The accuracy of a classifier, whether it is an ensemble or not, is directly influenced by the training data used in learning. In remote sensing, training data mislabeling is inevitable and faces a major challenge. This paper proposes a versatile data cleaning which handles the mislabeling problem by exploiting the ensemble concepts for identifying, then eliminating or correcting the mislabeled training data. A powerful ensemble method, random forest, is at the core of our filter design and helps to distinguish mislabeled data from uncorrupted data more accurately. The major contribution of this work lies on the explicit use of the hypothesis margin as a decision means to identify and eliminate or correct mislabeled training data in an ensemble learning framework. Another key development that makes our algorithm superior to existing approaches is a design that avoids rare class instances to be mistaken for class noise. This fundamental aspect makes our data cleaning system particularly suitable for remote sensing classification tasks which usually suffer from both mislabeling and imbalance problems. The effectiveness of our algorithm is demonstrated in performing mapping of land covers. The generalization performance of two major supervised noise-sensitive classifiers, boosting and K-nearest neighbors, is strengthened by effective class noise reduction. A comparative analysis is conducted with respect to random forest, deep convolutional neural networks, as well as two well-established ensemble-based class noise filters, the majority vote and the consensus vote filters. This analysis demonstrates that our approach is more accurate than deep convolutional neural networks (one-dimensional CNN, AlexNet, EfficientNet, ResNet50 and ShuffletNet) and the reference ensemble methods.

The characteristics of training data is a fundamental consideration when constructing any supervised classifier. Class mislabelling and imbalance are major training data issues that often adversely affect machine learning algorithms, including ensembles. This work proposes extended bagging algorithms to better handle noisy and multi-class imbalanced classification tasks. These algorithms upgrade the sampling procedure by taking benefit of the confidence in ensemble classification outcome. The underlying idea is that a bagging ensemble learning algorithm can achieve greater performance if it is allowed to choose the data from which it learns. The effectiveness of the proposed methods is demonstrated in performing classification on 10 various data sets.

Abstract The ensemble method is a powerful data mining paradigm, which builds a classification model by integrating multiple diversified component learners. Bagging is one of the most successful ensemble methods. It is made of bootstrap‐inspired classifiers and uses these classifiers to get an aggregated classifier. However, in bagging, bootstrapped training sets become more and more similar as redundancy is increasing. Besides redundancy, any training set is usually subject to noise. Moreover, the training set might be imbalanced. Thus, each training instance has a different impact on the learning process. This paper explores some properties of the ensemble margin and its use in improving the performance of bagging. We introduce a new approach to measure the importance of training data in learning, based on the margin theory. Then, a new bagging method concentrating on critical instances is proposed. This method is more accurate than bagging and more robust than boosting. Compared to bagging, it reduces the bias while generally keeping the same variance. Our findings suggest that (a) examples with low margins tend to be more critical for the classifier performance; (b) examples with higher margins tend to be more redundant; (c) misclassified examples with high margins tend to be noisy examples. Our experimental results on 15 various data sets show that the generalization error of bagging can be reduced up to 2.5% and its resilience to noise strengthened by iteratively removing both typical and noisy training instances, reducing the training set size by up to 75%.

The problem of class imbalance is often encountered in remote sensing data and has a negative effect on the classification performance of supervised classifiers even in ensemble models. The ensemble margin is a fundamental concept in ensemble learning with potential effectiveness in improving the classification of remote sensing data. This paper proposes a novel margin based extended random forest algorithm to address the class imbalance issues in the difficult context of remote sensing classification. This algorithm combines ensemble learning with data sampling. A comparative analysis is conducted with respect to standard random forest, undersampling and over-sampling combined ensembles.

The accuracy of a supervised classifier is directly influenced by the quality of the training data used. However, real-world data often suffers from mislabelling issues. To handle the mislabeling problem, we propose an ensemble margin-based mislabeled training data identification, elimination and correction approach based on data ordering. A powerful ensemble method, random forest, is at the core of our algorithms design. The effectiveness of our methods is demonstrated in performing mapping of land covers. A comparative analysis is conducted with respect to the majority vote filter, a popular ensemble-based mislabeled data filter.

Ensemble classifiers perform better than single classifiers and result in reduced generalisation error. Diversity across ensemble members is a key factor affecting classification performance. Here, an original exploration of the relationship between ensemble diversity and classification performance applied to large area remote sensing classification, using random forests, is undertaken. Results demonstrate how targeting lower margin training samples is both a strategy for inducing diversity in ensemble classifiers and achieving better classifier performance for difficult or rare classes, and a way to reduce data redundancy.

