Foivos Ntelemis

Foivos Ntelemis

Auto-generated profile, based on publicly available information

Foivos Ntelemis

Department of Computer Science, University of Surrey, Guildford, U.K.

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 May 2022 in Knowledge-Based Systems
Verified
Journal Article
Closed access
FNFoivos Ntelemis
Foivos Ntelemis
YJYaochu Jin
Yaochu Jin
STSpencer Thomas
Spencer Thomas

10 June 2021 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
FNFoivos Ntelemis
Foivos Ntelemis
YJYaochu Jin
Yaochu Jin
STSpencer Thomas
Spencer Thomas
Image clustering has recently attracted significant attention due to the increased availability of unlabeled datasets. The efficiency of traditional clustering algorithms heavily depends on the distance functions used and the dimensionality of the features. Therefore, performance degradation is often observed when tackling either unprocessed images or high-dimensional features extracted from processed images. To deal with these challenges, we propose a deep clustering framework consisting of a modified generative adversarial network (GAN) and an auxiliary classifier. The modification employs Sobel operations prior to the discriminator of the GAN to enhance the separability of the learned features. The discriminator is then leveraged to generate representations as to the input to an auxiliary classifier. An objective function is utilized to train the auxiliary classifier by maximizing the mutual information between the representations obtained via the discriminator model and the same representations perturbed via adversarial training. We further improve the robustness of the auxiliary classifier by introducing a penalty term into the objective function. This minimizes the divergence across multiple transformed representations generated by the discriminator model with a low dropout rate. The auxiliary classifier is implemented with a group of multiple cluster-heads, where a tolerance hyper-parameter is used to tackle imbalanced data. Our results indicate that the proposed method achieves competitive results compared with state-of-the-art clustering methods on a wide range of benchmark datasets including CIFAR-10, CIFAR-100/20, and STL10.

1 May 2023 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
FNFoivos Ntelemis
Foivos Ntelemis
YJYaochu Jin
Yaochu Jin
STSpencer Thomas
Spencer Thomas
Self-supervised learning (SSL) has become a popular method for generating invariant representations without the need for human annotations. Nonetheless, the desired invariant representation is achieved by utilizing prior online transformation functions on the input data. As a result, each SSL framework is customized for a particular data type, for example, visual data, and further modifications are required if it is used for other dataset types. On the other hand, autoencoder (AE), which is a generic and widely applicable framework, mainly focuses on dimension reduction and is not suited for learning invariant representation. This article proposes a generic SSL framework based on a constrained self-labeling assignment process that prevents degenerate solutions. Specifically, the prior transformation functions are replaced with a self-transformation mechanism, derived through an unsupervised training process of adversarial training, for imposing invariant representations. Via the self-transformation mechanism, pairs of augmented instances can be generated from the same input data. Finally, a training objective based on contrastive learning is designed by leveraging both the self-labeling assignment and the self-transformation mechanism. Despite the fact that the self-transformation process is very generic, the proposed training strategy outperforms a majority of state-of-the-art representation learning methods based on AE structures. To validate the performance of our method, we conduct experiments on four types of data, namely visual, audio, text, and mass spectrometry data and compare them in terms of four quantitative metrics. Our comparison results demonstrate that the proposed method is effective and robust in identifying patterns within the tested datasets.

14 April 2026 in ESMO Real World Data and Digital Oncology
Verified
Journal Article
Open access
FNF. Ntelemis
F. Ntelemis
SWS. Wolf
S. Wolf
CBC. Bass
C. Bass
AGA. Geraldes
A. Geraldes
JSJ. Schmidt
J. Schmidt
SAS. Arslan
S. Arslan
The assessment of estrogen/progesterone receptors (ER/PR) and human epidermal growth factor receptor 2 (HER2) is essential for managing breast cancer (BC) patients, as these biomarkers guide targeted therapies. Traditional methods such as immunohistochemistry (IHC) and in situ hybridisation (ISH) are widely used, but they are resource-intensive and time-consuming. To address these challenges, we developed PANProfiler Breast (PPB), an artificial intelligence (AI)-based digital test to analyse images of haematoxylin–eosin (H&E)-stained slides for BC biomarker assessment. PPB was validated on a large, diverse cohort of 3138 unique BC samples from five sites across the UK and United States. The validation included both biopsy and resection specimens, digitally scanned using multiple imaging platforms. Standard performance characteristics were assessed by comparison with IHC and ISH reference assays. PPB consistently demonstrated high performance across multiple cohorts, confirming its effectiveness for establishing the status of ER/PR and assessing HER2 negativity for BC patients. HER2 achieved an average sensitivity of 90.6% (±0.6) and a concordance of 93.7% (±7.8%) when determining the negative status. For ER, we measured an overall sensitivity and specificity of 98.2% (±0.3%) and 62.0% (±7.3%), averaged over all sites, respectively. For PR the averaged sensitivity and specificity across all sites were 97.9% (±1.0%) and 45.7% (±1.6%), respectively. Integrating PPB into routine clinical workflows could enable rapid and accurate assessment of BC biomarkers, improving access to targeted therapies and enhancing clinical decision making.