Rashik Shahriar Akash
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Rashik Shahriar Akash has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Radiful Islam
Department of Computer Science and Engineering, Daffodil International University, Dhaka, Bangladesh, daffodilvarsity.edu.bd
3 shared publications - Ahmed Wasif Reza
East West University
2 shared publications - Mohammad Shamsul Arefin
Department of Computer Science and Engineering, Chittagong University of Engineering and Technology, Chattogram, Bangladesh
2 shared publications - Abu Bakar Siddiq Tapu
Daffodil International University
1 shared publication - Atik Bhuiyan
Department of Computer Science and Engineering, Daffodil International University, Dhaka, Bangladesh, daffodilvarsity.edu.bd
1 shared publication - Awlad Hossen Rony
Health Informatics Research Lab, Department of Computer Science and Engineering, Daffodil International University, Dhaka, 1216, Bangladesh
1 shared publication
Latest Publications
- Open Access
- 17 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Parameter-Efficient Hybrid Quantum-Classical Learning for Diabetic Retinopathy Grading: Development and External Validation Study (Preprint)
BACKGROUND Diabetic retinopathy (DR) grading becomes more difficult when models are applied to images from new clinical sources. Differences in cameras, image quality, and patient populations can reduce model performance after transfer. Hybrid quantum-classical models can reduce the number of parameters that need to be trained, but previous quantum studies for DR have been generally evaluated on a single dataset. The key question is whether hybrid quantum-classical learning can be useful for cross-domain DR grading performance across new datasets while reducing task-specific training from millions of parameters to thousands. OBJECTIVE We developed a parameter-efficient hybrid quantum-classical model for DR grading and evaluated how well it transfers to external datasets. We compared it with fully fine-tuned classification and regression models and examined overall performance, grade-level errors, and confidence calibration. METHODS We trained all models on APTOS 2019 and tested them without further tuning on MESSIDOR-2 (n=1,744), IDRiD (n=516), and UoA (n=200). Our Hybrid-QC model uses a frozen ConvNeXt-Tiny encoder followed by a small classical projection, an 8-qubit variational quantum circuit, and a final classification layer. Hybrid-QC updates 25,093 parameters, compared with about 28.05 million trainable parameters in the fully fine-tuned ConvNeXt-Tiny classifier (ConvNeXt-Cls) and regression model (ConvNeXt-Reg), approximately 1,118 times fewer. We first evaluated no-DR versus advanced-DR classification and then the more difficult five-class grading. RESULTS For no-DR versus advanced-DR classification, Hybrid-QC achieved an area under the receiver operating characteristic curve (AUROC) of 0.986 on MESSIDOR-2, 0.989 on IDRiD, and 0.995 on UoA. All three models achieved AUROC above 0.98 on every external dataset, and no two models differed by more than 0.01. In five-class grading, Hybrid-QC achieved quadratic weighted kappa (QWK) values close to those of ConvNeXt-Cls on MESSIDOR-2 (0.608 vs 0.618) and IDRiD (0.719 vs 0.717). ConvNeXt-Reg achieved the highest QWK and the lowest mean absolute error on all three external datasets. Hybrid-QC achieved the highest Grade-1 F1 on MESSIDOR-2, IDRiD, and UoA (0.186, 0.162, and 0.400) and predicted fewer Grade-1 cases as Grade 0 than ConvNeXt-Cls. It also had a lower expected calibration error than ConvNeXt-Cls on MESSIDOR-2 (0.047 vs 0.296), IDRiD (0.141 vs 0.373), and UoA (0.100 vs 0.216), and remained better calibrated even after temperature scaling of ConvNeXt-Cls. CONCLUSIONS Hybrid-QC reduces the trainable parameter requirement from about 28 million to about 25 thousand while achieving binary AUROC within 0.01 of the fully fine-tuned models on all three external datasets and five-class QWK close to ConvNeXt-Cls on MESSIDOR-2 and IDRiD. ConvNeXt-Reg remained stronger for overall ordinal grading, while Hybrid-QC achieved the highest Grade-1 F1 on all three external datasets and lower calibration error than ConvNeXt-Cls after domain shift. These results support Hybrid-QC as a parameter-efficient approach to cross-domain DR grading.
