Tutorial “ Anomaly Detection in Images”

Presenter : Giacomo Boracchi, Politecnico di Milano, DEIB
Diego Carrera, STMicroelectronics

Abstract
Anomaly detection problems are ubiquitous in engineering: the prompt detection of anomalies is often a primary concern, since these might provide precious information for understanding the dynamics of a monitored process and for activating suitable countermeasures. In fact, anomalies are typically the most informative samples in an image (e.g., defects in images used for quality control) or in data streams (e.g., arrhythmias in ECG tracing or frauds in credit card transactions). Not surprisingly, detection problems have been widely investigated in the image analysis and pattern recognition communities and are key in application scenarios ranging from quality inspection to health monitoring. The tutorial presents a rigorous formulation of the anomaly-detection problem that fits many image analysis techniques and applications. The tutorial describes in detail the most important approaches in the literature, following the machine-learning perspective of supervised, semi-supervised and unsupervised monitoring tasks. Special emphasis will be given to anomaly detection methods based on learned models, which are often adopted to handle images and signals. These include deep learning models, in particular CNNs, as well as dictionaries yielding (convolutional) sparse representations. The tutorial is accompanied by various examples where anomaly detection algorithms are applied to solve real world problems. These include visual quality inspection algorithms to monitor chip manufacturing and nanofiber production, as well as algorithms to detect arrhythmia in EGC tracings.

Motivation
Detection problems are ubiquitous in science and engineering, spanning in many application domains. Unlike most application-oriented papers, that introduce detection problems without generality, this tutorial presents a rigorous formulation of anomaly detection problems, and provides a unifying view over many signal and image analysis algorithms. As such, the tutorial will be valuable for the many PhD students and researchers that face detection problems in different scenarios, including imaging, health monitoring and detection by classification. The tutorial has the major focus of anomaly detection methods based on learned models.

Slides

ICIAP 2019 Tutorial
Giacomo Boracchi, Diego Carrera Tutorial at ICIAP 2019;
September 9th 2019, Trento, Italy
(Slides);


References

[Carrera et al. 2016 a] ECG Monitoring in Wearable Devices by Sparse Models
Diego Carrera, Beatrice Rossi, Daniele Zambon, Pasqualina Fragneto, and Giacomo Boracchi
, Proceedings of European Conference on Machine Learning and Principles and Practice of Knowledge Discovery, ECML-PKDD 2016, Riva del Garda, Italy, September 19 - 23, Accepted, 16 pages
(Preprint)

[Carrera et al. 2016 b] Defect Detection in SEM Images of Nanofibrous Materials
Diego Carrera, Fabio Manganini, Giacomo Boracchi, Ettore Lanzarone
, IEEE Transactions on Industrial Informatics -- In Press, 11 pages, doi:10.1109/TII.2016.2641472
(Preprint), (Original), (Dataset),







 


 
de -->