Damiana Teliti

Damiana Teliti

Auto-generated profile, based on publicly available information

Damiana Teliti


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

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

Representative Publications

View All Publications
2 July 2026 in The Open Neuroimaging Journal
Verified
Journal Article
Open access
PDF file
DTDamiana Teliti
Damiana Teliti
OSOlsi Shehu
Olsi Shehu
BKBekir Karlik
Bekir Karlik
JKJasmin Kevrić
Jasmin Kevrić
Introduction/Objective: Glioblastoma (GBM) is the most lethal primary brain malignancy, with a 12-15-month median survival. This study externally validates quantum-inspired machine learning for GBM risk stratification using the BraTS 2020 cohort, assessing cross-domain and cross-tumor-type generalization. The aim was to validate the quantum radiomics methodology on an independent GBM dataset and evaluate cross-domain generalizability from genomics to radiomics applications. Methods: A total of 75 radiomics features were extracted from multi-modal MRI (T1, T1ce, T2, FLAIR) of 494 GBM patients. Quantum transformation expanded features to 715 dimensions. A calibrated ensemble model was trained for 12-month survival risk stratification with a 60/15/25% train/validation/test split. Results: Test performance achieved 96.80% accuracy [95% CI: 94.92%-98.68%], 99.97% AUC- ROC, 100% precision, 93.55% sensitivity, and 100% specificity. Perfect specificity (zero false positives) enables confident low-risk prediction. Performance outperformed publicized benchmarks: traditional radiomics (73-76%), simple deep learning (82-88%), and complex deep learning models (89-92%). The small validation-test gap (98.65% and 96.80%) also indicates adequate generalization. Discussion: The significant advantage over current methods suggests that quantum-inspired feature transformation is capable of capturing nonlinear patterns in radiomics data, which conventional methodologies fail to do. Perfect specificity has clinical significance, indicating that confident identification of low-risk patients can be achieved for decisions about treatment planning. The out-of-domain success of cross-domain from genomics to radiomics indicates domain independence for the quantum transform. Conclusion: Quantum radiomics generalizes to external validation with clinical-grade performance. Cross-domain validation indicates the domain-agnostic nature of quantum transformation as an improvement to biomedical data and its potential for clinical treatment planning.

21 July 2026 in The Open Neuroimaging Journal
Verified
Journal Article
Open access
PDF file
OSOlsi Shehu
Olsi Shehu
DTDamiana Teliti
Damiana Teliti
BKBekir Karlik
Bekir Karlik
JKJasmin Kevrić
Jasmin Kevrić
Introduction/Objective: Reliable identification of individuals from physiological signals is an active research direction in biometrics. Electroencephalography (EEG) is one such modality because brain signals carry individual-specific patterns. This study investigates whether a deep learning–based EEG identification framework can achieve reliable closed-set individual identification on a 109-subject benchmark under a standard subject-dependent evaluation protocol. Verification scenarios and the associated metrics (FAR, FRR, EER) are out of scope and are not reported. This work proposes and validates an attention-augmented Bi-GRU model for EEG-based individual identification, aiming to achieve high identification accuracy on a large multi-subject public dataset. Methods: Time-domain features were extracted from 64-channel EEG recordings collected during left-hand motor movement/imagery tasks from 109 subjects. A three-layer Bi-GRU model with a self-attention mechanism was trained for a 109-class identification task using a 70/15/15% subject-dependent train–validation–test split. Training robustness was supported through training-only SMOTE-based class balancing and Gaussian-noise augmentation. Results: Under the adopted subject-dependent split, the proposed model achieved a mean test accuracy of 96.68% ± 0.25% across three independent training runs (95% Student-t CI 96.04–97.31%), a top-5 accuracy of 99.37% ± 0.15%, and a small validation–test gap of +0.37 ± 0.90 percentage points. Performance was higher than several traditional machine learning and simpler deep learning baselines and was comparable to reported transformer and graph neural network results. Discussion: The attention mechanism highlighted discriminative temporal patterns from EEG signals; performance under the subject-dependent split indicated that our model was able to maintain stable identification performance across the evaluated recordings. Conclusion: The attention-enhanced Bi-GRU is able to capture discriminative temporal EEG patterns and shows strong identification performance under the adopted protocol, supporting further investigation of EEG-based biometric identification systems.