Nazira Abdillaeva
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Nazira Abdillaeva has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Ruslan Isaev
Paragon International University, Cambodia
3 shared publications - Shafee Ur Rehman
Faculty of Medicine, Ala-Too International University, Bishkek, Kyrgyzstan
2 shared publications - A. Habib
Ala-too International University, Kyrgyzstan
1 shared publication - A. Shahzad
Ala-too International University, Kyrgyzstan
1 shared publication - Asel A. Usenova
International Higher School of Medicine, Kyrgyzstan
1 shared publication - Burul Shambetova
Ala-too International University, Kyrgyzstan
1 shared publication
Latest Publications
- 4 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 445-450
Swelling-Aware Face Recognition for Hemodialysis Patients Using Contrastive Deep Learning
Facial recognition technologies offer a promising solution for secure and contactless patient identification in clinical environments, particularly in hemodialysis centers where patients require frequent and repeated visits. However, their reliability is significantly reduced in the presence of facial edema - a common side effect of renal replacement therapy - which alters facial geometry and degrades the accuracy of conventional models trained on healthy faces. This study introduces a clinically relevant and scalable deep learning framework for facial edema classification, forming the foundation of the HEM-ID biometric module for national deployment in Kyrgyzstan. A synthetic dataset was constructed by applying controlled warping and augmentation techniques to a filtered set of 910 neutral facial images, simulating three swelling levels: neutral, mild (swollen1), and moderate (swollen2). Each class was augmented to 1,820 images using affine transformations, color augmentation, and horizontal flips, resulting in a fully balanced dataset of $\mathbf{5, 4 6 0}$ images. We evaluated five deep learning architectures - ResNet50, Efficient-Net-B0, a self-supervised contrastive model (SimCLR), a supervised contrastive model (SupCon), and a custom lightweight CNN named HemEdem. All models were trained using transfer learning, early stopping, and standardized augmentation strategies. Quantitative evaluation shows that SupCon achieved the highest macro F1-score of 0.87 and accuracy of 88%, followed by SimCLR (accuracy 0.82, macro F1 0.80). Baseline CNNs performed lower (ResNet50: accuracy 0.76, macro F1 0.74; EfficientNet-B0: accuracy 0.74, macro F1 0.71), while the lightweight HemEdemCNN reached accuracy 0.68 and macro F1 0.68, reflecting its efficiency-oriented design. The proposed approach demonstrates the feasibility of deploying robust, interpretable, and privacy-preserving facial identification systems in resource-limited clinical settings. HEM-ID offers a scalable solution to ensure secure and reliable patient verification across distributed hemodialysis clinics, even in the presence of dynamic physiological changes.- 4 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 445-450
Swelling-Aware Face Recognition for Hemodialysis Patients Using Contrastive Deep Learning
Facial recognition technologies offer a promising solution for secure and contactless patient identification in clinical environments, particularly in hemodialysis centers where patients require frequent and repeated visits. However, their reliability is significantly reduced in the presence of facial edema - a common side effect of renal replacement therapy - which alters facial geometry and degrades the accuracy of conventional models trained on healthy faces. This study introduces a clinically relevant and scalable deep learning framework for facial edema classification, forming the foundation of the HEM-ID biometric module for national deployment in Kyrgyzstan. A synthetic dataset was constructed by applying controlled warping and augmentation techniques to a filtered set of 910 neutral facial images, simulating three swelling levels: neutral, mild (swollen1), and moderate (swollen2). Each class was augmented to 1,820 images using affine transformations, color augmentation, and horizontal flips, resulting in a fully balanced dataset of $\mathbf{5, 4 6 0}$ images. We evaluated five deep learning architectures - ResNet50, Efficient-Net-B0, a self-supervised contrastive model (SimCLR), a supervised contrastive model (SupCon), and a custom lightweight CNN named HemEdem. All models were trained using transfer learning, early stopping, and standardized augmentation strategies. Quantitative evaluation shows that SupCon achieved the highest macro F1-score of 0.87 and accuracy of 88%, followed by SimCLR (accuracy 0.82, macro F1 0.80). Baseline CNNs performed lower (ResNet50: accuracy 0.76, macro F1 0.74; EfficientNet-B0: accuracy 0.74, macro F1 0.71), while the lightweight HemEdemCNN reached accuracy 0.68 and macro F1 0.68, reflecting its efficiency-oriented design. The proposed approach demonstrates the feasibility of deploying robust, interpretable, and privacy-preserving facial identification systems in resource-limited clinical settings. HEM-ID offers a scalable solution to ensure secure and reliable patient verification across distributed hemodialysis clinics, even in the presence of dynamic physiological changes.- Open Access
- 16 March 2025
- Published by Innovative Research Publishing in International Journal of Innovative Research and Scientific Studies
- Vol. 8 (3) , 2206-2215
Meta-analysis: The Role of AI and Machine Learning in the Management of Hemodialysis Patient Data
Artificial Intelligence (AI) and Machine Learning (ML) technologies transform clinical decision processes in hemodialysis care. The research evaluates AI/ML models through a systematic assessment of their ability to forecast vital clinical outcomes and optimize dialysis treatment. The research team conducted database searches across Google Scholar, PubMed, IEEE Xplore, and Scopus for studies about AI applications in hemodialysis from 2014 through 2024. The research included peer-reviewed clinical studies that presented clear methodologies and performance metrics. The researchers selected 150 studies for inclusion following their full-text evaluation process. The QUADAS-2 tool evaluated study bias while the random-effects model performed the meta-analysis. AI/ML models achieved remarkable accuracy when forecasting mortality (AUC 0.92), hospitalization (accuracy 89%), and intradialytic hypotension (F1-score 0.81). Deep learning and reinforcement learning models achieved significant improvements in dialysis adequacy and access monitoring. The studies revealed data quality problems in 30% of cases while 65% of clinicians expressed doubts about model interpretability. AI/ML technologies demonstrate significant potential to enhance hemodialysis management through predictive modeling and therapy optimization. The successful clinical adoption of these technologies depends on resolving data quality problems and improving transparency and ethical standards.