Bagging is a simple and powerful ensemble method which relies on bootstrap sampling over training data to produce diversity. Indeed, ensembles generalise better when their members form a diverse and accurate set. In this paper, the margin theory is exploited to select training instances for bagging. The selection of training data is performed using a new iterative guided bagging algorithm exploiting low margin instances. This method has been successfully applied to image data. Results show that low margin instances have a major influence on forming an appropriate training set to build reliable ensemble classifiers, leading to a significant increase in both overall and per-class accuracies.

Mislabeled training data is a challenge to face in order to build a robust classifier whether it is an ensemble or not. This work handles the mislabeling problem by exploiting four different ensemble margins for identifying, then eliminating or correcting the mislabeled training data. Our approach is based on class noise ordering and relies on the margin values of misclassified data. The effectiveness of our ordering-based class noise removal and correction methods is demonstrated in performing image classification. A comparative analysis is conducted with respect to the majority vote filter, a reference ensemble-based class noise filter.

This work investigates the discriminative power of wavelet decomposition based texture features in forest cover classification. Our texture features are used as inputs in a random forests classifier. The performances of this tree-based ensemble classifier are assessed by classification accuracy as well as classification confidence provided by an unsupervised version of ensemble margin. The effectiveness of the proposed texture based multiple classifier system is demonstrated in performing mapping of very high resolution forest imagery. Traditional grey level co-occurrence matrix derived texture features are also evaluated through our ensemble classification framework for comparison.

In remote sensing, where training data are typically ground-based, mislabeled training data is inevitable. This work handles the mislabeling problem by exploiting the ensemble margin for identifying, then eliminating or correcting the mislabeled training data. The effectiveness of our class noise removal and correction methods is demonstrated in performing mapping of land covers. A comparative analysis is conducted with respect to the majority vote filter, a reference ensemble-based class noise filter.

Ensemble methods have been successfully used as a classification scheme. This work focuses on exploiting the margin theory to design better ensemble classifiers. We show that low margin instances have a major influence in building reliable classifiers. The margin paradigm is at the core of a new ordering-based mislabeled instance elimination method. The same margin framework, relying on an alternative definition of ensemble margin, is used to derive a novel ensemble diversity measure that has the property of revealing sources of diversity at data level. Our work has been successfully applied to image data.

This work introduces new ensemble margin criteria, to evaluate the performance of Random Forests (RF), in the context of large area land cover classification, using imbalanced and noisy training data. Experiments using binary and multiclass classification problems reveal insights into the behaviour of RF over big data, in which training data contains noise and may not be evenly distributed among classes. The margin-based RF performance evaluation is conducted using remote sensing and ancillary spatial data, across a 7.2 million hectare study area.

The potential of very high resolution Pléiades image texture for forest structure mapping was assessed on maritime pine stands in south-western France. A preliminary step showed that multi-linear regressions allow a reliable prediction of forest variables (such as crown diameter or tree height) from a set of features automatically selected among a huge number of Haralick texture features with various spatial parameterizations. In a second step, to assess Pléiades image texture contribution for classification, Random Forests (RF) classification was performed to discriminate four forest structure classes from recent reforestation to mature stand. Two texture feature selection strategies are compared: (1) the previous regression-based modelling using in situ tree measurements (2) the RF-variable importance using a visual photo-interpretation. Both methods produced comparable classification accuracies. The results highlight the contribution of processes automation and the need for using both Pléiades image resolutions (panchromatic and multispectral) to derive the best performing texture features.

This work exploits the margin theory to design better ensemble classifiers for remote sensing data. The margin paradigm is at the core of a new bagging algorithm. This method increases the classification accuracy, particularly in case of difficult classes, and significantly reduces the training set size. The same margin framework is used to derive a novel ensemble pruning algorithm. This method not only highly reduces the complexity of ensemble methods but also performs better than complete bagging in handling minority classes. Our techniques have been successfully used for the classification of remote sensing data.

This paper presents a new feature selection method which aims to effectively and efficiently map remote sensing data. An automated texture-based modelling procedure of forest structure variables is at the core of our approach. We show that texture features that are highly correlated to genuine physical parameters of forest structure have potential for building reliable classifiers. We demonstrate the effectiveness of our modelling-based texture feature selection method in performing mapping of very high resolution forest images. Our method outperforms Random Forest variable importance in terms of classification accuracy and computational complexity.

