A5102914203- 3D Surveying and Cultural Heritage
- Advanced Image and Video Retrieval Techniques
- Infrastructure Maintenance and Monitoring
- Precipitation Measurement and Analysis
- Remote-Sensing Image Classification
- Advanced Vision and Imaging
- Meteorological Phenomena and Simulations
- Video Surveillance and Tracking Methods
- Conservation Techniques and Studies
- Non-Destructive Testing Techniques
- Video Analysis and Summarization
- Remote Sensing and LiDAR Applications
- Tensor decomposition and applications
- Image Retrieval and Classification Techniques
- COVID-19 diagnosis using AI
- Radiomics and Machine Learning in Medical Imaging
- Robotics and Sensor-Based Localization
- Advanced Data Compression Techniques
- Air Quality Monitoring and Forecasting
- Climate variability and models
- Sparse and Compressive Sensing Techniques
- Augmented Reality Applications
- Image and Video Quality Assessment
- AI in cancer detection
- Video Coding and Compression Technologies
National Technical University of Athens
2016-2025
Institute of Communication and Computer Systems
2016-2023
National and Kapodistrian University of Athens
1999-2021
Faculty (United Kingdom)
2021
Cyprus Research and Innovation Center (Cyprus)
2018
Technical University of Crete
2008-2014
University of Crete
2009
Energy disaggregation, or Non-Intrusive Load Monitoring (NILM), describes various processes aiming to identify the individual contribution of appliances, given aggregate power signal. In this paper, a non-causal adaptive context-aware bidirectional deep learning model for energy disaggregation is introduced. The proposed model, CoBiLSTM, harnesses representational recurrent Long Short-Term Memory (LSTM) neural networks, while fitting two basic properties NILM problem which state art methods...
An interpretable deep learning framework for land use and cover classification (LULC) in remote sensing using SHAP is introduced. It utilizes a compact CNN model the of satellite images then feeds results to explainer so as strengthen results. The proposed applied Sentinel-2 containing 27000 pixel size 64 × operates on three-band combinations, reducing model's input data by 77% considering that 13 channels are available, while at same time investigating how different spectrum bands affect...
In this work, we present tensor-based linear and nonlinear models for hyperspectral data classification analysis. By exploiting principles of tensor algebra, introduce new architectures, the weight parameters which satisfies {\it rank}-1 canonical decomposition property. Then, learning algorithms to train both non-linear classifier in a way i) minimize error over training samples ii) coefficients The advantages proposed model is that it reduces number required thus respective properly model,...
Understanding the dynamics of deforestation and land uses neighboring areas is vital importance for design development appropriate forest conservation management policies. In this article, we approach as a multilabel classification (MLC) problem in an endeavor to capture various relevant from satellite images. To end, propose vision transformer model, ForestViT, which leverages benefits self-attention mechanism, obviating any convolution operations involved commonly used deep learning models...
Non-Intrusive Load Monitoring (NILM) describes the process of inferring consumption pattern appliances by only having access to aggregated household signal. Sequence-to-sequence deep learning models have been firmly established as state-of-the-art approaches for NILM, in an attempt identify appliance power signal into Exceeding limitations recurrent that widely used sequential modeling, this paper proposes a transformer-based architecture NILM. Our approach, called ELECTRIcity, utilizes...
The number of digital images that are available online today has reached unprecedented levels. Recent statistics showed by the end 2013 there were over 250 billion photographs stored in just one major social media sites, with a daily average upload 300 million photos. These photos, apart from documenting personal lives, often relate to experiences well-known places cultural interest, throughout several periods time. Thus viewpoint Cultural Heritage professionals, they constitute valuable and...
Recent studies indicated that detecting radiographic patterns on CT chest scans can yield high sensitivity and specificity for COVID-19 detection. In this work, we scrutinize the effectiveness of deep learning models semantic segmentation pneumonia infected area in images detection COVID-19. We explore efficacy U-Nets Fully Convolutional Neural Networks task using real-world data from patients. The results indicate are capable accurate despite class imbalance dataset man-made annotation...