- 16 March 2025
- Published by Innovative Research Publishing in International Journal of Innovative Research and Scientific Studies
- Vol. 8 (3) , 2206-2215
Open AccessMeta-analysis: The Role of AI and Machine Learning in the Management of Hemodialysis Patient Data
Artificial Intelligence (AI) and Machine Learning (ML) technologies transform clinical decision processes in hemodialysis care. The research evaluates AI/ML models through a systematic assessment of their ability to forecast vital clinical outcomes and optimize dialysis treatment. The research team conducted database searches across Google Scholar, PubMed, IEEE Xplore, and Scopus for studies about AI applications in hemodialysis from 2014 through 2024. The research included peer-reviewed clinical studies that presented clear methodologies and performance metrics. The researchers selected 150 studies for inclusion following their full-text evaluation process. The QUADAS-2 tool evaluated study bias while the random-effects model performed the meta-analysis. AI/ML models achieved remarkable accuracy when forecasting mortality (AUC 0.92), hospitalization (accuracy 89%), and intradialytic hypotension (F1-score 0.81). Deep learning and reinforcement learning models achieved significant improvements in dialysis adequacy and access monitoring. The studies revealed data quality problems in 30% of cases while 65% of clinicians expressed doubts about model interpretability. AI/ML technologies demonstrate significant potential to enhance hemodialysis management through predictive modeling and therapy optimization. The successful clinical adoption of these technologies depends on resolving data quality problems and improving transparency and ethical standards.
- 1 January 2025
- Published by Springer Nature
- p. 314-327
Comparative Analysis of Handwritten Text Recognition Across Cyrillic-Using Countries
- 1 January 2025
- Published by Springer Nature
- p. 314-327
Comparative Analysis of Handwritten Text Recognition Across Cyrillic-Using Countries
- Open Access
- 1 January 2025
- Published by FapUNIFESP (SciELO) in Brazilian Journal of Biology
- Vol. 85, e292947
The development of predictive biomarkers and immunologic markers for breast cancer: current status and future perspectives
Breast cancer is the leading cause of cancer-related mortality among women worldwide. The development of predictive biomarkers and immunologic markers has revolutionized breast cancer diagnosis and treatment, enabling personalized medicine approaches. This study aims to review the current status, challenges, and future directions of predictive and immunologic biomarkers in breast cancer. Predictive biomarkers such as HER2/neu expression, estrogen receptor (ER)/progesterone receptor (PR) status, and genomic signatures play a crucial role in treatment decisions, particularly for hormone receptor-positive and HER2-positive tumors. Immunologic markers, including PD-L1 expression, microsatellite instability (MSI), tumor mutational burden (TMB), and tumor-infiltrating lymphocytes (TILs), are essential for predicting responses to immunotherapy, especially in aggressive subtypes like triple-negative breast cancer (TNBC). However, challenges such as tumor heterogeneity, therapy resistance, and standardization of biomarker testing remain significant barriers. This study also explores emerging technologies like artificial intelligence (AI), genomic profiling, and liquid biopsy, which are expected to enhance precision medicine, improve early diagnosis, and optimize treatment strategies. The integration of these advanced molecular approaches holds great promise for improving patient outcomes and overcoming existing limitations in breast cancer management.
- 1 January 2025
- Published by FapUNIFESP (SciELO) in Brazilian Journal of Biology
- Vol. 85, e292947
Open AccessThe development of predictive biomarkers and immunologic markers for breast cancer: current status and future perspectives
Breast cancer is the leading cause of cancer-related mortality among women worldwide. The development of predictive biomarkers and immunologic markers has revolutionized breast cancer diagnosis and treatment, enabling personalized medicine approaches. This study aims to review the current status, challenges, and future directions of predictive and immunologic biomarkers in breast cancer. Predictive biomarkers such as HER2/neu expression, estrogen receptor (ER)/progesterone receptor (PR) status, and genomic signatures play a crucial role in treatment decisions, particularly for hormone receptor-positive and HER2-positive tumors. Immunologic markers, including PD-L1 expression, microsatellite instability (MSI), tumor mutational burden (TMB), and tumor-infiltrating lymphocytes (TILs), are essential for predicting responses to immunotherapy, especially in aggressive subtypes like triple-negative breast cancer (TNBC). However, challenges such as tumor heterogeneity, therapy resistance, and standardization of biomarker testing remain significant barriers. This study also explores emerging technologies like artificial intelligence (AI), genomic profiling, and liquid biopsy, which are expected to enhance precision medicine, improve early diagnosis, and optimize treatment strategies. The integration of these advanced molecular approaches holds great promise for improving patient outcomes and overcoming existing limitations in breast cancer management.