Apprentissage de représentations invariantes de domaines pour des données d'images hétérogènes L’apprentissage profond supervisé repose largement sur une grande quantité de données étiquetées, souvent difficile à obtenir. L'adaptation de domaine résout ce problème en appliquant les connaissances acquises à partir d'un jeu de données étiqueté à un autre jeu lié, mais non ou faiblement étiqueté. L'adaptation de domaine hétérogène est particulièrement complexe car les domaines se situent dans différents espaces. Ces méthodes sont très intéressantes pour les champs où une variété de capteurs est utilisée (comme la télédétection), capturant des images de différentes modalités. Cette thèse propose des approches novatrices pour l'Adaptation de Domaine d'Images Hétérogènes (ADIH) basées sur l'extraction de caractéristiques invariantes de domaines. La thèse examine différents scénarios de supervision dans le domaine cible : non supervisé, semi-supervisé et avec des contraintes. Les résultats montrent que l'approche proposée surpasse de manière cohérente les méthodes concurrentes.

AnA/B-Testevaluates the impact of a new technology by running it in a real production environment and testing its performance on a set of items. Recent development efforts aroundA/B-Testsrevolve around dynamic allocation. They allow for quicker determination of the best variation (A or B), thus saving money for the user. However, dynamic allocation by traditional methods requires certain assumptions, which are not always valid in reality. This is often due to the fact that the populations being tested are not homogeneous. This article reports on a new reinforcement learning methodology which has been deployed by the commercialA/B-Testplatform AB Tasty. We provide a new method that not only builds homogeneous groups of users, but also allows the best variation for these groups to be found in a short period of time. This article provides numerical results on AB Tasty's data, in addition to public datasets, tha demonstrate an improvement over traditional methods.

An important research effort has been recently dedicated to understand the decision mechanism of deep neural networks. Among them, Class Activation Mapping (CAM) and its variations have proved their capacity to obtain useful insights about Convolutional Neural Network (CNN) models’ decisions. However, these methods remain limited to the supervised case regardless of CNN-based advances in unsupervised tasks such as clustering. To fill this gap, we propose a new method called Grad-CeAM for centroid-based clustering methods used on CNN representation. Through an experimental study, we show that our method has the capacity to localize discriminating features used by a CNN model to create its representation and that it can be used to explain the clusters assignment. We also show that this method can be used in different application domains by providing uses cases on time series and images clustering.

The advent of high-resolution instruments for time-series sampling poses added complexity for the formal definition of thematic classes in the remote sensing domain-required by supervised methods-while unsupervised methods ignore expert knowledge and intuition. Constrained clustering is becoming an increasingly popular approach in data mining because it offers a solution to these problems, however, its application in remote sensing is relatively unknown. This article addresses this divide by adapting publicly available k-Means constrained clustering implementations to use the dynamic time warping (DTW) dissimilarity measure, which is thought to be more appropriate for time-series analysis. Adding constraints to the clustering problem increases accuracy when compared to unconstrained clustering. The output of such algorithms are homogeneous in spatially defined regions.

The binary partition tree (BPT) is a hierarchical data-structure that models the content of an image in a multiscale way. In particular, a cut of the BPT of an image provides a segmentation, as a partition of the image support. Actually, building a BPT allows for dramatically reducing the search space for segmentation purposes, based on intrinsic (image signal) and extrinsic (construction metric) information. A large literature has been devoted to the construction on such metrics, and the associated choice of criteria (spectral, spatial, geometric, etc.) for building relevant BPTs, in particular in the challenging context of remote sensing. But, surprisingly, there exists few works dedicated to evaluate the quality of BPTs, i.e. their ability to further provide a satisfactory segmentation. In this paper, we propose a framework for BPT quality evaluation, in a supervised paradigm. Indeed, we assume that ground-truth segments are provided by an expert, possibly with a semantic labelling and a given uncertainty. Then, we describe local evaluation metrics, BPT nodes / ground-truth segments fitting strategies, and global quality score computation considering semantic information, leading to a complete evaluation framework. This framework is illustrated in the context of BPT segmentation of multispectral satellite images.

