Scholarly Works
Research Output
Recent Publications
Gaze Prediction in Virtual Reality Without Eye Tracking Using Visual and Head Motion Cues
Publishing year: 2026
Conference paper
Lecture Notes in Computer Science, 2026, Volume 16396 LNCS
Scopus
Pages 399-410
Gaze prediction plays a critical role in Virtual Reality (VR) applications by reducing sensor-induced latency and enabling computationally demanding techniques such as foveated rendering, which rely on anticipating user attention. However, direct eye tracking is often unavailable due to hardware limitations or privacy concerns. To address this, we present a novel gaze prediction framework that combines Head-Mounted Display (HMD) motion signals with visual saliency cues derived from video frames. Our method employs UniSal, a lightweight saliency encoder, to extract visual features, which are then fused with HMD motion data and processed through a time-series prediction module. We evaluate two lightweight architectures, TSMixer and LSTM, for forecasting future gaze directions. Experiments on the EHTask dataset, along with deployment on commercial VR hardware, show that our approach consistently outperforms baselines such as Center-of-HMD and Mean Gaze. These results demonstrate the effectiveness of predictive gaze modeling in reducing perceptual lag and enhancing natural interaction in VR environments where direct eye tracking is constrained.
Visual Perception of Obstacles: Do Humans and Machines Focus on the Same Image Features?
Publishing year: 2024
Book chapter
19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, 2024, Volume 4
Repository
Pages 357-364
Contemporary cities are fractured by a growing number of barriers, such as on-going construction and infrastructure damages, which endanger pedestrian safety. Automated detection and recognition of such barriers from visual data has been of particular concern to the research community in recent years. Deep Learning (DL) algorithms are now the dominant approach in visual data analysis, achieving excellent results in a wide range of applications, including obstacle detection. However, explaining the underlying operations of DL models remains a key challenge in gaining significant understanding on how they arrive at their decisions. The use of heatmaps that highlight the focal points in input images that helped the models reach their predictions has emerged as a form of post-hoc explainability for such models. In an effort to gain insights into the learning process of DL models, we studied the similarities between heatmaps generated by a number of architectures trained to detect obstacles on sidewalks in images collected via smartphones, and eye-tracking heatmaps generated by humans as they detect the corresponding obstacles on the same data. Our findings indicate that the focus points of humans more closely align with those of a Vision Transformer architecture, as opposed to the other network architectures we examined in our experiments.
Thematic Modeling of UN Sustainable Development Goals: A Comparative Meta-based Approach
Publishing year: 2023
Conference paper
International Conference on Artificial Intelligence Applications and Innovations, 2023, Volume 676 IFIP
Repository
To effectively monitor academic progress towards United Nations Sustainable Development Goals (SDGs), accurate and up-to-date information is crucial. Thematic modeling with keyword-based queries have emerged as a promising tool for clustering the large amounts of publications available in academic research databases. However, the effectiveness of these bibliometric queries depends on the choice of keywords used, and there is no standardized set of keywords for tracking academic progress towards the SDGs. In this study, we performed a comparative analysis of the most used keyword-based queries, assessing their advantages and disadvantages, and identifying gaps and redundancies. Based on our findings, we performed a meta-based approach to develop a new set of keywords that is both concise and thematically representative.
Performance Assessment of Fine-Tuned Barrier Recognition Models in Varying Conditions
Publishing year: 2023
Conference paper
Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics, 2023, Volume 14185 LNCS
Scopus
Pages 172-181
Walking has been widely promoted by various medical institutions as a major contributor to physical activity that keeps people healthy. However, pedestrian safety remains a critical concern due to barriers present on sidewalks, such as bins, poles, and trees. Although pedestrians are generally cautious, these barriers can pose a significant risk to vulnerable groups, such as the visually impaired and elderly. To address this issue, accurate and robust computer vision models can be used to detect barriers on pedestrian pathways in real-time. In this study, we assess the performance of fine-tuned egocentric barrier recognition models under various conditions, such as lighting variations, angles of view, video frame rates and levels of obstruction. In this context, we collected a dataset of different barriers, and fine-tuned two representative image recognition models, assessing their performances on a set of videos taken from a predefined route. Our findings provide guidelines for retaining model performance for applications using barrier recognition models in varying environmental conditions.
On the Identification of Influential Topics in the Social Sciences Using Citation Analysis
Publishing year: 2023
Conference paper
Repository
In this study, we investigate the influential topics of publications in the field of Social Sciences using a novel citation analysis technique. These publications are indexed by Scopus, spanning the period from 2014 to 2022, and our analysis utilizes a graph-partitioning based topic modelling approach utilizing the citations of a publication, specifically the publication's cited papers titles. We focus on the top 1 % publications per year, based on citation count, considering the citation count metric as a concrete indicator of the influence a publication has on the community. While the study's findings reveal intriguing insights into the impact of the pandemic and the steady increase of innovation management, our primary contribution lies in the methodology we propose and utilize. This methodology can be readily adapted to examine other fields of study and can accommodate diverse types of publication information, such as abstracts, keywords, and more. Our approach involves constructing a bipartite graph that links publication IDs with their corresponding cited publications' titles keywords. By projecting this graph onto the keyword space, we identify and visually present the underlying topics that emerge from the data. This visualization offers a clear representation of the identified topics, enhancing the overall understanding of the research landscape in the Social Sciences field.
Assessing the Importance of Key Terms in UN SDGs Through Citizens Labs
Publishing year: 2023
Conference paper
Repository
This study employs a novel methodology to assess the importance of key terms in the United Nations (UN) Sustainable Development Goals (SDGs) by incorporating citizens' perspectives through keyword evaluation in citizens' labs. Representative keywords, sourced from reliable references like the Cyprus University of Technology SDG Keyword Mapping (CUT), are initially selected to represent the key concepts and objectives of each SDG. Citizens assess these keywords during a lab session, utilizing an interactive platform like Mentimeter. The participants dynamically provide their input by choosing the most significant SDG keywords according to their opinions consequently assigning importance weights to them. This approach fosters a more comprehensive understanding of the SDGs and aligns sustainable development initiatives with citizen priorities.
23
Unique publications
56.5%
Open access (13 publications; 1 without information)
43.5%
With other institutions (10 publications; 1 without information)
H-Index Reports
| Index type | Scopus | Google Scholar | Web of Science |
|---|---|---|---|
| H-index | 6 | 0 | 0 |
| H5-index | 5 | 0 | 0 |
| H10-index | 6 | 0 | 0 |
Publications per year and source
No description yet.
- Faculty
- Faculty of Engineering and Technology
- Department
- Department of Electrical Engineering, Computer Engineering and Informatics
- Title
- Research Fellow
