Emma Izquierdo-Verdiguier

Encoding Invariances in Remote Sensing Image Classification With SVM

This letter introduces a simple method for including invariances in support-vector-machine (SVM) remote sensing image classification. We design explicit invariant SVMs to deal with the particular characteristics of remote sensing images. The problem of including data invariances can be viewed as a problem of encoding prior knowledge, which translates into incorporating informative support vectors (SVs) that better describe the classification problem. The proposed method essentially generates new (synthetic) SVs from the obtained by training a standard SVM with the available labeled samples. Then, original and transformed SVs are used for training the virtual SVM introduced in this letter. We first incorporate invariances to rotations and reflections of image patches for improving contextual classification. Then, we include an invariance to object scale in patch-based classification. Finally, we focus on the challenging problem of including illumination invariances to deal with shadows in the images. Very good results are obtained when few labeled samples are available for classification. The obtained classifiers reveal enhanced sparsity and robustness. Interestingly, the methodology can be applied to any maximum-margin method, thus constituting a new research opportunity.

Semisupervised Kernel Feature Extraction for Remote Sensing Image Analysis

This paper presents a novel semisupervised kernel partial least squares (KPLS) algorithm for nonlinear feature extraction to tackle both land-cover classification and biophysical parameter retrieval problems. The proposed method finds projections of the original input data that align with the target variable (labels) and incorporates the wealth of unlabeled information to deal with low-sized or underrepresented data sets. The method relies on combining two kernel functions: the standard radial-basis-function kernel based on labeled information and a generative, i.e., probabilistic, kernel directly learned by clustering the data many times and at different scales across the data manifold. The construction of the kernel is very simple and intuitive: Two samples should belong to the same class if they consistently belong to the same clusters at different scales. The effectiveness of the proposed method is successfully illustrated in multi- and hyperspectral remote sensing image classification and biophysical parameter estimation problems. Accuracy improvements in the range between +5% and 15% over standard principal component analysis (PCA), +4% and 15% over kernel PCA, and +3% and 10% over KPLS are obtained on several images. The average gain in the root-mean-square error of +5% and reductions in bias estimates of +3% are obtained for biophysical parameter retrieval compared to standard PCA feature extraction.

Spectral clustering with the probabilistic cluster kernel

Abstract This letter introduces a probabilistic cluster kernel for data clustering. The proposed kernel is computed with the composition of dot products between the posterior probabilities obtained via GMM clustering. The kernel is directly learned from the data, is parameter-free, and captures the data manifold structure at different scales. The projections in the kernel space induced by this kernel are useful for general feature extraction purposes and are here exploited in spectral clustering with the canonical k-means. The kernel structure, informative content and optimality are studied. Analysis and performance are illustrated in several real datasets.

Cloud screening with combined MERIS and AATSR images

This paper presents a cloud screening algorithm based on ensemble methods that exploits the combined information from both MERIS and AATSR instruments on board ENVISAT in order to improve current cloud masking products for both sensors. The first step is to analyze the synergistic use of MERIS and AATSR images in order to extract some physically-based features increasing the separability of clouds and surface. Then, several artificial neural networks are trained using different sets of input features and different sets of training samples depending on acquisition and surface conditions. Finally, outputs of the trained neural networks are combined at the decision level to construct a more accurate and robust ensemble of classifiers. The proposed classifier is tested on more than 80 coregistered MERIS/AATSR images providing better classification accuracy than the official cloud flags and available operational cloud screening algorithms for MERIS and AATSR. Moreover, thanks to the synergy of both sensors, it correctly classifies critical cloud-screening problems such as snow and ice covers over land and sun-glint over ocean.

Optimized Kernel Entropy Components

This brief addresses two main issues of the standard kernel entropy component analysis (KECA) algorithm: the optimization of the kernel decomposition and the optimization of the Gaussian kernel parameter. KECA roughly reduces to a sorting of the importance of kernel eigenvectors by entropy instead of variance, as in the kernel principal components analysis. In this brief, we propose an extension of the KECA method, named optimized KECA (OKECA), that directly extracts the optimal features retaining most of the data entropy by means of compacting the information in very few features (often in just one or two). The proposed method produces features which have higher expressive power. In particular, it is based on the independent component analysis framework, and introduces an extra rotation to the eigen decomposition, which is optimized via gradient-ascent search. This maximum entropy preservation suggests that OKECA features are more efficient than KECA features for density estimation. In addition, a critical issue in both the methods is the selection of the kernel parameter, since it critically affects the resulting performance. Here, we analyze the most common kernel length-scale selection criteria. The results of both the methods are illustrated in different synthetic and real problems. Results show that OKECA returns projections with more expressive power than KECA, the most successful rule for estimating the kernel parameter is based on maximum likelihood, and OKECA is more robust to the selection of the length-scale parameter in kernel density estimation.

