A5067738598- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Handwritten Text Recognition Techniques
- Medical Image Segmentation Techniques
- Remote-Sensing Image Classification
- Image and Signal Denoising Methods
- Image Processing and 3D Reconstruction
- Neural Networks and Applications
- Earthquake Detection and Analysis
- Image Processing Techniques and Applications
- Natural Language Processing Techniques
- Weed Control and Herbicide Applications
- AI in cancer detection
- Fuzzy Logic and Control Systems
- Text and Document Classification Technologies
- Allelopathy and phytotoxic interactions
- Face and Expression Recognition
- Digital Imaging for Blood Diseases
- Radiomics and Machine Learning in Medical Imaging
- Vehicle License Plate Recognition
- Seismic Waves and Analysis
- Gait Recognition and Analysis
- Semiconductor Lasers and Optical Devices
- Human Pose and Action Recognition
- Video Analysis and Summarization
University of Patras
2014-2023
New York Institute of Technology
2021-2022
Agricultural University of Athens
2005-2021
University of Cyprus
2013
Chinese University of Hong Kong
2013
Applied Multilayers (United Kingdom)
2013
National and Kapodistrian University of Athens
2002-2006
Benaki Phytopathological Institute
2004
University of Bonn
2003
University College London
1984-1987
Automatically measuring a muscle’s cross-sectional area is an important application in clinical practice that has been studied extensively recent years for its ability to assess muscle architecture. Additionally, adequately segmented can be used estimate the echogenicity of muscle, another valuable parameter correlated with quality. This study assesses state-of-the-art convolutional neural networks and vision transformers automating this task new, large, diverse database. database consists...
Abstract The crop system implemented on organic farms requires that residues are incorporated into the soil following end of wheat cultivation. This leads to a reduction in moisture and creation favourable conditions for emergence weeds. In contrast, covering with 60 % was found maintain lead dry weed mass, population density frequency as well diversity, regardless initial flora before implementation different tillage systems. Finally, an increase prevalence broad‐leafed weeds observed level...
Abstract. This paper provides the overview of an integrated, multi-disciplinary effort to study effects 29 March 2006 total solar eclipse on environment, with special focus atmosphere. The has been visible over Eastern Mediterranean, and this occasion several research academic institutes organised co-ordinated experimental campaigns, at different distances from totality various environments in terms air quality. Detailed results findings are presented a number component scientific papers...
Despite the success of hand-crafted features in computer visioning for many years, nowadays, this has been replaced by end-to-end learnable that are extracted from deep convolutional neural networks (CNNs). Whilst CNNs can learn robust directly image pixels, they require large amounts samples and extreme augmentations. On contrary, features, like SIFT, exhibit several interesting properties as provide local rotation invariance. In work, a novel scheme combining strengths SIFT descriptors...
This paper proposes a methodology that incorporates principles from cluster analysis and graph representation to achieve efficient image segmentation results. More specifically, feature-based, inter-region dissimilarity relation is considered here in order determine the matrix graph-based scheme. The calculation of function between adjacent elementary regions based on proximity each region's feature vector main clusters are formed by samples space. In contrast typical approaches literature,...
Autoimmune diseases are proven to be connected with the occurrence of autoantibodies in patient serum. Antinuclear (ANAs) identification can accomplished a laboratory using indirect immunofluorescence (IIF) imaging. ANAs characterized by specific "visual" patterns on humane epithelial cell line (HEp-2). The stage is usually done trained and highly qualified physicians through visual inspection slides fluorescence microscope. presence subjectivity process, interobserver variability,...
Biological cues inherent in human motion play an important role the context of social communication. While recognizing gender other people is for humans, security, advertisement and population statistics systems could also benefit from such kind information. In this work first time we propose a method suitable real gait based recognition relying on poses estimated depth images. We provide evidence that pose representation by images greatly problem analysis. Given sequence, every frame...
A multiresolution color image segmentation approach is presented that incorporates the main principles of region-based and cluster-analysis approaches. The contribution This work may be divided into two parts. In first part, a multiscale dissimilarity measure proposed makes use feature transformation operation to interregion relations with respect their proximity clusters image. As part this process, an original also generate representation information using nonparametric clustering. second...
Frequency ramped diode laser sensing and measurement systems suffer from a variety of limitations noise sources. Nonlinearities in the frequency ramp produce unwanted sidebands spectra system output make accurate distance determination difficult domain. Thermally induced drifts prohibit long-term sensitive phase measurements even with reference interferometer. It is shown that due to fundamental linewidth not bias current determines floor most FMCW regimes away (1/ <tex...
This research work proposes an innovative processing scheme for the exploitation of eye movement dynamics on field biometrical identification. As mechanisms that derive movements highly depend each person's idiosyncrasies, cues reflect at a certain extent individual characteristics may be captured and subsequently deployed implementation robust identification system. Our methodology involves employment non - parametric statistical test, multivariate Wald Wolfowitz test (WW-test), in order to...
There is growing interest in studying the association of functional connectivity patterns with particular cognitive tasks. The ability graphs to encapsulate relational data has been exploited many related studies, where networks (sketched by different neural synchrony estimators) are characterized a rich repertoire graph-related metrics. We introduce commute times (CTs) as an alternative way capture true interplay between nodes graph (FCG). CT measure time taken for random walk setout and...
In this paper, a feature extraction method for offline signature verification is presented that harnesses the power of sparse representation (SR) in order to deliver state-of-the-art performance several datasets like CEDAR, MCYT-75, GPDS, and UTSIG. Beyond accuracy improvements, major parameters associated with SR; such as selected configuration, dictionary size, sparsity level, positivity priors are investigated. Besides, it evinced second-order statistics codes powerful pooling function...
In this work, we present a novel approach for training Generative Adversarial Networks (GANs). Using the attention maps produced by Teacher Network are able to improve quality of generated images as well perform weakly supervised object localization on images. To end, generate HEp-2 cells captured with Indirect Imunofluoresence (IIF) and study ability our network cell. First, demonstrate that whilst GANs can learn mapping between input domain target distribution efficiently, discriminator is...
Automatically delineating the deep and superficial aponeurosis of skeletal muscles from ultrasound images is important in many aspects clinical routine. In particular, finding muscle parameters, such as thickness, fascicle length or pennation angle, a time-consuming task requiring both human labour specialised knowledge. this study, multi-step solution for automating these tasks presented. A process to effortlessly extract automatically measuring thickness has been introduced first step....