Adeline T M Paiement - Profile on Academia.edu (original) (raw)

Uploads

inproceedings by Adeline T M Paiement

PerCom 2017 (Work in progress session), Mar 2017

Nutrition related health conditions can seriously decrease quality of life; a system able to moni... more Nutrition related health conditions can seriously decrease quality of life; a system able to monitor the kitchen activities and eating behaviour of patients could provide clinicians with important indicators for improving a patient's condition. To achieve this, the system has to reason about the person's actions and goals. To address this challenge, we present a behaviour recognition approach that relies on symbolic behaviour representation and probabilistic reasoning to recognise the person's actions, the type of meal being prepared and its potential impact on a patient's health. We test our approach on a cooking dataset containing unscripted kitchen activities recorded with various sensors in a real kitchen. The results show that the approach is able to recognise the sequence of executed actions and the prepared meal, to determine whether it is healthy, and to reason about the possibility of depression based on the type of meal.

In movement analysis frameworks, body pose may often be adequately represented in a simple, low-d... more In movement analysis frameworks, body pose may often be adequately represented in a simple, low-dimensional, and high-level space, while full body joints locations constitute excessively redundant and complex information. We propose a method for estimating body pose in such highlevel pose spaces directly from a depth image without relying on intermediate skeleton-based steps. Our method is based on a convolutional neural network (CNN) that maps the depth-silhouette of a person to its position in the pose space. We apply our method to a pose representation proposed in that was initially built from skeleton data. We find our estimation of pose to be consistent with the original one and suitable for use in the movement quality assessment framework of . This opens the possibility of a wider application of the movement analysis method to movement types and view-angles that are not supported by the skeleton tracking algorithm.

British Machine Vision Conference (BMVC), Sep 2014

We propose a general method for online estimation of the quality of movement from Kinect skeleton... more We propose a general method for online estimation of the quality of movement from Kinect skeleton data. A robust non-linear manifold learning technique is used to reduce the dimensionality of the noisy skeleton data. Then, a statistical model of normal movement is built from observations of healthy subjects, and the level of matching of new observations with this model is computed on a frame-by-frame basis following Markovian assumptions. The proposed method is validated on the assessment of gait on stairs.

A multi-modal sensor infrastructure for healthcare in a residential environment

2015 IEEE International Conference on Communication Workshop (ICCW), 2015

A multi-modal sensor infrastructure for healthcare in a residential environment

2015 IEEE International Conference on Communication Workshop (ICCW), 2015

Nutrition related health conditions can seriously decrease the quality of life, and a system able... more Nutrition related health conditions can seriously decrease the quality of life, and a system able to monitor the kitchen activities and eating behaviour of patients could provide clinicians important information towards the improvement of the patient's condition. We propose a symbolic model able to describe unscripted kitchen activities and eating behaviour of people in home settings. The model consists of an ontology that describes the problem domain and of a computational state space model that is able to reason in a probabilistic manner about the person's actions, goals, and causes of problems during action execution. To validate our model, we recorded 15 unscripted kitchen tasks involving 9 subjects and manually annotated the video data according to the proposed ontology schema. We then compared the model's ability to recognise people's activities and their goals by generating simulated noisy observations from the annotation of the experiments. The results showed that our model is able to recognise kitchen activities with an average accuracy of 0.8, when using specialised models, and 0.4 when using the general model.

Designing a Video Monitoring System for AAL applications: The SPHERE Case Study

Black Hole Motion as Catalyst of Orbital Resonances

Proceedings of the International Astronomical Union, 2007

ABSTRACT

Medical Image Understanding and Analysis (MIUA), 2010

The use of long-axis images in cardiac MRI segmentation is essential in order to locate the valve... more The use of long-axis images in cardiac MRI segmentation is essential in order to locate the valves and delineate the ventricles' volume accurately. However, depending on the imaging protocol used, long-axis images do not always provide enough support for straightforward segmentation. We show that it is possible to use both short-axis and long-axis images for segmentation, even in cases where the long-axis images do not cover the entire heart volume and have various orientations and spacings, and different gains and contrasts. We propose a method to achieve this goal, based on the simultaneous interpolation and segmentation of the data in a level set framework. Results on both synthetic and real images are presented.

