in computer vision tasks including VGG16, VGG19, ResNet50, and InceptionV3 and also two pre-trained models especially presented for signature processing tasks including SigNet and SigNet- F. Experiments have been conducted using two benchmark signature datasets: GPDS Synthetic signature dataset and MCYT- 75 as Latin signature datasets, and two Persian datasets: UTSig and FUM-PHSD. Obtained experimental results, in comparison with literature, verify the effectiveness of the models: VGG16 and SigNet for signature verification and the superiority of VGG16 in signature recognition task.

9 citations


Journal ArticleDOI
TL;DR: The aim of this thesis is to solve the face-detection in the first attempt using the Haar-cascade classifier from images containing simple and complex backgrounds by using a modified Haar cascade algorithm.
Abstract: Amid the previous three decades, the topic of image processing has gained vital name and recognition among researchers because of their frequent look in varied and widespread applications within the field of various branches of science and engineering. As an example, image processing is helpful to issues in signature recognition, digital video processing, remote sensing and finance. Image processing models are used for detecting the face. The aim of this thesis is to solve the face-detection in the first attempt using the Haar-cascade classifier from images containing simple and complex backgrounds. It is one of the preeminent detectors in terms of reliability and speed. We introduced a new method to deal with the frontal face images by using a modified Haar cascade algorithm. By using this algorithm, we can detect the image as well as the coordinates. The main attraction of this paper is to solve different types of images having one object, two objects, and three objects which can’t be solved by any of the existing methods but can be solved by our proposed method.

5 citations


Journal ArticleDOI
TL;DR: This research work was carried out in collaboration among all authors and author MR gave useful advices and managed the literature searches and wrote the first draft of the manuscript.
Abstract: This research work was carried out in collaboration among all authors. Author WS carried out the research work and wrote the protocol and author MR gave useful advices and managed the literature searches and wrote the first draft of the manuscript. Author ARW supervised, gave useful advices, comments, remarks and engagement through the learning process of the entire research. All authors read and approved the final manuscript.

4 citations


Journal ArticleDOI
TL;DR: A proposed strategy is compared with current technique by several performance metrics and the proposed LFNN technique efficiently recognize the face images and corresponding signature from the input databases than the existing technology.
Abstract: In biometrics, picking of the right methodology is a testing errand for recognition of a person. Due to the advantage of widely accepted identification, face, and signature-based biometric modality are selected as a significant pattern as compared with other modalities. Different Face and signature successions of a similar subject may contain varieties in determination, light, pose, facial appearances and sign position. These varieties add to the difficulties in planning a viable multimodal-based face and signature recognition algorithm. This paper proposed about the face and signature recognition method from a large dataset with the different pose and multiple features. Face recognition is the first stage of a system then the signature verification will be done. Here, data glove signaling means of signing process are taken into account to do signature verification system. Hence the proposed work have used Face, and the corresponding signature is detected from data glove signal patterns to features-level fusion for the verification system. The proposed Legion feature based verification method will be developed using four important steps like, i) Preprocessing, ii) feature extraction from face and data glove signals, iii) Legion feature based feature matching through Euclidean distance, iv) Legion feature Neural network (LFNN) fusion based on weighted summation formulae where two weights will be optimally found out using Legion optimization algorithm, vi) Recognition based on the final score. Finally, based on the feature library the face image and signature can be recognized. The comparability estimation is finished by utilizing least Euclidean separation fusion based LFNN to decide perceived and non-perceived images. Also, in a similar examination, a proposed strategy is compared with current technique by several performance metrics and the proposed LFNN technique efficiently recognize the face images and corresponding signature from the input databases than the existing technology.

4 citations


Journal ArticleDOI
20 Aug 2020
TL;DR: The proposed method is for extraction features (such as ;Energy, Contrast, Entropy, and Correlation) from Offline Signature Verification System and the results show that the (UTSig) and the combination of (NISDCC, CEDAR, SigComp2012).
Abstract: There has been challenging the pattern recognition that more attention needs to be paid to this area Offline Signature Verification (OSV), particularly when it is relied upon to popularize fully on the skillful frauds that are not accessible during the preparation. Its difficulties additionally incorporate little training tests and great intra-class divergence. At times the crude signature can incorporate additional pixel known as noises or may not be in the legitimate structure where preprocessing is obligatory. Insomuch as a signature is preprocessed accurately, it leads to a superior outcome for both signature matching and fraud disclosure. For example; an appropriate estimation of gamma value improves the contrast of the signature image, on another hand, Pre-preparing likewise comprises binarization, noise elimination, so forth...The proposed method is for extraction features (such as ;Energy, Contrast, Entropy, and Correlation) from Offline Signature Verification System. In this paper, the data processing deals with twain parallel styles viz signature training and signature testing analysis. Insomuch as that the extracted features from a signature picture doesn't powerful, this will cause higher verification error rates particularly for skillful fabrications in hacking the system. The results show that’s the (UTSig) and the combination of (NISDCC, CEDAR, SigComp2012).Comparing with the other researches, the results in this Paper is the best and the system is more efficient with (UTSig) signature which were 97%.

