In this correspondence, we propose a novel method to extract illumination insensitive features for face recognition under varying lighting called the gradient faces. Theoretical analysis shows gradient faces is an illumination insensitive measure, and robust to different illumination, including uncontrolled, natural lighting. In addition, gradient faces is derived from the image gradient domain such that it can discover underlying inherent structure of face images since the gradient domain explicitly considers the relationships between neighboring pixel points. Therefore, gradient faces has more discriminating power than the illumination insensitive measure extracted from the pixel domain. Recognition rates of 99.83% achieved on PIE database of 68 subjects, 98.96% achieved on Yale B of ten subjects, and 95.61% achieved on Outdoor database of 132 subjects under uncontrolled natural lighting conditions show that gradient faces is an effective method for face recognition under varying illumination. Furthermore, the experimental results on Yale database validate that gradient faces is also insensitive to image noise and object artifacts (such as facial expressions).

Face Recognition Under Varying Illumination Using Gradientfaces

In this paper, a new denoising method is proposed for hyperspectral data cubes that already have a reasonably good signal-to-noise ratio (SNR) (such as 600 : 1). Given this level of the SNR, the noise level of the data cubes is relatively low. The conventional image denoising methods are likely to remove the fine features of the data cubes during the denoising process. We propose to decorrelate the image information of hyperspectral data cubes from the noise by using principal component analysis (PCA) and removing the noise in the low-energy PCA output channels. The first PCA output channels contain a majority of the total energy of a data cube, and the rest PCA output channels contain a small amount of energy. It is believed that the low-energy channels also contain a large amount of noise. Removing noise in the low-energy PCA output channels will not harm the fine features of the data cubes. A 2-D bivariate wavelet thresholding method is used to remove the noise for low-energy PCA channels, and a 1-D dual-tree complex wavelet transform denoising method is used to remove the noise of the spectrum of each pixel of the data cube. Experimental results demonstrated that the proposed denoising method produces better denoising results than other denoising methods published in the literature.

Denoising of Hyperspectral Imagery Using Principal Component Analysis and Wavelet Shrinkage

Deep convolutional neural networks (DCNNs) have led to breakthrough results in numerous practical machine learning tasks, such as classification of images in the ImageNet data set, control-policy-learning to play Atari games or the board game Go, and image captioning. Many of these applications first perform feature extraction and then feed the results thereof into a classifier. The mathematical analysis of DCNNs for feature extraction was initiated by Mallat, 2012. Specifically, Mallat considered so-called scattering networks based on a wavelet transform followed by the modulus non-linearity in each network layer, and proved translation invariance (asymptotically in the wavelet scale parameter) and deformation stability of the corresponding feature extractor. This paper complements Mallat’s results by developing a theory that encompasses general convolutional transforms, or in more technical parlance, general semi-discrete frames (including Weyl–Heisenberg filters, curvelets, shearlets, ridgelets, wavelets, and learned filters), general Lipschitz-continuous non-linearities (e.g., rectified linear units, shifted logistic sigmoids, hyperbolic tangents, and modulus functions), and general Lipschitz-continuous pooling operators emulating, e.g., sub-sampling and averaging. In addition, all of these elements can be different in different network layers. For the resulting feature extractor, we prove a translation invariance result of vertical nature in the sense of the features becoming progressively more translation-invariant with increasing network depth, and we establish deformation sensitivity bounds that apply to signal classes such as, e.g., band-limited functions, cartoon functions, and Lipschitz functions.

A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction

Machine learning techniques for breast cancer computer aided diagnosis using different image modalities: A systematic review.

Deep convolutional neural networks have led to breakthrough results in numerous practical machine learning tasks such as classification of images in the ImageNet data set, control-policy-learning to play Atari games or the board game Go, and image captioning. Many of these applications first perform feature extraction and then feed the results thereof into a trainable classifier. The mathematical analysis of deep convolutional neural networks for feature extraction was initiated by Mallat, 2012. Specifically, Mallat considered so-called scattering networks based on a wavelet transform followed by the modulus non-linearity in each network layer, and proved translation invariance (asymptotically in the wavelet scale parameter) and deformation stability of the corresponding feature extractor. This paper complements Mallat's results by developing a theory that encompasses general convolutional transforms, or in more technical parlance, general semi-discrete frames (including Weyl-Heisenberg filters, curvelets, shearlets, ridgelets, wavelets, and learned filters), general Lipschitz-continuous non-linearities (e.g., rectified linear units, shifted logistic sigmoids, hyperbolic tangents, and modulus functions), and general Lipschitz-continuous pooling operators emulating, e.g., sub-sampling and averaging. In addition, all of these elements can be different in different network layers. For the resulting feature extractor we prove a translation invariance result of vertical nature in the sense of the features becoming progressively more translation-invariant with increasing network depth, and we establish deformation sensitivity bounds that apply to signal classes such as, e.g., band-limited functions, cartoon functions, and Lipschitz functions.

Image denoising with neighbour dependency and customized wavelet and threshold

This paper describes several techniques improving a Chinese character recognition system. Enhanced nonlinear normalization, feature extraction and tuning kernel parameters of support vector machine on a large data set with thousands of classes, contribute to improvement of the overall system performance. The enhanced nonlinear normalization method not only solves the aliasing problem in the original Yamada et al.'s nonlinear normalization method but also avoids the undue stroke distortion in the peripheral region of the normalized image. The support vector machine is for the first time tested on a large data set composed of several million samples and thousands of classes. The recognition system has achieved a high recognition rate of 99.0% on ETL9B, a handwritten Chinese character database.

An improved handwritten Chinese character recognition system using support vector machine

Semi-automatic computer aided lesion detection in dental X-rays using variational level set

Invariant pattern recognition using radon, dual-tree complex wavelet and Fourier transforms

http://users.encs.concordia.ca/~bui/pdf/pattern-rec-Chen-Bui-Krzy.pdf

Adam Krzyak

Papers

Image denoising with neighbour dependency and customized wavelet and threshold

An improved handwritten Chinese character recognition system using support vector machine

Semi-automatic computer aided lesion detection in dental X-rays using variational level set

Invariant pattern recognition using radon, dual-tree complex wavelet and Fourier transforms

Rotation invariant pattern recognition using ridgelets, wavelet cycle-spinning and Fourier features