Olti Qirici

Dr. Olti Qirici

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
9 September 2026 in Diagnostics
Verified
Journal Article
Open access
PDF file
OQOlti Qirici
Olti Qirici
EEEugen Enesi
Eugen Enesi
KBKleona Binjaku
Kleona Binjaku
AKAnduel Kuqi
Anduel Kuqi
IEIndrit Enesi
Indrit Enesi
ADArjan Durresi
Arjan Durresi
Intracranial aneurysm rupture is a major cause of hemorrhagic stroke and a significant contributor to human mortality. With the emergence of artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), a series of new approaches have been introduced to improve aneurysm detection, rupture assessment, risk prediction, and prognosis evaluation. This narrative review synthesizes the findings from 21 peer-reviewed studies identified through searches in PubMed, Scopus, Web of Science, and IEEE Xplore, covering the period from June to September 2025. The reviewed studies examined the use of AI for aneurysm detection, rupture assessment and risk prediction, and prognosis evaluation based on clinical, morphological, radiomic, and hemodynamic data. The reviewed studies applied different AI approaches, including conventional ML algorithms, convolutional neural networks, transformer-based models, radiomics, and multimodal models. Several studies reported encouraging performance in aneurysm detection, segmentation, rupture assessment, and prognosis evaluation, while some approaches showed better performance compared with traditional clinical scoring systems. However, important limitations remain, including the frequent use of small or single-center datasets, limited external validation, and heterogeneity in imaging protocols and feature-extraction methods. AI is showing strong potential in supporting the management of intracranial aneurysms. However, further research is needed, particularly through multicenter and external validation, improved model explainability and robustness, and better integration of these systems into clinical workflows.

1 January 2025
Verified
Conference Paper
Closed access
OQOlti Qirici
Olti Qirici
In the following paper, the solution that was designed and developed will be presented for the Collateral Allocation (and Optimization) process before implementing it, through Mixed Integer Linear Programming. The solution was intended to be used on a broader scope, which arises as a need as part of risk regulatory reporting, and specifically Basel IV. The implementation was done with an open-source library for Mathematical Programming, GLPK, on top of C++ for better performance. The performance measure proved that the solution was fast and reliable, therefore suitable to be used even for purposes other than academic curiosity.

1 January 2025
Verified
Conference Paper
Closed access
OQOlti Qirici
Olti Qirici
In this paper the author is going to present the results of the research for his doctoral degree, where a new technique has been realized and experimented with for the transition from speech to text. In this experiment, the author utilized the self-organizing maps (SOM), Kohonen Neural Networks, to produce a satisfactory result, which would generate text from speech for the Albanian numbering system. The result of such research might be used in IVR systems which still do not exist for the Albanian language. Since for numbers, it's always simpler to produce a satisfactory result because of the simplicity in the language model, it is possible to produce such a system with ease by utilizing the SOM. To produce the desired results and construct the Neural Network, the solution has been developed in Java. As an input to train the system, the author’s voice was utilized, in a 1-hour registration for a 6-hour model training. Considering the limitations in hardware, time and other resources, it was not possible to have additional training material, and the results were limited to the input provided, but good enough to acknowledge such techniques as a pretty good solution to the speech-to-text problem. The system produced stable and convergent results on several learning parameters, and the output was satisfactory. For solutions other than numbering or for a broader number of voices, genders and dialects, we would suggest longer training material and a longer training time which, taking into consideration the results of this experiment, most probably would provide still convergent.