Self-supervised learning for anomaly detection on time series: application to cellular data - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Self-supervised learning for anomaly detection on time series: application to cellular data

Résumé

This paper presents a new method for anomaly detec-tion in time series and its application to cellular data.These time series are computed from cell images ac-quired thanks to lens-free microscopy. In the context ofcellular biology, detecting abnormal cells is interestingfor any further analysis. Indeed, cells that deviate fromhealthy trajectories can further drive tissues towarddiseases [RAG+20]. It would be both time-consumingand costly to manually analyse each cell in a dataset often thoudands cells. To overcome this human process,we present a deep self-supervised approach to automat-ically detect abnormal cells from their dry mass timeseries. A 1D-convolutio nal neural network is trained topredict the dry mass of cells. An anomaly is detected ifthe mean squared error (MSE) between prediction andground truth is above a fixed threshold. This processbased on self-supervised learning is tested on a datasetof 9,100 time series of dry mass. The method succeedsin detecting abnormal time series with a precision of 96.6%.
Fichier non déposé

Dates et versions

hal-03655782 , version 1 (29-04-2022)

Identifiants

  • HAL Id : hal-03655782 , version 1

Citer

Romain Bailly, Marielle Malfante, Cédric Allier, Lamya Ghenim, Jerome I. Mars. Self-supervised learning for anomaly detection on time series: application to cellular data. Conférence sur L'apprentissage Automatique, Jun 2021, Saint Etienne, France. ⟨hal-03655782⟩
67 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More