In this article we present CoSC, a generic framework for collaborative segmentation and classification. The framework is guided by both radiometric homogeneity based criteria and implicit semantic criteria to segment and extract the objects of a given thematic class. We present a proof-of-concept case-study and show that CoSC is able to reach higher confidence for object classification and results in significant improvement of the whole segmentation.

Although agreement between the annotators who mark feature locations within images has been studied in the past from a statistical viewpoint, little work has attempted to quantify the extent to which this phenomenon affects the evaluation of foreground-background segmentation algorithms. Many researchers utilize ground truth (GT) in experimentation and more often than not this GT is derived from one annotator's opinion. How does the difference in opinion affects an algorithm's evaluation? A methodology is applied to four image-processing problems to quantify the interannotator variance and to offer insight into the mechanisms behind agreement and the use of GT. It is found that when detecting linear structures, annotator agreement is very low. The agreement in a structure's position can be partially explained through basic image properties. Automatic segmentation algorithms are compared with annotator agreement and it is found that there is a clear relation between the two. Several GT estimation methods are used to infer a number of algorithm performances. It is found that the rank of a detector is highly dependent upon the method used to form the GT, and that although STAPLE and LSML appear to represent the mean of the performance measured using individual annotations, when there are few annotations, or there is a large variance in them, these estimates tend to degrade. Furthermore, one of the most commonly adopted combination methods-consensus voting-accentuates more obvious features, resulting in an overestimation of performance. It is concluded that in some data sets, it is not possible to confidently infer an algorithm ranking when evaluating upon one GT.

Automated image analysis of multiplexed brightfield immunohistochemistry assays is a challenging objective. One central task of the analysis is the robust identification of the different stains in the image, called stain unmixing. Stain unmixing strongly depends on the method of image acquisition. Currently available multispectral cameras enable color unmixing of single fields of view (FoV), selected by matter experts (e.g. pathologists). Beyond the individual FoV approach, there is an increasing need to process larger regions or whole histopathological sections (whole slide imaging; WSI). Rapid color deconvolution in WSI is a challenge that is only partially solved. We propose a method based on a multilayer perceptron to compute dye-specific stain layers for chromogenic red and brown labeling in WSI.

Frequent high-resolution images will be provided by new satellites such as Venμs, SENTINEL-2 and Landsat Data Continuity Mission. Methods to handle this new type of data are currently developed (see [1] for an example). However, a more frequent observation of the surface of the Earth may be required for some applications. Moreover, the temporal resolution may be reduced by meteorological artifacts. In this work, we propose to take advantage of the higher temporal resolution of satellites with a lower spatial resolution to detect land-cover modification at a high spatial resolution. The proposed approach does not use any fusion step of the high- and low-resolution images. We show that the low spatial resolution satellite image time series (SITS) can be used in order to inform about the stability and relevance of the high spatial resolution classification. Experiments include a wide variety of resolution ratios and study the use of each ratio for the assessment of high resolution classification maps (computed from the high spatial resolution SITS).

In order to maintain and improve the quality of life of elderly and people with cognitive impairments in their home, we must elaborate assistive technologies that will alleviate the effects of cognitive decline. The use of ambient intelligence, such as smart homes, is a way to provide assistance services. However, managing these services is not a trivial task. In this paper, we propose an organizational centered multi-agent system (OCMAS) architecture in order to manage these services in a smart home. This approach allows to ease the management of services and the deployment of new services in the smart home system dynamically.

Image classification methods based on histogram analysis generally require to use relevant distances for histogram comparison. In this article, we propose a new distance devoted to compare histograms associated to semantic concepts linked by (dis)similarity correlations. This distance, whose computation relies on a hierarchical strategy, captures the multilevel semantic relations between these concepts. It also inherits from the low complexity properties of standard bin-to-bin distances, thus leading to fast and accurate results in the context of multiresolution image classification. Experiments performed on satellite images emphasize the relevance and usefulness of the proposed distance.