Real-time RGB-D Tracking with Depth Scaling Kernelised Correlation Filters and Occlusion Handling

A Comparative Home Activity Monitoring Study using Visual and Inertial Sensors

Real-time Estimation of Physical Activity Intensity for Daily Living

articles by Adeline T M Paiement

Registration and Modeling From Spaced and Misaligned Image Volumes

Integrated Segmentation and Interpolation of Sparse Data

IEEE Transactions on Image Processing, 2014

We address the two inherently related problems of segmentation and interpolation of 3D and 4D spa... more We address the two inherently related problems of segmentation and interpolation of 3D and 4D sparse data and propose a new method to integrate these stages in a level set framework. The interpolation process uses segmentation information rather than pixel intensities for increased robustness and accuracy. The method supports any spatial configurations of sets of 2D slices having arbitrary positions and orientations. We achieve this by introducing a new level set scheme based on the interpolation of the level set function by radial basis functions. The proposed method is validated quantitatively and/or subjectively on artificial data and MRI and CT scans and is compared against the traditional sequential approach, which interpolates the images first, using a state-of-the-art image interpolation method, and then segments the interpolated volume in 3D or 4D. In our experiments, the proposed framework yielded similar segmentation results to the sequential approach but provided a more robust and accurate interpolation. In particular, the interpolation was more satisfactory in cases of large gaps, due to the method taking into account the global shape of the object, and it recovered better topologies at the extremities of the shapes where the objects disappear from the image slices. As a result, the complete integrated framework provided more satisfactory shape reconstructions than the sequential approach.

Multiple Human Tracking in RGB-D Data: A Survey

A Comparative Study of Pose Representation and Dynamics Modelling for Online Motion Quality Assessment

Regular black hole motion and stellar orbital resonances

Monthly Notices of the Royal Astronomical Society, 2007

ABSTRACT

Integrated Registration, Segmentation, and Interpolation for 3D/4D Sparse Data

Calorie Counter: RGB-Depth Visual Estimation of Energy Expenditure at Home

phdtheses by Adeline T M Paiement

We address the problem of object modelling from 3D and 4D sparse data acquired as different seque... more We address the problem of object modelling from 3D and 4D sparse data acquired as different sequences which are misaligned with respect to each other. Such data may result from various imaging modalities and can therefore present very diverse spatial configurations and appearances. We focus on medical tomographic data, made up of sets of 2D slices having arbitrary positions and orientations, and which may have different gains and contrasts even within the same dataset. The analysis of such tomographic data is essential for establishing a diagnosis or planning surgery.

PerCom 2017 (Work in progress session), Mar 2017

British Machine Vision Conference (BMVC), Sep 2014

A multi-modal sensor infrastructure for healthcare in a residential environment

2015 IEEE International Conference on Communication Workshop (ICCW), 2015

A multi-modal sensor infrastructure for healthcare in a residential environment

2015 IEEE International Conference on Communication Workshop (ICCW), 2015

Designing a Video Monitoring System for AAL applications: The SPHERE Case Study

Black Hole Motion as Catalyst of Orbital Resonances

Proceedings of the International Astronomical Union, 2007

ABSTRACT

Medical Image Understanding and Analysis (MIUA), 2010

Real-time RGB-D Tracking with Depth Scaling Kernelised Correlation Filters and Occlusion Handling

A Comparative Home Activity Monitoring Study using Visual and Inertial Sensors

Real-time Estimation of Physical Activity Intensity for Daily Living

Registration and Modeling From Spaced and Misaligned Image Volumes

Integrated Segmentation and Interpolation of Sparse Data

IEEE Transactions on Image Processing, 2014

Multiple Human Tracking in RGB-D Data: A Survey

A Comparative Study of Pose Representation and Dynamics Modelling for Online Motion Quality Assessment