4 citations


Proceedings ArticleDOI
14 Sep 2020
TL;DR: This study tries to extract, investigate and compare the use of static and dynamic features using Five well-known classifiers such as AutoMLP, Naïve Bayes, Neural Network, Gradient Boosted Trees and Generalized Linear Model.
Abstract: Handwritten signatures are commonly used for the signification, authentication, and validation of people's important transactions and documents. However, this security measure could also be a threat at the same time. This study aims to develop a model to detect and recognize a signature. This study also tries to extract, investigate and compare the use of static and dynamic features using Five well-known (5) classifiers such as AutoMLP, Naive Bayes, Neural Network, Gradient Boosted Trees and Generalized Linear Model. The classifier that shows the most acceptable model is Neural Network which shows an accuracy rate of 92.88%.

3 citations


Journal ArticleDOI
TL;DR: A new methodology for offline handwritten signature verification and recognition based on the Shearlet transform and transfer learning is proposed to verify the effectiveness of the presented method in both signature verification
Abstract: Despite the growing growth of technology, handwritten signature has been selected as the first option between biometrics by users. In this paper, a new methodology for offline handwritten signature verification and recognition based on the Shearlet transform and transfer learning is proposed. Since, a large percentage of handwritten signatures are composed of curves and the performance of a signature verification/recognition system is directly related to the edge structures, subbands of shearlet transform of signature images are good candidates for input information to the system. Furthermore, by using transfer learning of some pre-trained models, appropriate features would be extracted. In this study, four pre-trained models have been used: SigNet and SigNet-F (trained on offline signature datasets), VGG16 and VGG19 (trained on ImageNet dataset). Experiments have been conducted using three datasets: UTSig, FUM-PHSD and MCYT-75. Obtained experimental results, in comparison with the literature, verify the effectiveness of the presented method in both signature verification and signature recognition.

2 citations


Proceedings ArticleDOI
10 Nov 2020
TL;DR: In this article, the authors proposed a model of handwritten signature, containing fuzzy features of curvature of discrete handwritten signature functions, and proposed handwritten signature reference template creation algorithm, characterized by the use of the potential method for constructing membership functions of fuzzy features.
Abstract: In this paper we proposed the approach for dynamic handwritten signatures recognition. We proposed a formal model of the handwritten signature, containing fuzzy features of curvature of discrete handwritten signature functions. We proposed handwritten signature reference template creation algorithm, characterized by the use of the potential method for constructing membership functions of fuzzy features. The choice of a rational set of features has been implemented, which allows to minimize the false accept rate (up to 0.05%), as well as a rational set that minimizes the equal error rate (up to 0.36%), which significantly exceeds the efficiency of existing handwritten signature recognition algorithms.


Book ChapterDOI
06 Sep 2020
TL;DR: In this article, the authors proposed an approach for recognition handwritten signatures based on dynamics characteristics and fuzzy sets theory, which made it possible to recognize handwritten signatures even in case of their fuzzy character.
Abstract: In this paper we proposed the approach for recognition handwritten signatures based on dynamics characteristics and fuzzy sets theory. We suggested the formal model of human handwritten signature, which contains several fuzzy features. We also suggested handwritten signature reference template creation and handwritten signature recognition algorithms. Suggested formal model was used as a basis. Method of potentials was used to obtain membership functions of fuzzy features which makes it possible to create a reference template even under conditions of a small capacity of training set. The research was conducted on the MCYT_Signature_100 signature collection which includes 2500 genuine signatures and 2500 skillful fakes. Suggested approach makes it possible to recognize handwritten signatures even in case of their fuzzy character. Accuracy evaluation results showed FAR value 2.8%, FRR value 0.4% for random fake patterns and FRR value 0.8% for skillful fake patterns, which is better than evaluation results of many others approaches.

