A Novel Intra-Class Distance-Based Signature Identification Algorithm Using Weighted Gabor Features and Dynamic Characteristics

作者：Tahmasebi Ava; Pourghassem Hossein^*

来源：Arabian Journal for Science and Engineering, 2013, 38(11): 3019-3029.

DOI：10.1007/s13369-012-0455-3

• Biometric
• Online signature
• Dynamic characteristic
• Gabor filter
• Distance measure
• Intra-class distance density distribution function

Biometric recognition systems have important roles in everyday activities. Online signature verification is one of the effective biometric solutions. Due to using dynamic characteristics of the online signature, it is more robust to copy problems. This paper presents a novel intra-class distance-based signature identification algorithm based on the combination of pressure and position features of an individual's signature. In this algorithm, the pressure feature is combined with the position feature as a weight of the signature image that is called weighted signature image. By utilizing the Gabor filter on the weighted signature images, local features are extracted using a blocking scheme. Then, a distance-based strategy is employed for the signature identification process. In this strategy, the optimum decision boundary of the distance measure is selected based on the genuine and imposter intra-class distance density distribution functions. The proposed algorithm is evaluated on an SVC 2004 database containing 1,600 signatures from 40 different users. The obtained results show the effectiveness and efficiency of our algorithm in comparison with other common approaches.

• 出版日期2013-11