Regular black hole motion and stellar orbital resonances

Monthly Notices of the Royal Astronomical Society, 2007

ABSTRACT

Integrated Registration, Segmentation, and Interpolation for 3D/4D Sparse Data

Calorie Counter: RGB-Depth Visual Estimation of Energy Expenditure at Home

Interest is growing in using Human-Centered design to enhance compatibility of AI within human en... more Interest is growing in using Human-Centered design to enhance compatibility of AI within human environments. These design techniques are valid for eliciting human-centered design requirements, however, they often paint the scenario that AI interaction design is a one-way process in which user behaviour is captured to improve interactions and user experiences. Such an approach does not consider real-world settings in which Human-AI environments involve multiple stakeholders, with contrasting needs, which could impact the interactivity, usability and usefulness of Human-AI environments. In this paper, we present a framework for incorporating multiple-stakeholders perspectives into the design of Human-AI environments, designed to establish a common dialogue between end-users' needs for Human-AI interaction and AI developers' practical limitations. This is a work in progress project and in our future work we plan to follow this iterative prototyping approach to develop a real-world continuous-learning Human-AI detection system for online grooming.

Synthetically generated clouds on ground-based solar observations

Zenodo (CERN European Organization for Nuclear Research), Feb 28, 2023

Research Square (Research Square), Sep 24, 2023

The study and prediction of space weather entails the analysis of solar images showing structures... more The study and prediction of space weather entails the analysis of solar images showing structures of the Sun's atmosphere. When imaged from the Earth's ground, images may be polluted by terrestrial clouds which hinder the detection of solar structures. We propose a new method to remove cloud shadows, based on a U-Net architecture, and compare classical supervision with conditional GAN. We evaluate our method on two different imaging modalities, using both real images and a new dataset of synthetic clouds. Quantitative assessments are obtained through image quality indices (RMSE, PSNR, SSIM). We demonstrate improved results with regards to the traditional cloud removal technique and a sparse coding baseline, on different cloud types and textures.

2023 18th International Conference on Machine Vision and Applications (MVA)

We explore the use of deep learning to localise galactic structures in low surface brightness (LS... more We explore the use of deep learning to localise galactic structures in low surface brightness (LSB) images. LSB imaging reveals many interesting structures, though these are frequently confused with galactic dust contamination, due to a strong local visual similarity. We propose a novel unified approach to multi-class segmentation of galactic structures and of extended amorphous image contaminants. Our panoptic segmentation model combines Mask R-CNN with a contaminant specialised network and utilises an adaptive preprocessing layer to better capture the subtle features of LSB images. Further, a human-in-the-loop training scheme is employed to augment ground truth labels. These different approaches are evaluated in turn, and together greatly improve the detection of both galactic structures and contaminants in LSB images.

Neural Networks

We explore different strategies to integrate prior domain knowledge into the design of graph neur... more We explore different strategies to integrate prior domain knowledge into the design of graph neural networks (GNN). Our study is supported by a usecase of estimating the potential energy of chemical systems (molecules and crystals) represented as graphs. We integrate two elements of domain knowledge into the design of the GNN to constrain and regularise its learning, towards higher accuracy and generalisation. First, knowledge on the existence of different types of relations/graph edges (e.g. chemical bonds in our case study) between nodes of the graph is used to modulate their interactions. We formulate and compare two strategies, namely specialised message production and specialised update of internal states. Second, knowledge of the relevance of some physical quantities is used to constrain the learnt features towards a higher physical relevance using a simple multi-task learning (MTL) paradigm. We explore the potential of MTL to better capture the underlying mechanisms behind the studied phenomenon. We demonstrate the general applicability of our two knowledge integrations by applying them to three architectures that rely on different mechanisms to propagate information between nodes and to update node states. Our implementations are made publicly available. To support these experiments, we release three new datasets of out-of-equilibrium molecules and crystals of various complexities.

