Face recognition using combined multiple feature extraction based on Fourier-Mellin approach for single example image per person

Chen, Yee Ming; Chiang, Jen-Hong

doi:10.1016/j.patrec.2010.03.018

Cited by 23 publications

(10 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We noticed that some classification results on the ORL database and the Yale database are also reported for a multiple-feature method (denoted as MFM) [14]. Here, we can present only a rough comparison because the image sizes and the experimental set-ups are different for our MR_2DLDA method and the MFM method.…”

Section: Resultsmentioning

confidence: 99%

“…Here, we can present only a rough comparison because the image sizes and the experimental set-ups are different for our MR_2DLDA method and the MFM method. It can be seen from Tables 9 and 10 in [14] that, the classification rates of the ORL database and the Yale database are 71% and 0.69%, respectively, when the first image of each individual is used as the training sample. However, we can see from Table 4 that the corresponding classification rates of the MR_2DLDA method are 71.39% and 80%, respectively.…”

Section: Resultsmentioning

confidence: 99%

“…Multiple-feature fusion is also an effective approach for the one-sample-per-person face recognition problem. A combination of the frequency invariant features and the moment invariant features [14], and a fusion of the directionality of edges and the intensity facial features [15], are proposed for face recognition with a single training sample. Instead of using 2D representation, a 3D model-based method is an important approach to the one-sample-per-person face recognition problem.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Face Recognition with Multi-Resolution Spectral Feature Images

et al. 2013

View full text Add to dashboard Cite

The one-sample-per-person problem has become an active research topic for face recognition in recent years because of its challenges and significance for real-world applications. However, achieving relatively higher recognition accuracy is still a difficult problem due to, usually, too few training samples being available and variations of illumination and expression. To alleviate the negative effects caused by these unfavorable factors, in this paper we propose a more accurate spectral feature image-based 2DLDA (two-dimensional linear discriminant analysis) ensemble algorithm for face recognition, with one sample image per person. In our algorithm, multi-resolution spectral feature images are constructed to represent the face images; this can greatly enlarge the training set. The proposed method is inspired by our finding that, among these spectral feature images, features extracted from some orientations and scales using 2DLDA are not sensitive to variations of illumination and expression. In order to maintain the positive characteristics of these filters and to make correct category assignments, the strategy of classifier committee learning (CCL) is designed to combine the results obtained from different spectral feature images. Using the above strategies, the negative effects caused by those unfavorable factors can be alleviated efficiently in face recognition. Experimental results on the standard databases demonstrate the feasibility and efficiency of the proposed method.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Face Recognition with Multi-Resolution Spectral Feature Images

et al. 2013

View full text Add to dashboard Cite

show abstract

“…They produce a larger number of coefficients as compared to other moments for the same maximum order, thus is more suitable in case of small images. The application areas include pattern recognition [18], character recognition [19], edge location [20], face recognition [21], image reconstruction [22] and so on. The RHFMs were introduced by Ren et al [23].The important properties possessed by RHFM are magnitude invariance, robustness to image noise, orthogonality and reconstruction capability.…”

Section: Introductionmentioning

confidence: 99%

Empirical Analysis of Rotation Invariance in Moment Coefficients

Ranade¹,

Anand²

2015

IJCA

View full text Add to dashboard Cite

Moments can be viewed as powerful image descriptors that capture global characteristics of an image. The magnitude of the moment coefficients is said to be invariant under geometrical transformations like rotation which makes them suitable for most of the recognition applications. But in practice, the invariance of moment coefficients is compromised due to the errors in computation. This paper presents an empirical study of some popularly used moment functions to find out the robust coefficients under rotation. The selected robust coefficients are used in face recognition under in-plane rotation. Experimental results demonstrate that the performance of the proposed method comes at par with the performance of the traditional method by using lesser number of moment coefficients and thus results in significant saving in the feature extraction time.

show abstract

“…The high order moments are, however, prone to numerical integration error and numerical instability. Due to these attractive features, the OFMMs are used in a variety of applications that include pattern recognition [14,21], optical character recognition [9], fingerprint verification [1], object recognition [18], edge detection [3], face recognition [4,6], image reconstruction [12,17], etc.…”

Section: Introductionmentioning

confidence: 99%

Accurate Computation of Orthogonal Fourier-Mellin Moments

Singh

Upneja

2012

J Math Imaging Vis

View full text Add to dashboard Cite

Orthogonal Fourier-Mellin moments (OFMMs) suffer from geometric error and the numerical integration error. The geometric error arises when the square image is mapped into a unit disk and the mapping does not become perfect. The numerical integration error arises when the double integration is approximated by the zeroth order summation. In this paper, we propose methods which reduce these errors. The geometric error is reduced by considering the arc-grids lying on the boundary of the unit disk and the square grids lying completely inside the disk. The numerical integration error is reduced by Gaussian numerical integration, for which a simple computational framework is provided. The relative contributions of geometric error and numerical integration error to the total error are also analyzed. It is observed that the geometric error is significant only for the small images whereas the magnitude of numerical integration is significantly high for all image sizes, which increases with the order of moments. A simple computational framework which is similar to the conventional zeroth order approximation is also proposed which not only reduces numerical integration error but also reduces geometric error without considering arc-grids. The improved accuracy of OFMMs are shown to provide better image reconstruction, numerical stability and rotation and scale invariance. Exhaustive experimental results on a variety of real images have shown the efficacy of the proposed methods.

show abstract

Face recognition using combined multiple feature extraction based on Fourier-Mellin approach for single example image per person

Cited by 23 publications

References 25 publications

Face Recognition with Multi-Resolution Spectral Feature Images

Face Recognition with Multi-Resolution Spectral Feature Images

Empirical Analysis of Rotation Invariance in Moment Coefficients

Accurate Computation of Orthogonal Fourier-Mellin Moments

Contact Info

Product

Resources

About