2 code implementations • 17 Apr 2024 • Francesco Pro, Nikolaos Dionelis, Luca Maiano, Bertrand Le Saux, Irene Amerini
Nowadays the accurate geo-localization of ground-view images has an important role across domains as diverse as journalism, forensics analysis, transports, and Earth Observation.
1 code implementation • 17 Apr 2024 • Nikolaos Dionelis, Francesco Pro, Luca Maiano, Irene Amerini, Bertrand Le Saux
In this paper, we develop a new model for semantic segmentation of unlabelled images, the Non-annotated Earth Observation Semantic Segmentation (NEOS) model.
1 code implementation • 9 Jan 2024 • Casper Fibaek, Luke Camilleri, Andreas Luyts, Nikolaos Dionelis, Bertrand Le Saux
Massive amounts of unlabelled data are captured by Earth Observation (EO) satellites, with the Sentinel-2 constellation generating 1. 6 TB of data daily.
no code implementations • 30 Nov 2021 • Nikolaos Dionelis, Mehrdad Yaghoobi, Sotirios A. Tsaftaris
By including our boundary, FROB reduces the threshold linked to the model's few-shot robustness; it maintains the OoD performance approximately independent of the number of few-shots.
1 code implementation • 28 Oct 2021 • Nikolaos Dionelis, Mehrdad Yaghoobi, Sotirios A. Tsaftaris
OMASGAN addresses the rarity of anomalies by generating strong and adversarial OoD samples on the distribution boundary using only normal class data, effectively addressing mode collapse.
no code implementations • 29 Sep 2021 • Nikolaos Dionelis, Mehrdad Yaghoobi, Sotirios A. Tsaftaris
We propose a self-supervised learning few-shot confidence boundary methodology based on generative and discriminative models, including classification.
no code implementations • 24 Jul 2021 • Nikolaos Dionelis, Mehrdad Yaghoobi, Sotirios A. Tsaftaris
In this paper, we create a GAN-based tail formation model for anomaly detection, the Tail of distribution GAN (TailGAN), to generate samples on the tail of the data distribution and detect anomalies near the support boundary.
Generative Adversarial Network Unsupervised Anomaly Detection
no code implementations • 21 Jul 2021 • Nikolaos Dionelis, Mehrdad Yaghoobi, Sotirios A. Tsaftaris
We propose an invertible-residual-network-based model, the Boundary of Distribution Support Generator (BDSG).