Abstract Over the past few years, positron emission tomography/computed tomography (PET/CT) imaging for computer-aided diagnosis has received increasing attention. Supervised deep learning architectures are usually employed detection of abnormalities, with anatomical localization, especially in case CT scans. However, main limitations supervised paradigm include (i) large amounts data required model training, and (ii) assumption fixed network weights upon training completion, implying that...
Non-Intrusive Load Monitoring (NILM) describes the extraction of individual consumption pattern a domestic appliance from aggregated household consumption.Nowadays, NILM research focus is shifted towards practical applications, such as edge deployment, to accelerate transition greener energy future.NILM applications at eliminate privacy concerns and data transmission-related problems.However, resource restrictions pose additional challenges NILM.NILM approaches are usually not designed run...
Targeted nature-based small-scale interventions is an approach commonly adopted by urban developers. The public acceptance of their implementation could be improved participation, emphasizing residents or shopkeepers located close to the areas interest. In this work, we propose a methodology that combines 3D technology, based on open data sources, user-generated content, software and game engines for both minimizing time cost whole planning process enhancing citizen participation. proposed...
An adaptive algorithm for extracting foreground objects from background in videophone or videoconference applications is presented. The uses a neural network architecture that classifies the video frames regions of interest (ROI) and non-ROI areas, also being able to automatically adapt its performance scene changes. incorporated motion-compensated discrete cosine transform (MC-DCT)-based coding schemes, allocating more bits ROI than areas. Simulation results are presented, using Claire...
Abstract. The advent of technology in digital cameras and their incorporation into virtually any smart mobile device has led to an explosion the number photographs taken every day. Today, images stored online available freely reached unprecedented levels. It is estimated that 2011, there were over 100 billion just one major social media sites. This growing exponentially. Moreover, advances fields Photogrammetry Computer Vision have significant breakthroughs such as Structure from Motion...
Intangible Cultural Heritage is a mainspring of cultural diversity and as such it should be focal point in heritage preservation safeguarding endeavours. Nevertheless, although significant progress has been made digitization technology regards tangible assets especially the area 3D reconstruction, e-documentation intangible not seen comparable progress. One main reasons associated lies challenges involved systematic e-digitisation assets, performing arts. In this paper, we present at...
Modeling ionospheric variability throughout a proper total electron content (TEC) parameter estimation is demanding, however, crucial, process for achieving better accuracy and rapid convergence in precise point positioning (PPP). In particular, the single-frequency PPP (SF-PPP) method lacks due to difficulty of dealing adequately with error sources. order apply ionosphere corrections techniques, such as SF-PPP, external information global maps (GIMs) crucial. this article, we propose deep...
A context-aware adaptive data cube (C2A-DC) framework based on Earth Observation (EO) for environmental monitoring to mitigate Climate Change (CC) effects is proposed. It has the property of combining DC formation, calculation Remote Sensing (RS) operations, deep and machine learning algorithms classification harmonization by updating layers using information obtained from previous all applied EO originating a satellite selected stakeholder. Moreover, proposed in sense that it allows tasks...
Abstract. Outdoor large-scale cultural sites are mostly sensitive to environmental, natural and human made factors, implying an imminent need for a spatio-temporal assessment identify regions of potential interest (material degradation, structuring, conservation). On the other hand, in Cultural Heritage research quite different actors involved (archaeologists, curators, conservators, simple users) each diverse needs. All these statements advocate that 5D modelling (3D geometry plus time...
Abstract Recent studies indicated that detecting radiographic patterns on CT chest scans can yield high sensitivity and specificity for COVID-19 detection. In this work, we scrutinize the effectiveness of deep learning models semantic segmentation pneumonia infected area in images detection COVID-19. We explore efficacy U-Nets Fully Convolutional Neural Networks task using real-world data from patients. The results indicate are capable accurate despite class imbalance dataset man-made...
Merging satellite products and ground-based measurements is often required for obtaining precipitation datasets that simultaneously cover large regions with high density are more accurate than pure products. Machine statistical learning regression algorithms regularly utilized in this endeavor. At the same time, tree-based ensemble adopted various fields solving problems accuracy low computational costs. Still, information on which algorithm to select correcting contiguous United States (US)...