A5062716152- Functional Brain Connectivity Studies
- Cell Image Analysis Techniques
- Medical Imaging Techniques and Applications
- Anomaly Detection Techniques and Applications
- Advanced MRI Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Health, Environment, Cognitive Aging
- Advanced Neuroimaging Techniques and Applications
- Machine Learning in Healthcare
- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Biomedical Text Mining and Ontologies
- Bioinformatics and Genomic Networks
Sorbonne Université
2021-2024
Pitié-Salpêtrière Hospital
2021-2024
Centre National de la Recherche Scientifique
2021-2024
Assistance Publique – Hôpitaux de Paris
2021-2024
Institut du Cerveau
2021-2024
Inserm
2021-2024
Sorbonne Paris Cité
2022
Université Sorbonne Nouvelle
2021
Institut national de recherche en informatique et en automatique
2021
We present Clinica ( www.clinica.run ), an open-source software platform designed to make clinical neuroscience studies easier and more reproducible. aims for researchers (i) spend less time on data management processing, (ii) perform reproducible evaluations of their methods, (iii) easily share results within institution with external collaborators. The core is a set automatic pipelines processing analysis multimodal neuroimaging (currently, T1-weighted MRI, diffusion PET data), as well...
Unsupervised anomaly detection is a popular approach for the analysis of neuroimaging data as it allows identifying wide variety anomalies from unlabelled data. It relies on reconstructing subject-specific model healthy appearance to which subject's image can be compared detect anomalies. In literature, common rely analysing residual between real and its pseudo-healthy reconstruction. This however has limitations partly due reconstructions being imperfect lack natural thresholding mechanism....
Unsupervised anomaly detection using deep learning models is a popular computer-aided diagnosis approach because it does not need annotated data and restricted to the of disease seen during training. Such consists in first distribution free images. Images presenting anomalies are then detected as outliers this distribution. These approaches have been widely applied neuroimaging detect sharp localized such tumors or white matter hyper-intensities from structural MRI. In work, we aim FDG PET...
Over the past years, pseudo-healthy reconstruction for unsupervised anomaly detection has gained in popularity. This approach great advantage of not requiring tedious pixel-wise data annotation and offers possibility to generalize any kind anomalies, including that corresponding rare diseases. By training a deep generative model with only images from healthy subjects, will learn reconstruct images. is then compared input detect localize anomalies. The evaluation such methods often relies on...
In this paper, we present ClinicaDL, an open-source software platform that aims at enhancing the reproducibility and rigor of research for deep learning in neuroimaging. We first provide overview then focus on recent advances. Features aim addressing three key issues field: lack reproducibility, methodological flaws plague many published studies difficulties using neuroimaging datasets people with little expertise application area. Key existing functionalities include automatic data...
Unsupervised anomaly detection is a popular approach for the analysis of neuroimaging data as it allows to identify wide variety anomalies from unlabelled data. It relies on building subject-specific model healthy appearance which subject's image can be compared detect anomalies. In literature, common rely analysing residual between and its pseudo-healthy reconstruction. This however has limitations partly due reconstructions being imperfect lack natural thresholding mechanism. Our proposed...