Fengxu Guan

Learn More
To carry out the finger vein recognition effectively and quickly, according to the characteristics of horizontal and vertical two-dimensional linear discriminant analysis ((2D)<sup>2</sup>LDA) reducing the dimensions, an algorithm of finger vein recognition is proposed. The algorithm is two direction weighted (2D)<sup>2</sup>LDA ((W2D)<sup>2</sup>LDA) based(More)
Based on the characteristics of wavelet transform (WT), wavelet moment (WM), horizontal and vertical twodimensional principal component analysis ((2D)PCA), a new method of finger vein recognition is proposed. Firstly, the original images are decomposed into high-frequency and lowfrequency components through WT, and the wavelet moment is extracted. Secondly(More)
Due to the restriction of acquisition equipment and other reasons, the finger vein line appeared broken, torsion, translation and other deformation in local area. Therefore, the traditional methods got low recognition accuracy. In this paper, combining the advantages of modular method and bi-directional weighted B2DPCA with eigenvalue normalization; the(More)
In this paper, 3D sound localization is presented by fusing audio information with video signals. This localization system consists of two subsystems, an audio setup with 3 microphones in a triangular format and a video source (a camera), all mounted on a pan-tilt unit. The audio subsystem is to generate two independent time differences(More)
The advantages of finger vein are in vivo identification, internal features, non-contact and security, but it's identified features are insufficient. Finger crease is in the surface of finger skin, and it has some stable biological features. So a finger has two kinds of biological features. Combined with the advantages and the disadvantages of finger vein(More)
In order to carry on multi target recognition quickly and effectively, according to the characteristics of SIFT, PCASIFT and SURF algorithm, a method of multi target recognition based on SURF algorithm and OpenCV is studied. The method includes multi angles recognition and multi targets recognition. Firstly, feature points were extracted single target or(More)
  • 1