A5024874767- Misinformation and Its Impacts
- Face recognition and analysis
- Neural dynamics and brain function
- Functional Brain Connectivity Studies
- Topic Modeling
- Sentiment Analysis and Opinion Mining
- Generative Adversarial Networks and Image Synthesis
- Hate Speech and Cyberbullying Detection
- Humor Studies and Applications
- Adversarial Robustness in Machine Learning
- Multimodal Machine Learning Applications
- Complex Systems and Time Series Analysis
- Fashion and Cultural Textiles
- Consumer Perception and Purchasing Behavior
- Spam and Phishing Detection
- Visual Attention and Saliency Detection
- Domain Adaptation and Few-Shot Learning
- EEG and Brain-Computer Interfaces
- Biometric Identification and Security
- Explainable Artificial Intelligence (XAI)
- Image Enhancement Techniques
- Digital Media Forensic Detection
- Speech Recognition and Synthesis
- Chaos control and synchronization
- Aesthetic Perception and Analysis
Information Technologies Institute
2019-2025
Centre for Research and Technology Hellas
2019-2025
Harokopio University of Athens
2024
Information Technology Institute
2019-2023
China Philanthropy Research Institute
2021-2023
Aristotle University of Thessaloniki
2014-2019
This article presents FRCSyn-onGoing, an ongoing challenge for face recognition where researchers can easily benchmark their systems against the state of art in open common platform using large-scale public databases and standard experimental protocols. FRCSyn-onGoing is based on Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. first international aiming to explore use real synthetic data independently, also fusion, order address existing limitations technology....
Despite the widespread adoption of face recognition technology around world, and its remarkable performance on current benchmarks, there are still several challenges that must be covered in more detail. This paper offers an overview Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. is first international challenge aiming to explore use synthetic data address existing limitations technology. Specifically, FRCSyn targets concerns related privacy issues, demographic...
Abstract Multimedia content has become ubiquitous on social media platforms, leading to the rise of multimodal misinformation (MM) and urgent need for effective strategies detect prevent its spread. In recent years, challenge detection (MMD) garnered significant attention by researchers mainly involved creation annotated, weakly or synthetically generated training datasets, along with development various deep learning MMD models. However, problem unimodal bias been overlooked, where specific...
Granger causality and variants of this concept allow the study complex dynamical systems as networks constructed from multivariate time series. In work, a large number measures used to form series are assessed. These in domain, such model-based information measures, frequency phase domain. The aims also compare bivariate linear nonlinear well use dimension reduction measures. latter is particular relevant high-dimensional For performance low high dimensional coupled considered discrete...
Abstract Estimating the preferences of consumers is utmost importance for fashion industry as appropriately leveraging this information can be beneficial in terms profit. Trend detection a challenging task due to fast pace change industry. Moreover, forecasting visual popularity new garment designs even more demanding lack historical data. To end, we propose MuQAR, Multimodal Quasi-AutoRegressive deep learning architecture that combines two modules: (1) multimodal multilayer perceptron...
With the expansion of social media and increasing dissemination multimedia content, spread misinformation has become a major concern. This necessitates effective strategies for multimodal detection (MMD) that detect whether combination an image its accompanying text could mislead or misinform. Due to data-intensive nature deep neural networks labor-intensive process manual annotation, researchers have been exploring various methods automatically generating synthetic - which we refer as...
Bias in computer vision systems can perpetuate or even amplify discrimination against certain populations. Considering that bias is often introduced by biased visual datasets, many recent research efforts focus on training fair models using such data. However, most of them heavily rely the availability protected attribute labels dataset, which limits their applicability, while label-unaware approaches, i.e., approaches operating without labels, exhibit considerably lower performance. To...
Hate speech is a societal problem that has significantly grown through the Internet. New forms of digital content such as image memes have given rise to spread hate using multimodal means, being far more difficult analyse and detect compared unimodal case. Accurate automatic processing, analysis understanding this kind will facilitate endeavor hindering proliferation world. To end, we propose MemeFier, deep learning-based architecture for fine-grained classification Internet memes, utilizing...
Transcranial magnetic stimulation combined with electroencephalogram (TMS-EEG) can be used to explore the dynamical state of neuronal networks. In patients epilepsy, TMS induce epileptiform discharges (EDs) a stochastic occurrence despite constant parameters. This observation raises possibility that pre-stimulation period contains multiple covert states brain excitability some which are associated generation EDs.To investigate whether interictal "high excitability" upon produce EDs and...
In patients with Genetic Generalized Epilepsy (GGE), transcranial magnetic stimulation (TMS) can induce epileptiform discharges (EDs) of varying duration. We hypothesized that (a) the ED duration is determined by dynamic states critical network nodes (brain areas) at early post-TMS period, and (b) brain connectivity changes before, during after ED, as well within ED.EEG recordings from two GGE were analyzed. For hypothesis (a), characteristics dynamics stage are measured univariate...
Measures of Granger causality on multivariate time series have been used to form the so-called networks. A network represents interdependence structure underlying dynamical system or coupled systems, and its properties are quantified by indices. In this work, it is investigated whether indices networks generated an appropriate measure can discriminate different coupling structures. The information based partial mutual from mixed embedding (PMIME) networks, a large number ranked according...
The study of connectivity patterns a system's variables, such as multi-channel electroencephalograms (EEG), is utmost importance towards better understanding its internal evolutionary mechanisms. Here, the problem estimating network from multivariate time series in presence prominent unobserved variables addressed. causality measure partial mutual information mixed embedding (PMIME), designed to estimate direct lag-causal effects many observed adapted also zero-lag effects, so-called...
Internet memes are a special type of digital content that is shared through social media. They have recently emerged as popular new format media communication. often multimodal, combining text with images and aim to express humor, irony, sarcasm, or sometimes convey hatred misinformation. Automatically detecting important since it enables tracking cultural trends issues related the spread harmful content. While can take various forms belong different categories, such image macros, labeled...
Out-of-context (OOC) misinformation poses a significant challenge in multimodal fact-checking, where images are paired with texts that misrepresent their original context to support false narratives. Recent research evidence-based OOC detection has seen trend towards increasingly complex architectures, incorporating Transformers, foundation models, and large language models. In this study, we introduce simple yet robust baseline, which assesses MUltimodal SimilaritiEs (MUSE), specifically...
Granger causality and variants of this concept allow the study complex dynamical systems as networks constructed from multivariate time series. In work, a large number measures used to form series are assessed. For this, realizations on high dimensional coupled considered performance is evaluated, seeking for that closest true network system. particular, comparison focuses reduce state space dimension when many variables observed. Further, linear nonlinear reduction compared standard measure...