Signature recognition

About: Signature recognition is a research topic. Over the lifetime, 2138 publications have been published within this topic receiving 37605 citations.

A Watermarking Technique Using Discrete Curvelet Transform for Security of Multiple Biometric Features

Abstract: The robustness and security of the biometric watermarking approach can be improved by using a multiple watermarking. This multiple watermarking proposed for improving security of biometric features and data. When the imposter tries to create the spoofed biometric feature, the invisible biometric watermark features can provide appropriate protection to multimedia data. In this paper, a biometric watermarking technique with multiple biometric watermarks are proposed in which biometric features of fingerprint, face, iris and signature is embedded in the image. Before embedding, fingerprint, iris, face and signature features are extracted using Shen-Castan edge detection and Principal Component Analysis. These all biometric watermark features are embedded into various mid band frequency curvelet coefficients of host image. All four fingerprint features, iris features, facial features and signature features are the biometric characteristics of the individual and they are used for cross verification and copyright protection if any manipulation occurs. The proposed technique is fragile enough; features cannot be extracted from the watermarked image when an imposter tries to remove watermark features illegally. It can use for multiple copyright authentication and verification.

Chinese Handwritten Legal Amount Recognition with HMM-Based Approach

Abstract: A hidden Markov model (HMM) based method for Chinese legal amount recognition is presented in this paper. In the training phase, gradient feature is extracted from sliding windows and character HMMs are trained with single character images. In the recognition phase, the text line image is segmented using sentence HMM, which is constructed by character HMMs according to a strict language model. The main difference between our proposed method and traditional methods is that our segmentation is guided by language model, which can solve many tough segmentation problems. Moreover, we combine the HMM-based method with traditional OCR method to improve the recognition accuracy. Experiments have been performed on 4,709 legal amount text line images extracted from real-life bank checks. The recognition rate is 97.13%.

Three-dimensional human face recognition

Abstract: In this communication, we propose a technique for three-dimensional (3D) human face recognition. The 3D shape information of the faces is utilized to synthesize two-dimensional spatial functions, called the signature functions. Face recognition task is completed by carrying out cross-correlation between the signature functions of the faces and analyzing the correlation peaks. High correlation peak signifies true class recognition and low or no peak signifies false class rejection. Preliminary experimental results are presented that demonstrate the feasibility.

GAIT RECOGNITION USING SVM and NN

Abstract: Recognition of any individual is a task to identify people. Human recognition methods such as face, fingerprints and iris generally require a cooperative subject and close proximity or physical contact. These methods are not able to recognize an individual at a distance therefore recognition using gait is relatively new biometric technique without these disadvantages. Human identification using Gait is method to identify an individual by the way he walk or manner of moving on foot of human. Gait recognition is a type of biometric recognition and related to the behavioural characteristics of biometric recognition. Gait offers ability of distance recognition. Different parameters are used such as distance between head and pelvis and distance between feet and one another additional parameter used by us. However the majority of current approaches are model free which is simple and fast but we will use model based approach for feature extraction and for matching of parameters with database sequences. After matching of parameters CCR (Correct Classification Rate) will be obtained using SVM, NN technique. Some experimental results will show the effectiveness of proposed system.

Pixel Based Off-line Signature Verification System

Abstract: The verification of handwritten signatures is one of the oldest and the most popular authentication methods all around the world. As technology improved, different ways of comparing and analyzing signatures become more and more sophisticated. Since the early seventies, people have been exploring how computers can fully take over the task of signature verification and tried different methods. However, none of them is satisfactory enough and time consuming too. Therefore, our proposed pixel based off- line signature verification system is one of the fastest and easiest ways to authenticate any handwritten signature we have ever found. For signature acquisition, we have used scanner. Then we have divided the signature image into 2D array and calculated the hexadecimal RGB value of each pixel. After that, we have calculated the total percentage of matching. If the percentage of matching is more than 90, the signature is considered as valid otherwise invalid. We have experimented on more than 35 signatures and the result of our experiment is quite impressive. We have made the whole system web based so that the signature can be verified from anywhere. The average execution time for signature verification is only 0.00003545 second only.