This work focuses on the topic of machine learning, specifically implementation of a program for automated classification using deep learning. This work compares different trainable models of neural networks and describes practical solutions encountered during their implementation.

The rapid digitalization of the healthcare domain in recent years highlighted the need for advanced predictive methods particularly based upon deep learning methods. Deep learning methods which are capable of dealing with time- series data have recently emerged in various fields such as natural language processing, machine translation, and the Intensive Care Unit (ICU). The recent applications of deep learning in ICU have increasingly received attention, and it has shown promising results for different clinical tasks; however, there is still a need for the benchmark models as far as a handful of public datasets are available in ICU. In this thesis, a novel benchmark model of four clinical tasks on a multi-center publicly available dataset is presented; we employed deep learning models to predict clinical studies. We believe this benchmark model can facilitate and accelerate the research in ICU by allowing other researchers to build on top of it. Moreover, we investigated the effectiveness of the proposed method to predict the risk of delirium in the varying observation and prediction windows, the variable ranking is provided to ease the implementation of a screening tool for helping caregivers at the bedside. Ultimately, an attention-based interpretable neural network is proposed to predict the outcome and rank the most influential variables in the model predictions’ outcome. Our experimental findings show the effectiveness of the proposed approaches in improving the application of deep learning models in daily ICU practice.

Cette thèse de doctorat traite de la reconnaissance automatique du Langage français Parlé Complété (LPC), version française du Cued Speech (CS), à partir de l’image vidéo et sans marquage de l’information préalable à l’enregistrement vidéo. Afin de réaliser cet objectif, nous cherchons à extraire les caractéristiques de haut niveau de trois flux d’information (lèvres, positions de la main et formes), et fusionner ces trois modalités dans une approche optimale pour un système de reconnaissance de LPC robuste. Dans ce travail, nous avons introduit une méthode d’apprentissage profond avec les réseaux neurono convolutifs (CNN)pour extraire les formes de main et de lèvres à partir d’images brutes. Un modèle de mélange de fond adaptatif (ABMM) est proposé pour obtenir la position de la main. De plus, deux nouvelles méthodes nommées Modified Constraint Local Neural Fields (CLNF Modifié) et le model Adaptive Ellipse Model ont été proposées pour extraire les paramètres du contour interne des lèvres (étirement et ouverture aux lèvres). Le premier s’appuie sur une méthode avancée d’apprentissage automatique (CLNF) en vision par ordinateur. Toutes ces méthodes constituent des contributions significatives pour l’extraction de caractéristiques du LPC. En outre, en raison de l’asynchronie des trois flux caractéristiques du LPC, leur fusion est un enjeu important dans cette thèse. Afin de le résoudre, nous avons proposé plusieurs approches, y compris les stratégies de fusion au niveau données et modèle avec une modélisation HMM dépendant du contexte. Pour obtenir le décodage, nous avons proposé trois architectures CNNs-HMMs. Toutes ces architectures sont évaluées sur un corpus de phrases codées en LPC en parole continue sans aucun artifice, et la performance de reconnaissance CS confirme l’efficacité de nos méthodes proposées. Le résultat est comparable à l’état de l’art qui utilisait des bases de données où l’information pertinente était préalablement repérée. En même temps, nous avons réalisé une étude spécifique concernant l’organisation temporelle des mouvements de la main, révélant une avance de la main en relation avec l’emplacement dans la phrase. En résumé, ce travail de doctorat propose les méthodes avancées d’apprentissage automatique issues du domaine de la vision par ordinateur et les méthodologies d’apprentissage en profondeur dans le travail de reconnaissance CS, qui constituent un pas important vers le problème général de conversion automatique de CS en parole audio. / This PhD thesis deals with the automatic continuous Cued Speech (CS) recognition basedon the images of subjects without marking any artificial landmark. In order to realize thisobjective, we extract high level features of three information flows (lips, hand positions andshapes), and find an optimal approach to merging them for a robust CS recognition system.We first introduce a novel and powerful deep learning method based on the ConvolutionalNeural Networks (CNNs) for extracting the hand shape/lips features from raw images. Theadaptive background mixture models (ABMMs) are also applied to obtain the hand positionfeatures for the first time. Meanwhile, based on an advanced machine learning method Modi-fied Constrained Local Neural Fields (CLNF), we propose the Modified CLNF to extract theinner lips parameters (A and B ), as well as another method named adaptive ellipse model. Allthese methods make significant contributions to the feature extraction in CS. Then, due tothe asynchrony problem of three feature flows (i.e., lips, hand shape and hand position) in CS,the fusion of them is a challenging issue. In order to resolve it, we propose several approachesincluding feature-level and model-level fusion strategies combined with the context-dependentHMM. To achieve the CS recognition, we propose three tandem CNNs-HMM architectureswith different fusion types. All these architectures are evaluated on the corpus without anyartifice, and the CS recognition performance confirms the efficiency of our proposed methods.The result is comparable with the state of the art using the corpus with artifices. In parallel,we investigate a specific study about the temporal organization of hand movements in CS,especially about its temporal segmentation, and the evaluations confirm the superior perfor-mance of our methods. In summary, this PhD thesis applies the advanced machine learningmethods to computer vision, and the deep learning methodologies to CS recognition work,which make a significant step to the general automatic conversion problem of CS to sound.The future work will mainly focus on an end-to-end CNN-RNN system which incorporates alanguage model, and an attention mechanism for the multi-modal fusion.

The aim of this work was the development of algorithms for detection segmentation and classification of difficult to define bone metastatic cancerous lesions from spinal CT image data. For this purpose, the patient database was created and annotated by medical experts. Successively, three methods were proposed and developed; the first of them is based on the reworking and combination of methods developed during the preceding project phase, the second method is a fast variant based on the fuzzy k-means cluster analysis, the third method uses modern machine learning algorithms, specifically deep learning of convolutional neural networks. Further, an approach that elaborates the results by a subsequent random forest based meta-analysis of detected lesion candidates was proposed. The achieved results were objectively evaluated and compared with results achieved by algorithms published by other authors. The evaluation was done by two objective methodologies, technical voxel-based and clinical object-based ones. The achieved results were subsequently evaluated and discussed.