Signature recognition

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

Using speaker group dependent modelling to improve fusion of fragmentary classifier decisions

Abstract: Current speech-controlled human computer interaction is purely based on spoken information. For a successful interaction, additional information such as the individual skills, preferences and actual affective state of the user are often mandatory. The most challenging of these additional inputs is the affective state, since affective cues are in general expressed very sparsely. The problem can be addressed in two ways. On the one hand, the recognition can be enhanced by making use of already available individual information. On the other hand, the recognition is aggravated by the fact that research is often limited to a single modality, which in real-life applications is critical since recognition may fail in case sensors do not perceive a signal. We address the problem by enhancing the acoustic recognition of the affective state by partitioning the user into groups. The assignment of a user to a group is performed at the beginning of the interaction, such that subsequently a specialized classifier model is utilized. Furthermore, we make use of several modalities, acoustics, facial expressions, and gesture information. The combination of decisions not affected by sensor failures from these multiple modalities is achieved by a Markov Fusion Network. The proposed approach is studied empirically using the LAST MINUTE corpus. We could show that compared to previous studies a significant improvement of the recognition rate can be obtained.

Face and ear fusion recognition based on multi-agent

Abstract: Data fusion is one of the most important problems in current image processing field. Non-invasive characteristic of ear and profile face recognition contrary to other biometric recognition, unique ear features and ubiety about face and ear of 3D human head ensure the feasibility for fusing face and ear recognition. A new approach in decision fusion is proposed, the method uses less data than other fusion and has a faster recognition rate. The fusion based on face and ear recognition is a meaningful attempt to explore a novel method of biometric recognition. Eyes are parallel to the recognition process of facial patterns, but actual computer architecture is serial. At present, multi-biometrics authentication systems have not a uniform frame construction. The process of 3D face recognition is described by using recent multi-agent system theory for the first time. A multi-agent face recognizing structure model (MAFRSM) for 3D face recognition is proposed. Experiment data show that the MAFRSM can effectively enhance 3D face recognition rate.

Online zones' identification using signature baseline

Abstract: Signature zones' identification has been used in signature recognition and verification. The identification of an offline signature requires the whole image of the signature to be processed without considering other features in the signature. One of the signature features that are frequently used as the precondition for other subsequent algorithms in signature recognition and verification is the baseline. Since the representation of the offline signature is different than the online signature, a different approach is used to identify the online signature zones. This approach takes the information obtained from the baseline as a consideration for developing the zones' identification algorithms. The results shows that the algorithms developed have 85.73% correctness by average based on the zones definition and 37.18% accuracy by average compared to the opinion of a graphologist.

Online Signature Recognition Based on Pseudo-Inked Signature Image Template

Abstract: As human–robot interaction is widely and increasingly used, automated user verification has become a necessary condition for system access. Signature recognition is one of the representative methods for user verification. In this paper, a novel method using Pseudo-Inked Signature for online signature recognition is proposed. Pseudo-Inked Signature consists of three types of information of pen pressure value, pen tilting angle, and pen theta angle during online signature writing. We propose a fusion method for three different types of information by mimicking the inked effect of real pen writing. Besides a style of penmanship, Pseudo-Inked Signature reflects the characteristics of handwriting behavior. Therefore, it can make different Pseudo-Inked Signature even though the original signature images from different users look very similar to each other. Similarly, it can also make more similar Pseudo-Inked Signatures even though the original signature images from the same user look somewhat different to each other. In addition, since only one gray-scale image is dealt with to represent the signature style of a person by Pseudo-Inked Signature image, it is efficient and very easy to handle. Finally, we tested user verification experiments using k-NN classifier. The experimental results show that Pseudo-Inked Signature is good enough for the real application.