- 17 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessParameter-Efficient Hybrid Quantum-Classical Learning for Diabetic Retinopathy Grading: Development and External Validation Study (Preprint)
BACKGROUND Diabetic retinopathy (DR) grading becomes more difficult when models are applied to images from new clinical sources. Differences in cameras, image quality, and patient populations can reduce model performance after transfer. Hybrid quantum-classical models can reduce the number of parameters that need to be trained, but previous quantum studies for DR have been generally evaluated on a single dataset. The key question is whether hybrid quantum-classical learning can be useful for cross-domain DR grading performance across new datasets while reducing task-specific training from millions of parameters to thousands. OBJECTIVE We developed a parameter-efficient hybrid quantum-classical model for DR grading and evaluated how well it transfers to external datasets. We compared it with fully fine-tuned classification and regression models and examined overall performance, grade-level errors, and confidence calibration. METHODS We trained all models on APTOS 2019 and tested them without further tuning on MESSIDOR-2 (n=1,744), IDRiD (n=516), and UoA (n=200). Our Hybrid-QC model uses a frozen ConvNeXt-Tiny encoder followed by a small classical projection, an 8-qubit variational quantum circuit, and a final classification layer. Hybrid-QC updates 25,093 parameters, compared with about 28.05 million trainable parameters in the fully fine-tuned ConvNeXt-Tiny classifier (ConvNeXt-Cls) and regression model (ConvNeXt-Reg), approximately 1,118 times fewer. We first evaluated no-DR versus advanced-DR classification and then the more difficult five-class grading. RESULTS For no-DR versus advanced-DR classification, Hybrid-QC achieved an area under the receiver operating characteristic curve (AUROC) of 0.986 on MESSIDOR-2, 0.989 on IDRiD, and 0.995 on UoA. All three models achieved AUROC above 0.98 on every external dataset, and no two models differed by more than 0.01. In five-class grading, Hybrid-QC achieved quadratic weighted kappa (QWK) values close to those of ConvNeXt-Cls on MESSIDOR-2 (0.608 vs 0.618) and IDRiD (0.719 vs 0.717). ConvNeXt-Reg achieved the highest QWK and the lowest mean absolute error on all three external datasets. Hybrid-QC achieved the highest Grade-1 F1 on MESSIDOR-2, IDRiD, and UoA (0.186, 0.162, and 0.400) and predicted fewer Grade-1 cases as Grade 0 than ConvNeXt-Cls. It also had a lower expected calibration error than ConvNeXt-Cls on MESSIDOR-2 (0.047 vs 0.296), IDRiD (0.141 vs 0.373), and UoA (0.100 vs 0.216), and remained better calibrated even after temperature scaling of ConvNeXt-Cls. CONCLUSIONS Hybrid-QC reduces the trainable parameter requirement from about 28 million to about 25 thousand while achieving binary AUROC within 0.01 of the fully fine-tuned models on all three external datasets and five-class QWK close to ConvNeXt-Cls on MESSIDOR-2 and IDRiD. ConvNeXt-Reg remained stronger for overall ordinal grading, while Hybrid-QC achieved the highest Grade-1 F1 on all three external datasets and lower calibration error than ConvNeXt-Cls after domain shift. These results support Hybrid-QC as a parameter-efficient approach to cross-domain DR grading.
- Open Access
- 1 January 2025
- Published by Wiley in Journal of Tropical Medicine
- Vol. 2025 (1)
Early Dengue Prediction in Bangladesh: A Comparative Study With Feature Analysis, Explainable Artificial Intelligence, and Model Optimization
Dengue fever presents a growing public health challenge in tropical and subtropical regions, where early detection is crucial for effective intervention. This study conducts a comprehensive comparative analysis of 13 machine learning and deep learning models for nonclinical, symptom‐based dengue prediction, focusing on the Bangladeshi population. Using a dataset of 500 patient records with 22 symptom‐based features, we evaluated a wide spectrum of classifier algorithms, including tree‐based (e.g., random forest, extra trees, bagging), linear (logistic regression, SGDClassifier), and an instance‐based classifier. Our comparative evaluation revealed that a custom‐built, hyperparameter‐tuned artificial neural network (ANN) achieved the highest accuracy of 97.5%, significantly outperforming all other models. While tree‐based models like random forest also demonstrated strong performance (93.2%), other classifiers showed considerably lower efficacy. To ensure transparency in our top‐performing model, SHapley Additive exPlanations (SHAP) was employed, identifying critical predictors such as retro‐ocular pain, swollen eyelids, and age. This study not only establishes the superiority of a well‐tuned ANN for this task but also demonstrates the value of broad model comparison and explainability in building reliable diagnostic tools for public health.
