Palmprint Matching Using Line Features

Abstract

making a model of a hand. Delta points and minutiae ABSTRACT only can be extracted from the fine-resolution images. Lines and texture are the most clearly observable Abstract -Palmprint-based personal identification features in low-resolution palmprint images (such as is regarded as an effective method for automatically 100 dpi). Principal lines and wrinkles, called line recognizing a person's identity. It is well-known for features, are more appealing than texture for the several advantages such as stable line features, low-human vision. When human beings compare two resolution imaging, low-cost capturing device, and user palmprint images, they instinctively compare line friendly. This paper presents a new approach to identify features. This action motivates us to develop a palmprint using modified line-based Hausdorff Distance. matching scheme for the palm lines.Line Edge Map (LEM) of palmprint is extracted as the A line contains various information including feature used for identifying. Length, position, and length, position, orientation, width, and magnitude. orientation information of the palm lines are fully Line position is often considered as an important utilized for palmprint matching. In contrast to the fe peciall for line-based asraches. It is existing methods, our approach achieves higher feature, especially for line-based approaches. It is accuracy. It is able to yield 100% identification worthwhile to note that no previous palmprint accuracy and 0% EER using the public palmprint research has investigated the length, position, and databse provided by PolyU.orientation information of the palm lines for

Joint Line and Orientation Features in 3D Palmprint

“…There are two main approaches to 2D palmprint recognition: line-based approach (Li et al 2006;Wu et al 2006;Huang et al 2008) and texture-based approach (Kong and Zhang 2004;Sun et al 2005). For the representative linebased methods, Li et al proposed a modified line-based Hausdorff Distance for palmprint identification (Li et al 2006); Wu et al proposed a set of directional line detectors and used them to extract the palm lines for palmprint matching (Wu et al 2006); Huang et al proposed a modified finite Radon transform to extract the principal lines in palmprint (Huang et al 2008).…”

“…For the representative linebased methods, Li et al proposed a modified line-based Hausdorff Distance for palmprint identification (Li et al 2006); Wu et al proposed a set of directional line detectors and used them to extract the palm lines for palmprint matching (Wu et al 2006); Huang et al proposed a modified finite Radon transform to extract the principal lines in palmprint (Huang et al 2008). With respect to the texture-based methods, the most representative one may be Competitive Coding (CompCode) scheme proposed by Kong and Zhang (2004), where a series of directional Gabor: filters were used to extract the orientation features of palmprint.…”