Book ChapterDOI
25 Nov 2020
TL;DR: In this paper, the authors proposed an offline signature recognition system using a multiple algorithm approach using Local Binary Pattern (LBP) and Grey Level Co-occurrence Matrix (GLCM).
Abstract: Signature is a biometric trait that has piqued the interest of researchers. This is due to its high rate of acceptability. Offline signature in particular, has been around for a while and hence its suitability as a biometric trait. This paper proposes an offline signature recognition system using a multiple algorithm approach. The system accepts handwritten signature, filters the signature and crops the signature region. The Local Binary Pattern (LBP) of the signature image is then obtained. After this, Grey Level Co-occurrence Matrix (GLCM) is applied. Statistical features are then extracted. The difference in the stored features and the extracted features was obtained. The output is compared with a threshold for discrimination. This research aims at improving the performance of offline signature recognition using its textural features. The designed system gave an FRR and FAR of 8.6%, 4.6% respectively for MYCT signature database and 8.8%, 5.2% for GPDS signature database.

Book ChapterDOI
01 Jan 2020
TL;DR: A signature recognition systembased in a multi-section vector quantization with a handwriting text recognition system based in self-organizing maps and DTW and the application of a logarithmic transformation for signature scores previous normalize them is proposed.
Abstract: In this paper, we analyze the combined application of signatures and capital handwriting in a biometric recognition application. We combine a signature recognition system based in a multi-section vector quantization with a handwriting text recognition system based in self-organizing maps and DTW. Due to the need to normalize the scores before the combination, we study the effect of different normalization methods and we propose the application of a logarithmic transformation for signature scores previous normalize them. Experimental results show that the identification rate raises from 86.11% using capital letter words and 96.95% using signatures up to 99.72% with a fusion of both traits. Minimum detection cost function (DCF) also improves, from 3.56 and 3.51%, respectively, up to 1.0% using the fusion of both traits.


Proceedings ArticleDOI
26 May 2020
TL;DR: The results of the studies of the correlation-extreme methods for signature recognition and methods for compression of hyperspectral images (HSI) with controlled losses depending on the templates selection technique, on the threshold value, onThe type of threshold used and on the selected type of invariant transformation are presented.
Abstract: This paper presents the results of the studies of the correlation-extreme methods for signature recognition and methods for compression of hyperspectral images (HSI) with controlled losses depending on the templates selection technique, on the threshold value, on the type of threshold used and on the selected type of invariant transformation. Recommendations are formulated for the selection of the said parameters in order to more effectively form a set of templates. Experiments have been carried out on real HSI fragments.

Patent
21 Apr 2020
TL;DR: In this article, a service verification method, device and system based on handwritten signature recognition is presented, which comprises the steps that user information and a first handwritten signature image input by a user in the first service operation execution process are collected; querying a corresponding second handwritten signature images from a handwritten signature library according to the user information, storing user information acquired in advance in a second service operation, and comparing the first handwritten signatures with the corresponding handwritten signatures, and determining whether the user executing the first signature image and the second signature image are the same user or not according to a comparison
Abstract: The invention discloses a service verification method, device and system based on handwritten signature recognition. The method comprises the steps that user information and a first handwritten signature image input by a user in the first service operation execution process are collected; querying a corresponding second handwritten signature image from a handwritten signature image library according to the user information, the handwritten signature image library storing user information acquired in advance in a second service operation execution process and the corresponding handwritten signature image; comparing the first handwritten signature image with the second handwritten signature image, and determining whether the user executing the first service operation and the user executing the second service operation are the same user or not according to a comparison result; and when the user executing the first service operation and the user executing the second service operation are the same user, continuing to execute the first service operation. Whether the service operation is the operation of the user or not is verified based on the handwritten signature, and the safety of theservice system can be greatly improved.


26 Apr 2020
TL;DR: This work presents a system for offline signature verification, where the user has to submit a number of signatures that are used to extract two types of features, statistical features and structural features to train propagation neural network in its verification stage.
Abstract: Biometrics refer to the system of authenticating identities of humans, using features like retina scans, thumb and fingerprint scanning, face recognition and also signature recognition. Signatures are a simple and natural method of verifying a person’s identity. It can be saved as an image and verified by matching, using neural networks. Signature verification can be offline or online. In this work, we present a system for offline signature verification. The user has to submit a number of signatures that are used to extract two types of features, statistical features and structural features. A vector obtained from each of them is used to train propagation neural network in the verification stage. A test signature is then taken from the user, to compare it with those the network had been trained with. A test experiment was carried out with two sets of data. One set is used as a training set for the propagation neural network in its verification stage. This set with four signatures form each user is used for the training purpose. The second set consists of one sample of signature for each of the 20 persons is used as a test set for the system. A negative identification test was carried out using a signature of one person to test others’ signatures. The experimental results for the accuracy showed excellent false reject rate and false acceptance rate.

