Grégoire Lefebvre

In order to exploit representations of time-series signals, such as physiological signals, it is essential that these representations capture relevant information from the whole signal. In this work, we propose to use a Transformer-based model to process electrocardiograms (ECG) for emotion recognition. Attention mechanisms of the Transformer can be used to build contextualized representations for a signal, giving more importance to relevant parts. These representations may then be processed with a fully-connected network to predict emotions. To overcome the relatively small size of datasets with emotional labels, we employ self-supervised learning. We gathered several ECG datasets with no labels of emotion to pre-train our model, which we then fine-tuned for emotion recognition on the AMIGOS dataset. We show that our approach reaches state-of-the-art performances for emotion recognition using ECG signals on AMIGOS. More generally, our experiments show that transformers and pre-training are promising strategies for emotion recognition with physiological signals.

Geomagnetic field fingerprinting is gradually substituting Bluetooth and WiFi fingerprinting since the magnetic field is ubiquitous and independent of any infrastructure. Many studies have used Convolutional Neural Networks (CNNs) to develop indoor positioning systems. Most of these networks use actual magnetic values to build fingerprints. The main source of diminished accuracy is that these CNNs cannot solve the distribution issue of the same magnetic field values. To remedy this limitation, there is a recent interest in applying CNNs to sequences of actual and past data, but no comparative studies have shown the performance contribution of this alternative. In this paper, we propose a CNN-based magnetic fingerprinting system using Recurrence Plots (RPs) as sequence fingerprints. To fairly compare the proposed system with an existing solution treating instantaneous magnetic data, the same real-world data in an indoor environment are used. Testing results show location classification accuracies of 94.92% and 95.46% for the cases of using one RP and three RPs, respectively. As for the localisation error, results show that sequence pattern recognition results in at least a seven-fold decrease in mean distance error.

Autonomy and adaptability are essential components of ambient intelligence. For example, in smart homes, proactive acting and occupants advising, adapted to current and future contexts of living, are essential to go beyond limitations of previous domotic services. To reach such autonomy and adaptability, ambient systems need to automatically grasp their users’ ambient context. In particular, users’ activities and availabilities for communication are valuable pieces of contextual information that can help such systems to adapt to user needs and behaviours. While significant research work exists on activity recognition in homes, less attention has been given to prediction of future activities, as well as to availability recognition and prediction in general. In this article, we investigate several Dynamic Bayesian Network (DBN) architectures for activity and availability prediction of occupants in homes, including our novel model, called Past SItuations to predict the NExt Situation (PSINES). This predictive architecture utilizes context information, sensor event aggregations, and latent user cognitive states to accurately predict future home situations based on previous situations. We experimentally evaluate PSINES, as well as intermediate DBN architectures, on multiple state-of-the-art datasets, with prediction accuracies of up to 89.52% for activity and 82.08% for availability on the Orange4Home dataset.

Notre article décrit les résultats de travaux utilisant l'information de localisation afin de reconnaître les activités domestiques. Nous explorons l'utilisation de l'apprentissage supervisé pour (...)