Skip to main content
Anomaly detection is the practice of detecting rare patterns in data, which deviate beyond a particular distribution of expected behaviour. Anomalies by definition are rare, Automated mechanisms that are able to detect anomalies would be valuable as they have the potential to significantly reduce human operators time and therefore minimise errors and costs.
Classic machine learning approaches tackle the problem by measuring the distance between data points with respect to the area occupied by the majority of normal data. However, this is more challenging within computer vision and more sophisticated models are required to detect anomalies. My research aims to develop new models for recognising arbitrary anomalies within images and/or videos using combinations of various machine learning and deep learning techniques.
I am PhD student working on Anomaly Detection in Computer Vision using Deep Learning Models. My research thus far has been in Image Anomaly Detection and has recently begun branching into the field of Video Anomaly Detection.
I am also a full-time professional Senior Data Scientist operating in the financial sector, where I develop Deep Learning models for Entity Resolution. I have worked for financial giants such as Barclays Bank and Ernst & Young within the departments of Cards and Payments, Financial Crime, and Entity Resolution & Data Enrichment.
Lappas, Demetris, Argyriou, Vasileios and Makris, Dimitrios (2021) Fourier transformation autoencoders for anomaly detection. In: 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP); 06 - 11 Jun 2021, Toronto, Ontario, Canada (Held online).