Journal Article
TL;DR: A deep learning approach for offline signature verification to prevent the fraud signatures by fake peoples is presented and deep learning with the help of Convolution Neural Network is done.
Abstract: Now a day’s signature is becomes a most important biometric authentication technique. In banks or at the other necessary documents, signature plays an important role to authenticate the person. In this technique, we are going to present a deep learning approach for offline signature verification to prevent the fraud signatures by fake peoples. We are going to do deep learning with the help of Convolution Neural Network (CNN). In this study, we are going to collect dataset of different signatures from the different angles. Signature is taken as an input in the form of image. For signature recognition, it is important to make structural and some geometrical calculation getting to extract special features from the signatures then we train a man-made neural network on these features from different signers. Finally, the extracted features from the tested signature are compared with the previously trained features and that we know the signer.

Patent
05 Mar 2020
TL;DR: In this article, methods for semantic processing of data files including detecting formats of data embedded in the data files and converting the data to formats compatible with a data analysis tool are presented.
Abstract: Methods are provided for semantic processing of data files including detecting formats of data embedded in the data files and converting the data to formats compatible with a data analysis tool. The method may comprise determining if the data file comprises signature characteristics associated with a known data format and, if so, determining a set of data manipulation operations associated with the known data format to convert the data file to a compatible format for the data analysis tool. The method may further comprise semantically analyzing components of the data files to assess formatting across a required set of criterions needed by the data analysis tool and determining sets of data manipulation operations to perform to convert the data file to a compatible format.

Patent
28 Feb 2020
TL;DR: In this article, a handwritten signature recognition method based on acoustics is proposed, which consists of collecting data, calculating a signature track, and comparing the signature track with an identification model, and judging the authenticity of the to-be-verified data.
Abstract: The invention relates to the field of voice recognition technology and action signal recognition technology, in particular to a handwritten signature recognition method based on acoustics The methodmainly comprises the following steps of: S1, collecting data; S2, calculating a signature track; S3, performing feature extraction and model training; S4, performing data collection on a to-be-verified user, calculating a signature track, performing feature extraction, and taking extracted features as to-be-verified data; and S5, comparing the to-be-verified data with an identification model, andjudging the authenticity of the to-be-verified data According to the invention, only common acoustic components in existing intelligent equipment are utilized and can be called in software level withno need for extra hardware support; and moreover, the method is convenient to use, can accurately identify legal signatures and illegal signatures, can reduce the cost of handwritten signature recognition well, can solve problems such as inconvenience in use and is very wide in application prospect

Book ChapterDOI
14 Oct 2020
TL;DR: In this article, a signature detection algorithm and its subsequent signature identification using a deep learning model for processing images based on a convolutional neural network was presented, where a binary classification has been performed to predict text or signature and signature classifications to determine the author of this signature.
Abstract: The purpose of the article is to present a simple signature detection algorithm and its subsequent signature identification using a deep learning model for processing images based on a convolutional neural network. To solve the task of the image recognition, a binary classification has been performed to predict text or signature and signature classifications to determine the author of this signature. The proposed algorithm is interesting in the preliminary processing of scanned documents with signatures in order to extract the area with the signature and transfer it to the trained model. The research results are presented for documents of the same type, in which the signature is located in the same place. To select a specific element in the document we are using the tensor-slicing operations on Numpy arrays. To extract areas with text and signature, OpenCV tools are used. The results on the ready-made neural network model studies on a small dataset are presented in this article. Good results have been achieved in recognizing the famous writers’ signatures. The proposed algorithm demonstrates the possibility of using the classical convolution network model for solving specific practical problems. The studies can be recommended to students in the study of neural networks to understand the basics of deep learning and apply a ready-made model as a template for solving practical problems in the field of computer vision.

