Current students

The detection of anomalies in natural images is of utmost importance in computer vision and has a lot of practical applications, ranging from industrial quality control to medical diagnosis. This task also presents different core challenges (unbalanced dataset, concept drift, selection bias) that researchers are trying to face investigating the application of unsupervised machine learning algorithms. Alongside, in recent years, the advent of deep learning has led towards significant achievements in different computer vision tasks such as image classification, segmentation, and object detection. Thus, this research project aims to apply the latest advanced artificial intelligence techniques to anomaly detection over images, with particular attention to possible industrial applications.