Semisupervised kernel orthonormalized partial least squares

This paper presents a semisupervised kernel orthonormalized partial least squares (SS-KOPLS) algorithm for non-linear feature extraction. The proposed method finds projections that minimize the least squares regression error in Hilbert spaces and incorporates the wealth of unlabeled information to deal with small size labeled datasets. The method relies on combining a standard RBF kernel using labeled information, and a generative kernel learned by clustering all available data. The positive definiteness of the kernels is proven, and the structure and information content of the derived kernels is studied. The effectiveness of the proposed method is successfully illustrated in standard UCI database classification, Olivetti face database manifold learning, and in high-dimensional hyperspectral satellite image segmentation. High accuracy gains are obtained over KPLS in terms of expressive power of the extracted non-linear features. Matlab code is available at http://isp.uv.es for the interested readers.

Semisupervised nonlinear feature extraction for image classification

Feature extraction is of paramount importance for an accurate classification of remote sensing images. Techniques based on data transformations are widely used in this context. However, linear feature extraction algorithms, such as the principal component analysis and partial least squares, can address this problem in a suboptimal way because the data relations are often nonlinear. Kernel methods may alleviate this problem only when the structure of the data manifold is properly captured. However, this is difficult to achieve when small-size training sets are available. In these cases, exploiting the information contained in unlabeled samples together with the available training data can significantly improve data description by defining an effective semisupervised nonlinear feature extraction strategy. We present a novel semisupervised Kernel Partial Least Squares (KPLS) algorithm for non-linear feature extraction. The method relies on combining two kernel functions: the standard RBF kernel using labeled information and a generative kernel directly learned by clustering the data. The effectiveness of the proposed method is successfully illustrated in multi- and hyper-spectral remote sensing image classification: accuracy improvements between +15 - 20% over standard PCA and +10% over advanced kernel PCA and KPLS for both images is obtained. Matlab code is available at http://isp.uv.es for the interested readers.

Kernels for Remote Sensing Image Classification

Classification of images acquired by airborne and satellite sensors is a very challenging problem. These remotely sensed images usually acquire information from the scene at different wavelengths or spectral channels. The main problems involved are related to the high dimensionality of the data to be classified and the very few existing labeled samples, the diverse noise sources involved in the acquisition process, the intrinsic nonlinearity and non-Gaussianity of the data distribution in feature spaces, and the high computational cost involved to process big data cubes in near real time. The framework of statistical learning in general, and of kernel methods in particular, has gained popularity in the last decade. New methods have been presented to address all these problem specificities. We will review the main existing algorithmic proposals to cope with the spatial homogeneity of images, to take advantage of the manifold structure with semisupervised learning, to encode invariances, and to deal with one-class and multitemporal problems. The design of a suitable regularized functional has led to the successful solution of many remote sensing image classification problems. This article reviews the main advances for remote sensing image supervised classification with kernels through illustrative examples. n n nKeywords: n ngeoscience; nimage processing; nkernel methods; nremote sensing

A kernel regression approach to cloud and shadow detection in multitemporal images

Earth observation satellites will provide in the next years time series with enough revisit time allowing a better surface monitoring. In this work, we propose a cloud screening and a cloud shadow detection method based on detecting abrupt changes in the temporal domain. It is considered that the time series follows smooth variations and abrupt changes in certain spectral features will be mainly due to the presence of clouds or cloud shadows. The method is based on linear and nonlinear regression analysis; in particular we focus on the regularized least squares and kernel regression methods. Experiments are carried out using Landsat 5 TM time series acquired over Albacete (Spain), and comparative results with the Fmask approach [1] show the potential of exploiting the temporal domain.

Including invariances in SVM remote sensing image classification

This paper introduces a simple method to include invariances in support vector machine (SVM) for remote sensing image classification. We rely on the concept of virtual support vectors, by which the SVM is trained with both the selected support vectors and synthetic examples encoding the invariance of interest. The algorithm is very simple and effective, as demonstrated in two particularly interesting examples: invariance to the presence of shadows and to rotations in patchbased image segmentation. The improved accuracy (around +6% both in OA and Cohens κ statistic), along with the simplicity of the approach encourage its use and extension to encode other invariances and other remote sensing data analysis applications.