Dissertation
01 Jun 2020
TL;DR: In this paper, a new flexible enhanced fuzzy min-max (FEFMM) model is proposed to overcome limitations related to accuracy issue and four new procedures are introduced: a new training strategy to avoid generating unnecessary overlapped regions.
Abstract: In the attempts of building an efficient classifier model, various hybrid computational intelligence models have been introduced. Among these, the enhanced fuzzy min-max (EFMM) model was one of the most recent models coming with many essential features like the ability to provide online learning processes and handling the forgetting problem. Although EFMM has been proven to be one of the most premier models for undertaking the pattern classification problems, issues related to its learning process, concerning the overlap between the hyperboxes, random expansion coefficient value (user-defined) and hyperbox contraction remain unsolved. Therefore, two stages of improvements are introduced in this research to overcome the current limitations and improve classification performance in terms of accuracy and complexity. In the first stage, a new flexible enhanced fuzzy min-max (FEFMM) model is proposed to overcome limitations related to accuracy issue. Hence, four new procedures are introduced. First, a new training strategy to avoid generating unnecessary overlapped regions. Second, a new flexible expansion procedure to replace the expansion coefficient user-defined parameter with a self-adaptive value to produce more accurate decision boundaries. Third, a new overlap test rule is applied during the testing phase to identify any possible containment overlap case and activate the contraction process (if necessary). Fourth, a new contraction procedure to overcome the containment overlap and avoiding the data distortion problem (missing hyperbox information). In the second stage, a new pruning strategy is proposed to further enhance the performance of the proposed model in regards to overcome the network complexity problem. Hence, the resulting model is known as FEFMM-based pruning strategy (FEFMM-PS). The usefulness of both stages is evaluated systematically using a series of experiments using several benchmark datasets. Sixteen data sets are used in the evaluation process. These data sets are obtained from the UCI machine learning repository and the selection of these data sets is related to cover examples of different levels of difficulties, input and output classes, features, and a number of instances. The performance of FEFMM-PS in these experiments are then quantified using statistical measures where the bootstrap and k-fold cross-validation methods have been adopted. The results demonstrate the efficiency of FEFMM in handling pattern classification problems and providing a superior performance of classification accuracy as compared to the other network structures from the same variants such as EFMM, FMM variants and also non-FMM related models. Concerning the FEFMM-PS, the finding reveals that the model (FEFMM-PS) is able to solve network complexity problem and presents better classification accuracy as compared to FEFMM and other models from the literature. The proposed models FEFMM and FEFMM-PS can be applied in several application areas to further assess their applicability, such as face recognition, speaker recognition, signature recognition, and text classification.

Book ChapterDOI
12 Dec 2020
TL;DR: In this paper, the authors proposed different methods or techniques that help in increasing efficiency and gives the effectiveness of the overall system, which is one of the most extensively used biometric traits for verification of certain individuals.
Abstract: Attestation plays a crucial role to manage surveillance. So, need for authentication increases briskly. Because of the latest improvements in technology in a time where data rules everything, there is a high priority for security systems based on different biometric traits. Signature is one of the most extensively used biometric traits for verification of certain individuals. Signature plays an important role in different areas of application such as finance, banking, commercial, etc. However, in many cases, forged signatures may also be considered as original Signatures. Signatures are behavioral biometric traits and are mostly unique for each individual. Therefore, the signature verification system is one of the crucial systems that have to be produced with higher speeds and provide higher accuracy. Here in this paper, we propose different methods or techniques that help in increasing efficiency and gives the effectiveness of the overall system. We propose a method for each stage of the verification process and compare them with other alternative methods available.

Proceedings ArticleDOI
09 Dec 2020
TL;DR: In this paper, a convolutional neural network is trained on preprocessed signature images and tested on four different datasets with N number of individuals and M number of signatures for each individual and contains signatures that differ from each other in many aspects like the type of signature, its readability, etc.
Abstract: Handwritten signature is one of the essential biometric parameters widely used for document validation and verification. Other methods such as fingerprints, iris/retina scanning, face, and voice recognition, although more accurate, need special equipment. The purpose of the research is to demonstrate an appropriate and reliable technology organizations may use to recognize signatures automatically. Convolutional neural networks are trained on preprocessed signature images. The code was developed using MATLAB, and results indicate our method to provide promising results and have contributed by extending the technique to be reliable. The CNN is tested with 4 different datasets with N number of individuals and M number of signatures for each individual and contains signatures that differ from each other in many aspects like the type of signature, its readability, etc. We used our CNN to train and test on all the datasets to observe the performance and make interesting observations of our implementation. The network performed reasonably well on all datasets, which is presented in the results section.

Proceedings ArticleDOI
03 Sep 2020
TL;DR: This paper’s main objective is to design and implement a signature recognition system on a single board computer, i.e., Raspberry Pi 3 equipped with an LCD touchscreen, and showed that the recognition rate of 99.77% was achieved using a confidence level threshold of 85% during testing.
Abstract: Signatures play a crucial role in human life as part of their identity. Nowadays, there is a growing interest in the smart home system using the Internet of Things (IoT). Furthermore, signature recognition and verification can play essential roles in finance, banking, home system, insurance, and others. This paper’s main objective is to design and implement a signature recognition system on a single board computer, i.e., Raspberry Pi 3 equipped with an LCD touchscreen. First, the acquired signature image was cropped and resized. Next, a binary image was extracted as features to train the artificial neural network (ANN). The trained ANN was used to classify the input signature to determine whether the signature is genuine or forged. Results showed that the recognition rate of 99.77% was achieved using a confidence level threshold of 85% during testing.