Alex C. Kot

Eigenfeature Regularization and Extraction in Face Recognition

This work proposes a subspace approach that regularizes and extracts eigenfeatures from the face image. Eigenspace of the within-class scatter matrix is decomposed into three subspaces: a reliable subspace spanned mainly by the facial variation, an unstable subspace due to noise and finite number of training samples, and a null subspace. Eigenfeatures are regularized differently in these three subspaces based on an eigenspectrum model to alleviate problems of instability, overfitting, or poor generalization. This also enables the discriminant evaluation performed in the whole space. Feature extraction or dimensionality reduction occurs only at the final stage after the discriminant assessment. These efforts facilitate a discriminative and a stable low-dimensional feature representation of the face image. Experiments comparing the proposed approach with some other popular subspace methods on the FERET, ORL, AR, and GT databases show that our method consistently outperforms others.

Two-Dimensional Polar Harmonic Transforms for Invariant Image Representation

This paper introduces a set of 2D transforms, based on a set of orthogonal projection bases, to generate a set of features which are invariant to rotation. We call these transforms Polar Harmonic Transforms (PHTs). Unlike the well-known Zernike and pseudo-Zernike moments, the kernel computation of PHTs is extremely simple and has no numerical stability issue whatsoever. This implies that PHTs encompass the orthogonality and invariance advantages of Zernike and pseudo-Zernike moments, but are free from their inherent limitations. This also means that PHTs are well suited for application where maximal discriminant information is needed. Furthermore, PHTs make available a large set of features for further feature selection in the process of seeking for the best discriminative or representative features for a particular application.

Quality Measures of Fingerprint Images

In an automatic fingerprint identification system, it is desirable to estimate the image quality of the fingerprint image before it is processed for feature extraction. This helps in deciding on the type of image enhancements that are needed and in deciding on thresholds for the matcher in the case that dynamic thresholds are used. In this paper, we propose a Gabor-feature based method for determining the quality of the fingerprint images. An image is divided into N w?wblocks. Gabor features of each block are computed first, then the standard deviation of the m Gabor features is used to determine the quality of this block. The results are compared with an existing model of quality estimation. Our analysis shows that our method can estimate the image quality accurately.

Distance-reciprocal distortion measure for binary document images

In this letter, we present a novel objective distortion measure for binary document images. This measure is based on the reciprocal of distance that is straightforward to calculate. Our results show that the proposed distortion measure matches well to subjective evaluation by human visual perception.

Accurate Detection of Demosaicing Regularity for Digital Image Forensics

In this paper, we propose a novel accurate detection framework of demosaicing regularity from different source images. The proposed framework first reversely classifies the demosaiced samples into several categories and then estimates the underlying demosaicing formulas for each category based on partial second-order derivative correlation models, which detect both the intrachannel and the cross-channel demosaicing correlation. An expectation-maximization reverse classification scheme is used to iteratively resolve the ambiguous demosaicing axes in order to best reveal the implicit grouping adopted by the underlying demosaicing algorithm. Comparison results based on syntactic images show that our proposed formulation significantly improves the accuracy of the regenerated demosaiced samples from the sensor samples for a large number of diversified demosaicing algorithms. By running sequential forward feature selection, our reduced feature sets used in conjunction with the probabilistic support vector machine classifier achieve superior performance in identifying 16 demosaicing algorithms in the presence of common camera post demosaicing processing. When applied to real applications, including camera model and RAW-tool identification, our selected features achieve nearly perfect classification performances based on large sets of cropped image blocks.

Fingerprint reference-point detection

A robust fingerprint recognition algorithm should tolerate the rotation and translation of the fingerprint image. One popular solution is to consistently detect a unique reference point and compute a unique reference orientation for translational and rotational alignment. This paper develops an effective algorithm to locate a reference point and compute the corresponding reference orientation consistently and accurately for all types of fingerprints. To compute the reliable orientation field, an improved orientation smoothing method is proposed based on adaptive neighborhood. It shows better performance in filtering noise while maintaining the orientation localization than the conventional averaging method. The reference-point localization is based on multiscale analysis of the orientation consistency to search the local minimum. The unique reference orientation is computed based on the analysis of the orientation differences between the radial directions from the reference point, which are the directions of the radii emitted from the reference point with equivalent angle interval, and the local ridge orientations along these radii. Experimental results demonstrate that our proposed algorithm can consistently locate a unique reference point and compute the reference orientation with high accuracy for all types of fingerprints.