Random forests ensemble classifier showed to be suitable for classifying multisource data such as lidar and RGB image for urban scene mapping. However, two major problems remain: (1) the class boundaries are not well classified, a common issue in classification (2) the data are highly imbalanced raising another issue more specific to urban scenes. In this paper, we propose a new ensemble method based on the margin paradigm to improve the classification accuracy of minor classes. Random forests classifier is used in a two-pass methodology with an improved capability for classifying imbalanced data.

Support Vector Machines (SVMs) are popular for pattern classification. However, training a SVM requires large memory and high processing time, especially for large datasets, which limits their applications. To speed up their training, we present a new efficient support vector selection method based on ensemble margin, a key concept in ensemble classifiers. This algorithm exploits a new version of the margin of an ensemble-based classification and selects the smallest margin instances as support vectors. Our experimental results show that our method reduces training set size significantly without degrading the performance of the resulting SVMs classifiers.

We aim at recognizing a set of dance gestures from contemporary ballet. Our input data are motion trajectories followed by the joints of a dancing body provided by a motion-capture system. It is obvious that direct use of the original signals is unreliable and expensive. Therefore, we propose a suitable tool for nonuniform sub-sampling of spatio-temporal signals. The key to our approach is the use of polygonal approximation to provide a compact and efficient representation of motion trajectories. Our dance gesture recognition method involves a set of Hidden Markov Models (HMMs), each of them being related to a motion trajectory followed by the joints. The recognition of such movements is then achieved by matching the resulting gesture models with the input data via HMMs. We have validated our recognition system on 12 fundamental movements from contemporary ballet performed by four dancers.

A suitable detection/reconstruction approach is proposed for removing line scratches from degraded motion picture films. The detection procedure consists of two steps. First, a simple 1D-extrema detector provides line scratch candidates. Line artifacts persist across several frames. Therefore, to reject false detections, the detected scratches are tracked over the sequence using a Kalman filter. A new Bayesian restoration technique, dealing with both low and high frequencies around and inside the detected artifacts, is investigated to achieve a near invisible restoration of damaged areas.

Line scratches are common degradations in motion picture films. This paper presents an efficient method for line scratches detection strengthened by a Kalman filter. A new interpolation technique, dealing with both low and high frequencies (i.e. film grain) around the line artifacts, is investigated to achieve a nearby invisible reconstruction of damaged areas. Our line scratches detection and removal techniques have been validated on several film sequences.

This paper presents a robust technique to detect local deteriorations of old cinematographic films. This method relies on spatio-temporal information and combines two different detectors: a morphological detector which uses spatial properties of deteriorations, and a dynamic detector based on motion estimation techniques. Our deterioration detector has been validated on several film sequences and turned out to be a powerful tool for digital film restoration.

This paper deals with the recovery of 3D information using a single mobile camera in the context of active vision. First, we propose a general revisited formulation of the structure-from-known-motion issue. Within the same formalism, we handle various kinds of 3D geometrical primitives such as points, lines, cylinders, spheres, etc. We also aim at minimizing effects of the different measurement errors which are involved in such a process. More precisely, we mathematically determine optimal camera configurations and motions which lead to a robust and accurate estimation of the 3D structure parameters. We apply the visual servoing approach to perform these camera motions using a control law in closed-loop with respect to visual data. Real-time experiments dealing with 3D structure estimation of points and cylinders are reported. They demonstrate that this active vision strategy can very significantly improve the estimation accuracy.

This paper deals with the recovery of 3D information using a single mobile camera in the context of active vision. We propose a general revisited formulation of the structure-from-motion issue, and we determine adequate camera configurations and motions which lead to a robust and accurate estimation of the 3D structure parameters. We apply the visual servoing approach to perform these camera motions. Real-time experiments dealing with the 3D structure estimation of points and cylinders are reported, and demonstrate that this active vision strategy can very significantly improve the estimation accuracy.>

A method for the reconstruction and localization of geometrical primitives using active dynamic vision is presented. The approach is based on the use of the interaction matrix related to the visual data describing a primitive. Next, active vision is considered by computing adequate camera motions with a control law in closed-loop with respect to visual data. Simulation results on the localization of a sphere are presented and show that active vision can to a large extent improve the accuracy of the structure estimation.>