However, the quality of inked palmprint image is very poor, therefore, researchers’ interest later turned to online palmprint recognition. Zhang et al. proposed the first online low-resolution palmprint selleck kinase inhibitor recognition system, and published a palmprint image database, i.e., the PolyU database [5]. After that, research on palmprint recognition grew rapidly. In order to acquire low-resolution palmprint images, different devices were exploited. Ribaric et al. [11] used a digital scanner to collect palmprint images. Zhang et al. [5] and Sun et al. [12] developed CCD camera-based special devices for palmprint acquisition, respectively. Kumar et al. captured hand images using a digital camera [13]. In their works [5,11�C13], the palmprint images were captured in the contact manner.
Recently, there are more studies on contact-free palmprint recognition. Usually, web-cameras [14], cameras in smart phones, panel PCs, or notebook PCs were used to collect contact-free palmprint images.So far, many approaches have been proposed for low-resolution palmprint recognition. Kong et al. [6] made a survey of these approaches and divided them into several different categories such as texture based, palm line based, subspace learning based, orientation coding based, correlation based, local image descriptor based, and multi-feature based, respectively. From the literature [6], it can be seen that most research works have focused on feature extraction and matching. In order to improve the recognition performance, other strategies were exploited. For example, Zhang et al.
[15] proposed multi-spectral based palmprint recognition. Here, it should be noted that all of the previous studies of palmprint recognition only used one device to collect palmprint images. That is, the training set and test set were captured using a same device.In this paper, we investigate the problem of Palmprint Recognition Across Different Devices (PRADD), which has not been well studied so far. In fingerprint-based biometrics, the problem of biometric sensor interoperability has been investigated [16�C18]. Biometric sensor interoperability refers to the ability of a system to compensate for the variability introduced in the biometric data of an individual due to the deployment of different sensors [18]. From the literature [16�C18], Anacetrapib it can be seen that poor inter-sensor performance has been reported for fingerprint recognition.
With the wide applications of palmprint recognition and the popularization of all kinds of cameras, there is a high possibility that a person’s palmprint images would be captured by different devices. Therefore, the problem of PRADD http://www.selleckchem.com/products/ldk378.html needs to be carefully studied. The technique of PRADD has the following potential applications: (1) Remote enrollment in a palmprint based distributed biometrics system.