An intelligent pressure sensor using neural networks

In this paper, we propose a scheme of an intelligent capacitive pressure sensor (CPS) using an artificial neural network (ANN). A switched-capacitor circuit (SCC) converts the change in capacitance of the pressure-sensor into an equivalent voltage. The effect of change in environmental conditions on the CPS and subsequently upon the output of the SCC is nonlinear in nature. Especially, change in ambient temperature causes response characteristics of the CPS to become highly nonlinear, and complex signal processing may be required to obtain correct readout. The proposed ANN-based scheme incorporates intelligence into the sensor. It is revealed from the simulation studies that this CPS model can provide correct pressure readout within /spl plusmn/1% error (full scale) over a range of temperature variations from -20/spl deg/C to 70/spl deg/C. Two ANN schemes, direct modeling and inverse modeling of a CPS, are reported. The former modeling technique enables an estimate of the nonlinear sensor characteristics, whereas the latter technique estimates the applied pressure which is used for direct digital readout. When there is a change in ambient temperature, the ANN automatically compensates for this change based on the distributive information stored in its weights.

No-Reference Image Blur Assessment Based on Discrete Orthogonal Moments

Blur is a key determinant in the perception of image quality. Generally, blur causes spread of edges, which leads to shape changes in images. Discrete orthogonal moments have been widely studied as effective shape descriptors. Intuitively, blur can be represented using discrete moments since noticeable blur affects the magnitudes of moments of an image. With this consideration, this paper presents a blind image blur evaluation algorithm based on discrete Tchebichef moments. The gradient of a blurred image is first computed to account for the shape, which is more effective for blur representation. Then the gradient image is divided into equal-size blocks and the Tchebichef moments are calculated to characterize image shape. The energy of a block is computed as the sum of squared non-DC moment values. Finally, the proposed image blur score is defined as the variance-normalized moment energy, which is computed with the guidance of a visual saliency model to adapt to the characteristic of human visual system. The performance of the proposed method is evaluated on four public image quality databases. The experimental results demonstrate that our method can produce blur scores highly consistent with subjective evaluations. It also outperforms the state-of-the-art image blur metrics and several general-purpose no-reference quality metrics.

Fingerprint Retrieval for Identification

This paper presents a front-end filtering algorithm for fingerprint identification, which uses orientation field and dominant ridge distance as retrieval features. We propose a new distance measure that better quantifies the similarity evaluation between two orientation fields than the conventional Euclidean and Manhattan distance measures. Furthermore, fingerprints in the data base are clustered to facilitate a fast retrieval process that avoids exhaustive comparisons of an input fingerprint with all fingerprints in the data base. This makes the proposed approach applicable to large databases. Experimental results on the National Institute of Standards and Technology data base-4 show consistent better retrieval performance of the proposed approach compared to other continuous and exclusive fingerprint classification methods as well as minutia-based indexing schemes

Integer DCTs and fast algorithms

A method is proposed to factor the type-II discrete cosine transform (DCT-II) into lifting steps and additions. After approximating the lifting matrices, we get a new type-II integer discrete cosine transform (IntDCT-II) that is float-point multiplication free. Based on the relationships among the various types of DCTs, we can generally factor any DCTs into lifting steps and additions and then get four types of integer DCTs, which need no float-point multiplications. By combining the polynomial transform and the one-dimensional (1-D) integer cosine transform, a two-dimensional (2-D) integer discrete cosine transform is proposed. The proposed transform needs only integer operations and shifts. Furthermore, it is nonseparable and requires a far fewer number of operations than that used by the corresponding row-column 2-D integer discrete cosine transform.