The extraction of complex urban patterns from Very High Spatial Resolution (VHSR) images presents several challenges related to the complexity of the data. Based on the availability of images of a same scene at various resolutions (Medium to Very High Spatial resolutions), a hierarchical approach has been recently proposed to segment/classify objects of interest in a top-down fashion in order to determine patterns from VHSR images. To perform, this method requires the interactive definition of segmentation examples for each considered resolution image. In the context of large dataset processing, such interactive task becomes time consuming. To deal with this issue, we propose in this article, an extension of the domain adaptation paradigm enabling the transfer of the segmentation examples defined on a source dataset to automatically process a target one. Experiments performed on urban images provide satisfactory results which may be further used for operational needs.

Satellite Image Time Series are becoming increasingly available and will continue to do so in the coming years thanks to the launch of space missions which aim at providing a coverage of the Earth every few days with high spatial resolution. In the case of optical imagery, it will be possible to produce land use and cover change maps with detailed nomenclatures. It has been shown that the Dynamic Time Warping similarity measure is a consistent tool for the comparison of radiometric profiles of temporal evolution. Actually, it makes it possible to compare time series with both different lengths and different sampling. This property allows us to make the most of partially cloud-covered images, but also to transfer the knowledge learned on an agronomical year in order to classify the next year without using reference data. This article pursues this work on satellite image time series analysis and focuses on the introduction of constraints in the distance in order to fit to the expert's knowledge about the observed phenomena.

Satellite Image Time Series are becoming increasingly available and will continue to do so in the coming years thanks to the launch of space missions which aim at providing a coverage of the Earth every few days with high spatial resolution. In the case of optical imagery, it will be possible to produce land use and cover change maps with detailed nomenclatures. It has been shown that the Dynamic Time Warping (DTW) similarity measure makes it possible to compare radiometric time series with different lengths and sampling. This work aims at showing that DTW is also able to capture distorted phenomena sensed over a long satellite image time series. This article details the analysis of a satellite image time series sensed over 20 years; we show that DTW makes it possible to extract static phenomena, as well as distorted ones such as urbanized areas.

Satellite Image Time Series (SITS, for short) are useful resources for Earth monitoring. Upcoming satellites will provide a global coverage of the Earth's surface with a short revisit time (five days); a huge amount of data to analyze will be produced. In order to be able to analyze efficiently and accurately these images, new methods have to be designed. In this article, we propose to combine a spatio-temporal segmentation pre-processing method - quasi-flat zones, which have been recently extended to video analysis - and the distortion power of DTW to simplify the representation of the SITS, in order to reduce both the time and the memory consumption. Experiments carried out on a series of 46 images show that the memory consumption can be reduced by an order of magnitude without reducing the relevance of the analysis.

Satellite Image Time Series are becoming increasingly available and will continue to do so in the coming years thanks to the launch of space missions which aim at providing a coverage of the Earth every few days with high spatial resolution. In the case of optical imagery, it will be possible to produce land use and cover change maps with detailed nomenclatures. However, due to meteorological phenomena, such as clouds, these time series will become irregular in terms of temporal sampling, and one will need to compare time series with different lengths. In this paper, we present an approach to image time series analysis which is able to deal with irregularly sampled series and which also allows the comparison of pairs of time series where each element of the pair has a different number of samples. We present the dynamic time warping from a theoretical point of view and illustrate its capabilities with two applications to real-time series.

Satellite Image Time Series (SITS) analysis is an important domain with various applications in land study. In the coming years, both high temporal and high spatial resolution SITS will be available. This article aims at providing both temporal and spatial analysis of SITS. We propose first segmenting each image of the series, and then using these segmentations in order to characterize each pixel of the data with a spatial dimension (i.e. with contextual information). Providing spatially characterized pixels, pixel-based temporal analysis can be performed. Experiments carried out with this methodology show the relevance of this approach and the significance of the resulting extracted patterns in the context of the analysis of SITS.

Satellite Image Time Series are becoming increasingly available and will continue to do so in the coming years thanks to the launch of space missions which aim at providing a coverage of the Earth every few days with high spatial resolution. In the case of optical imagery, it will be possible to produce land use and cover change maps with detailed nomenclatures. However, due to meteorological phenomena, such as clouds, these time series will become irregular in terms of temporal sampling and one will need to compare time series with different lengths. In this paper we present an approach to image time series analysis which is able to deal with irregularly sampled series and which also allows the comparison of pairs of time series where each element of the pair has a different number of samples. We present the Dynamic Time Warping from a theoretical point of view and illustrate its capabilities for domain adaptation.