- 1 January 2025
- Published by Wiley in Journal of Tropical Medicine
- Vol. 2025 (1)
Open AccessEarly Dengue Prediction in Bangladesh: A Comparative Study With Feature Analysis, Explainable Artificial Intelligence, and Model Optimization
Dengue fever presents a growing public health challenge in tropical and subtropical regions, where early detection is crucial for effective intervention. This study conducts a comprehensive comparative analysis of 13 machine learning and deep learning models for nonclinical, symptom‐based dengue prediction, focusing on the Bangladeshi population. Using a dataset of 500 patient records with 22 symptom‐based features, we evaluated a wide spectrum of classifier algorithms, including tree‐based (e.g., random forest, extra trees, bagging), linear (logistic regression, SGDClassifier), and an instance‐based classifier. Our comparative evaluation revealed that a custom‐built, hyperparameter‐tuned artificial neural network (ANN) achieved the highest accuracy of 97.5%, significantly outperforming all other models. While tree‐based models like random forest also demonstrated strong performance (93.2%), other classifiers showed considerably lower efficacy. To ensure transparency in our top‐performing model, SHapley Additive exPlanations (SHAP) was employed, identifying critical predictors such as retro‐ocular pain, swollen eyelids, and age. This study not only establishes the superiority of a well‐tuned ANN for this task but also demonstrates the value of broad model comparison and explainability in building reliable diagnostic tools for public health.
- Open Access
- 1 January 2024
- Published by SAGE Publications in Digital Health
- Vol. 10
CerviXpert: A multi-structural convolutional neural network for predicting cervix type and cervical cell abnormalities
Objectives: Cervical cancer, a leading cause of cancer-related deaths among women globally, has a significantly higher survival rate when diagnosed early. Traditional diagnostic methods like Pap smears and cervical biopsies rely heavily on the skills of cytologists, making the process prone to errors. This study aims to develop CerviXpert, a multi-structural convolutional neural network designed to classify cervix types and detect cervical cell abnormalities efficiently. Methods: We introduced CerviXpert, a computationally efficient convolutional neural network model that classifies cervical cancer using images from the publicly available SiPaKMeD dataset. Our approach emphasizes simplicity, using a limited number of convolutional layers followed by max-pooling and dense layers, trained from scratch. We compared CerviXpert’s performance against other state-of-the-art convolutional neural network models, including ResNet50, VGG16, MobileNetV2, and InceptionV3, evaluating them on accuracy, computational efficiency, and robustness using five-fold cross-validation. Results: CerviXpert achieved an accuracy of 98.04% in classifying cervical cell abnormalities into three classes (normal, abnormal, and benign) and 98.60% for five-class cervix type classification, outperforming MobileNetV2 and InceptionV3 in both accuracy and computational demands. It demonstrated comparable results to ResNet50 and VGG16, with significantly reduced computational complexity and resource usage. Conclusion: CerviXpert offers a promising solution for efficient cervical cancer screening and diagnosis, striking a balance between accuracy and computational feasibility. Its streamlined architecture makes it suitable for deployment in resource-constrained environments, potentially improving early detection and management of cervical cancer.
- 1 January 2024
- Published by SAGE Publications in Digital Health
- Vol. 10
Open AccessCerviXpert: A multi-structural convolutional neural network for predicting cervix type and cervical cell abnormalities
Objectives: Cervical cancer, a leading cause of cancer-related deaths among women globally, has a significantly higher survival rate when diagnosed early. Traditional diagnostic methods like Pap smears and cervical biopsies rely heavily on the skills of cytologists, making the process prone to errors. This study aims to develop CerviXpert, a multi-structural convolutional neural network designed to classify cervix types and detect cervical cell abnormalities efficiently. Methods: We introduced CerviXpert, a computationally efficient convolutional neural network model that classifies cervical cancer using images from the publicly available SiPaKMeD dataset. Our approach emphasizes simplicity, using a limited number of convolutional layers followed by max-pooling and dense layers, trained from scratch. We compared CerviXpert’s performance against other state-of-the-art convolutional neural network models, including ResNet50, VGG16, MobileNetV2, and InceptionV3, evaluating them on accuracy, computational efficiency, and robustness using five-fold cross-validation. Results: CerviXpert achieved an accuracy of 98.04% in classifying cervical cell abnormalities into three classes (normal, abnormal, and benign) and 98.60% for five-class cervix type classification, outperforming MobileNetV2 and InceptionV3 in both accuracy and computational demands. It demonstrated comparable results to ResNet50 and VGG16, with significantly reduced computational complexity and resource usage. Conclusion: CerviXpert offers a promising solution for efficient cervical cancer screening and diagnosis, striking a balance between accuracy and computational feasibility. Its streamlined architecture makes it suitable for deployment in resource-constrained environments, potentially improving early detection and management of cervical cancer.
- 20 December 2023
- Published by Springer Nature
- p. 181-195
A Review on the Impacts of Social Media on the Mental Health
- 20 December 2023
- Published by Springer Nature
- p. 181-195
A Review on the Impacts of Social Media on the Mental Health
- 15 December 2023
- Published by Springer Nature
- p. 379-391
A Comprehensive Review on Family Budget Management
- 15 December 2023
- Published by Springer Nature
- p. 379-391
A Comprehensive Review on Family Budget Management