Palmprint identification using restricted fusion

Jia, Wei; Liu, Bin; Chau, Kwok Wing

doi:10.1016/j.amc.2008.05.024

Cited by 30 publications

(9 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, a special kind of information is obtained by fusing two or more types of information mentioned above. From this perspective, the available methods [2][3][4][5][6][7][8][9][10][11][12][13][14][15] for palmprint recognition can be divided into five categories [1] on the basis of the extracted features:…”

Section: і Introductionmentioning

confidence: 99%

Palmprint recognition by fusion of multi-color components

Zhang

Sun

Qiu

et al. 2009

2009 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery

View full text Add to dashboard Cite

Palmprint Recognition Systems (PRS) with multi-feature can increase the recognition performance of PRS. For such a purpose, effectively distinguished information with low dimension is important in fusion for palmprint recognition. Usually, color information is a beneficial feature in image retrieval. But it is always ignored in recognition. This paper proposes a novel palmprint recognition algorithm based on multi-color components fusion. Firstly, an improved DCT-based approach is used to extract color component features from RGB space, YIQ space and HSI space, respectively. Secondly, the extracted useful component features are fused to serial-fused feature vectors. Then, classification is performed by the nearest neighbor classifier. Experimental results show that the proposed fusion algorithm obtains higher recognition rates compared to single component feature. Keywords-palmprint discrete cosine transform feature fusion multi-color components І. INTRODUCTIONWith the rapid development of biometric, palmprint recognition attracts much attention due to its many advantages such as large information and high consumers' acceptability. Current recognition approaches mainly use four types of palmprint information, which are texture information, line information, appearance information and orientation information. In addition, a special kind of information is obtained by fusing two or more types of information mentioned above. From this perspective, the available methods [2][3][4][5][6][7][8][9][10][11][12][13][14][15] for palmprint recognition can be divided into five categories [1] on the basis of the extracted features: Texture-based approaches, e.g., 2D-Gabor filter [2], discrete Fourier transform [3], wavelet transform [4] and radon transform [5], etc; Line-based approaches, e.g., directional line detectors [6], sober operator [7], radon transform[8] and multi-resolution filter [9], etc; Appearance-based approaches, e.g., principle component analysis (PCA) [10], linear discriminate analysis (LDA) [11], locality preserving projection (LPP) [12] and kernel principle component analysis (KPCA) [13], etc; Orientation-based approaches, e.g., Gabor filter [14], etc; Multi-feature based approaches [15]; Different features, such as texture and line [15], can be fused into one single feature vector with feature layer fusion (FLF). FLF methods are divided into four classes: serial fusion, parallel fusion, weight fusion, and kernel based fusion. Serial fusion technique is straightforward, which contains the whole discriminatory information. A system characterized by the serial combination of multiple features can be a good trade-off between verification time, performance and acceptability. Experimental results confirm that proper fusion [15] is a promising approach that may increase the accuracy of systems.Distinguished feature extraction from original palmprint image space is inefficient because that pixel value is high auto-correlation. The principle advantage of discrete cosine transform (DCT) [16] is efficient in removal ...

show abstract

Section: і Introductionmentioning

confidence: 99%

Palmprint recognition by fusion of multi-color components

Zhang

Sun

Qiu

et al. 2009

2009 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery

View full text Add to dashboard Cite

show abstract

“…However, the Orientation based approaches loss the local minutia information. Recently, using information fusion technologies, multi-feature and multimodel based approaches have been proposed [12][13][14], which lead to relatively satisfying performance for these approaches depend on more features.…”

Section: Introductionmentioning

confidence: 99%

Palmprint Verification Based on Orthogonal Code

Song

Liu

2010

2010 Third International Conference on Information and Computing

View full text Add to dashboard Cite

In this paper, we propose a novel palmprint verification scheme named as Orthogonal Code based on multiple 2D Orthogonal Gabor Phase feature. A palmprint image contains various features with multiple orientations, so only a single orientation in the palmprint can not distinguish a subject well. The 2D Orthogonal Gabor filters with different orientations are employed to extract texture information and a phase-coding scheme is used to represent the palmprint. At the matching stage, Hamming distance is used as a similarity metric. The experimental results of verification on Hong Kong Polytechnic University Palmprint Database show that the proposed approach can operate at a high genuine acceptance rate of 99.3% and a low false acceptance rate of 4×10 -5 %, which outperforms the state-of-the-art Competitve code, OrdinalCode and Robust Line Orientation Code. Keywords-palmprint verification; 2D orthogonal Gabor filter; phase codingI.

show abstract

“…To extract a ROI from a hand image, two approaches can be used. The first approach extracts a square region [12,13,19,21] whereas the second performs a circular region on the hand image.…”

Section: Extraction Of the Palmprint Roimentioning

confidence: 99%

“…From the existing literature, several methods have been proposed to date. Primarily, these methods are based upon morphological characteristics [1,21]. For example, Zhang et al [10] use a Gabor filter to extract what has been called "Palm code" [10].…”

Section: Introductionmentioning

confidence: 99%

A Robust Technique to Characterize the Palmprint using Radon transform and Delaunay triangulation

Bouchemha¹,

Naït‐Ali²,

Doghmane³

2010

IJCA

View full text Add to dashboard Cite

For the purpose of biometric applications, we explore in this paper a new robust approach to characterizing palmprint features. Instead of processing the acquired image in the spatial domain, the proposed technique extracts palmprint features using Radon transform and a geometric Delaunay triangulation jointly. In such a process, Radon transform enables the extraction of directional characteristics from the palm of the hand. Afterwards, the most significant information is structured using Delaunay triangulation, thus providing a specific palmprint signature. In order to compare the uniqueness as well as the stability of the palmprint signature, Hausdorff distance has been used as a criterion of similarity. As will be shown in this paper, the palmprint signature is very robust even when considering a low Signal-to-Noise Ratio (SNR). Promising results are obtained from a local database containing 200 palmprint images. This technique is mainly appropriate for authentication applications.

show abstract

Palmprint identification using restricted fusion

Cited by 30 publications

References 31 publications

Palmprint recognition by fusion of multi-color components

Palmprint recognition by fusion of multi-color components

Palmprint Verification Based on Orthogonal Code

A Robust Technique to Characterize the Palmprint using Radon transform and Delaunay triangulation

Contact Info

Product

Resources

About