2022 IEEE International Conference on Image Processing (ICIP)

In this paper, we address the challenge of segmenting global contaminants in large images. The pr... more In this paper, we address the challenge of segmenting global contaminants in large images. The precise delineation of such structures requires ample global context alongside understanding of textural patterns. CNNs specialise in the latter, though their ability to generate global features is limited. Attention measures long range dependencies in images, capturing global context, though at a large computational cost. We propose a gridded attention mechanism to address this limitation, greatly increasing efficiency by processing multiscale features into smaller tiles. We also enhance the attention mechanism for increased sensitivity to texture orientation, by measuring correlations across features dependent on different orientations, in addition to channel and positional attention. We present results on a new dataset of astronomical images, where the task is segmenting large contaminating dust clouds.

Detection and parameter estimation of type II solar radio bursts

HAL (Le Centre pour la Communication Scientifique Directe), Sep 16, 2019

Solar active regions localization over multi-spectral observations

HAL (Le Centre pour la Communication Scientifique Directe), Sep 16, 2019

arXiv (Cornell University), Apr 12, 2022

Reconstructions of Lagrangian drift, for example for objects lost at sea, are often uncertain due... more Reconstructions of Lagrangian drift, for example for objects lost at sea, are often uncertain due to unresolved physical phenomena within the data. Uncertainty is usually overcome by introducing stochasticity into the drift, but this approach requires specific assumptions for modelling uncertainty. We remove this constraint by presenting a purely data-driven framework for modelling probabilistic drift in flexible environments. Using ocean circulation model simulations, we generate probabilistic trajectories of object location by simulating uncertainty in the initial object position. We train an emulator of probabilistic drift over one day given perfectly known velocities and observe good agreement with numerical simulations. Several loss functions are tested. Then, we strain our framework by training models where the input information is imperfect. On these harder scenarios, we observe reasonable predictions although the effects of data drift become noticeable when evaluating the models against unseen flow scenarios. Source code and data is available at .

Online grooming (OG) of children is a pervasive issue in an increasingly interconnected world. We... more Online grooming (OG) of children is a pervasive issue in an increasingly interconnected world. We explore various complementary methods to incorporate Corpus Linguistics (CL) knowledge into accurate and interpretable Deep Learning (DL) models. They provide an implicit text normalisation that adapts embedding spaces to the groomers' usage of language, and they focus the DNN's attention onto the expressions of OG strategies. We apply these integrations to two architecture types and improve on the state-of-the-art on a new OG corpus.

SN Computer Science, 2022

Precisely detecting solar active regions (AR) from multi-spectral images is a challenging task ye... more Precisely detecting solar active regions (AR) from multi-spectral images is a challenging task yet important in understanding solar activity and its influence on space weather. A main challenge comes from each modality capturing a different location of these 3D objects, as opposed to more traditional multi-spectral imaging scenarios where all image bands observe the same scene. We present a multi-task deep learning framework that exploits the dependencies between image bands to produce 3D AR detection where different image bands (and physical locations) each have their own set of results. Different feature fusion strategies are investigated in this work, where information from different image modalities is aggregated at different semantic levels throughout the network. This allows the network to benefit from the joint analysis while preserving the band-specific information. We compare our detection method against baseline approaches for solar image analysis (multi-channel coronal ho...

Astronomy & Astrophysics, 2022

Context. The identification and characterization of low surface brightness (LSB) stellar structur... more Context. The identification and characterization of low surface brightness (LSB) stellar structures around galaxies such as tidal debris of ongoing or past collisions is essential to constrain models of galactic evolution. So far most efforts have focused on the numerical census of samples of varying sizes, either through visual inspection or more recently with deep learning. Detailed analyses including photometry have been carried out for a small number of objects, essentially because of the lack of convenient tools able to precisely characterize tidal structures around large samples of galaxies. Aims. Our goal is to characterize in detail, and in particular obtain quantitative measurements, of LSB structures identified in deep images of samples consisting of hundreds of galaxies. Methods. We developed an online annotation tool that enables contributors to delineate the shapes of diffuse extended stellar structures with precision, as well as artifacts or foreground structures. All ...