Satellite Image Time Series (SITS) provide us with precious information on land cover evolution. By studying these series of images we can both understand the changes of specific areas and discover global phenomena that spread over larger areas. Changes that can occur throughout the sensing time can spread over very long periods and may have different start time and end time depending on the location, which complicates the mining and the analysis of series of images. This work focuses on frequent sequential pattern mining (FSPM) methods, since this family of methods fits the above-mentioned issues. This family of methods consists of finding the most frequent evolution behaviors, and is actually able to extract long-term changes as well as short term ones, whenever the change may start and end. However, applying FSPM methods to SITS implies confronting two main challenges, related to the characteristics of SITS and the domain's constraints. First, satellite images associate multiple measures with a single pixel (the radiometric levels of different wavelengths corresponding to infra-red, red, etc.), which makes the search space multi-dimensional and thus requires specific mining algorithms. Furthermore, the non evolving regions, which are the vast majority and overwhelm the evolving ones, challenge the discovery of these patterns. We propose a SITS mining framework that enables discovery of these patterns despite these constraints and characteristics. Our proposal is inspired from FSPM and provides a relevant visualization principle. Experiments carried out on 35 images sensed over 20 years show the proposed approach makes it possible to extract relevant evolution behaviors.

With the multiplication of satellite images with complementary spatial and spectral resolution, a major issue in the classification process is the simultaneous use of several images. In this context, the objective of this letter is to propose a new method which uses information contained in both spatial resolutions. The main idea is that on one hand, the semantic level associated with an image depends on its spatial resolution, and on the other hand, information given by these images is complementary. The goal of this multiresolution image method is to automatically build a classification using knowledge extracted from both images, by unsupervised way and without preprocessing image fusion. The method is tested by using a Quickbird (2.8 m) and a SPOT-4 (20 m) image on the urban area of Strasbourg (France). The experiments have shown that the results are better than a classical unsupervised classification on each image and comparable to a supervised region-based classification on the high-spatial-resolution image.

This paper presents a way to combine knowledge obtained from a clustering algorithm and from an ontology. Using the both sources of information allows to improve the results of the knowledge discovery process. The basic property of clustering algorithms, which is to group similar objects, is the key of this approach. We use it to extend the knowledge given by an ontology. Indeed, this knowledge can be partial or not enough accurate, and clustering can then be used to fill this lack of information. We also present results and validation in the field of remote sensing image interpretation.

Feature weighting is an aspect of increasing importance in clustering because data are becoming more and more complex. In this paper, we propose new feature weighting methods based on genetic algorithms. These methods use the cost function defined in LKM as a fitness function. We present new methods based on Darwinian, Lamarckian, and Baldwinian evolution. For each one of them, we describe evolutionary and coevolutionary versions. We compare classical hill-climbing optimization with these six genetic algorithms on different datasets. The results show that the proposed methods, except Darwinian methods, are always better than the LKM algorithm.

The multiplication of Very High Resolution (spatial or spectral) remote sensing images appears to be an opportu- nity to identify objects in urban and periurban areas. The classification methods applied in the object-oriented image analysis approach could be based on the use of domain knowledge. A major issue in these approaches is domain knowledge formalization and exploitation. In this paper, we propose a recognition method based on an ontology which has been developed by experts of the domain. In order to give objects a semantic meaning, we have developed a matching process between an object and the concepts of the ontology. Experiments are made on a Quickbird image. The quality of the results shows the effectiveness of the proposed method.

This article deals with the design of a hybrid learning system. This system integrates different kinds of unsupervised learning methods and gives a set of class hierarchies as the result. The classes in these hierarchies are very similar. The method occurrences compare their results and automatically refine them to try to make them converge towards a unique hierarchy that unifies all the results. Thus, the system decreases the importance of the initial choices made when initializing an unsupervised learning (the choice of the method and its parameters) and to solve some of the limitations of the methods used such as an imposed number of classes, a non-hierarchical result, or the size of the hierarchy.