SPHERE_H130_dataset

The dataset contains both RGB and depth images, and the data from two accelerometers for activity... more The dataset contains both RGB and depth images, and the data from two accelerometers for activity recognition in home environments. The dataset was presented in: L. Tao, T. Burghardt, S. Hannuna, M. Camplani, A. Paiement, D. Damen, M. Mirmehdi, I. Craddock. A Comparative Home Activity Monitoring Study using Visual and Inertial Sensors, 17th International Conference on E-Health Networking, Application and Services, 635-638, 2015. (The size of the repository is relatively big. We suggest users who do not need whole dataset downloading each folder separately.) Access to this dataset is restricted due to the involvement of identifiable participants.

Proceedings of the 10th International Conference on Pattern Recognition Applications and Methods, 2021

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific r... more HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

This document is made available in accordance with publisher policies. Please cite only the publi... more This document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Full terms of use are available:

PAIEMENT ET AL.: ONLINE QUALITY ASSESSMENT OF HUMAN MOVEMENT 1 Online quality assessment of human movement from skeleton data

We propose a general method for online estimation of the quality of movement from Kinect skeleton... more We propose a general method for online estimation of the quality of movement from Kinect skeleton data. A robust non-linear manifold learning technique is used to reduce the dimensionality of the noisy skeleton data. Then, a statistical model of normal move-ment is built from observations of healthy subjects, and the level of matching of new observations with this model is computed on a frame-by-frame basis following Marko-vian assumptions. The proposed method is validated on the assessment of gait on stairs. 1

Simulation of molecular and crystal systems enables insight into interesting chemical properties ... more Simulation of molecular and crystal systems enables insight into interesting chemical properties that benefit processes ranging from drug discovery to material synthesis. However these simulations can be computationally expensive and time consuming despite the approximations through Density Functional Theory (DFT). We propose the Valence Interaction Message Passing Neural Network (VIMPNN) to approximate DFT’s ground-state energy calculations. VIMPNN integrates physics prior knowledge such as the existence of different interatomic bounds to estimate more accurate energies. Furthermore, while many previous machine learning methods consider only stable systems, our proposed method is demonstrated on unstable systems at different atomic distances. VIMPNN predictions can be used to determine the stable configurations of systems, i.e. stable distance for atoms – a necessary step for the future simulation of crystal growth for example. Our method is extensively evaluated on a augmented ver...

SPHERE Unscripted kitchen activities

Environmental sensor data of several people performing unscripted cooking activities in the Spher... more Environmental sensor data of several people performing unscripted cooking activities in the Sphere kitchen. Access to this dataset is restricted and is available to bona fide researchers only. Please email data-bris@bristol.ac.uk to apply for access.

DS-KCF tracker code (SHAPE version)

This repository contains the source code (MATLAB version including shape) of the DS-KCF tracker p... more This repository contains the source code (MATLAB version including shape) of the DS-KCF tracker published in S. Hannuna, M. Camplani, J. Hall, M. Mirmehdi, D. Damen, T. Burghardt, A. Paiement, L. Tao, DS-KCF: A real-time tracker for RGB-D data, Journal of Real-Time Image Processing (2016)

ArXiv, 2021

Deep Neural Networks (DNNs) have often supplied state-of-the-art results in pattern recognition t... more Deep Neural Networks (DNNs) have often supplied state-of-the-art results in pattern recognition tasks. Despite their advances, however, the existence of adversarial examples have caught the attention of the community. Many existing works have proposed methods for searching for adversarial examples within fixed-sized regions around training points. Our work complements and improves these existing approaches by adapting the size of these regions based on the problem complexity and data sampling density. This makes such approaches more appropriate for other types of data and may further improve adversarial training methods by increasing the region sizes without creating incorrect labels.