Amani Yousef Owda
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Amani Yousef Owda has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Majdi Owda
Faculty of Artificial Intelligence and Data Science, UNESCO Chair in Data Science for Sustainable Development, Arab American University, Ramallah P600, Palestine
24 shared publications - Mobarak Abumohsen
Information Technologies Group, School of Telecommunication Engineering, University of Vigo, 36310 Vigo, Spain
8 shared publications - Neil Anthony Salmon
MMW Sensors Ltd
7 shared publications - Ahmad Abumihsan
Faculty of Artificial Intelligence and Data Science, Arab American University, Ramallah P600, Palestine
4 shared publications Nacer Ddine Rezgui
4 shared publications- Alexander J. Casson
Department of Electrical and Electronic Engineering, University of Manchester, Manchester, UK
3 shared publications
Latest Publications
- Open Access
A Novel Hybrid Deep Learning Algorithm for DDoS Attack Detection on Software-Defined Networking
Software-defined networking (SDN) is an advanced network architecture designed to transform conventional networks to meet evolving requirements. Unlike traditional networks, SDN separates the control and data planes, making it efficient to monitor, configure, and optimize network resources. However, due to its centralized nature, SDN is prone to several attack vectors. Among them, distributed denial of service (DDoS) attacks are the most potent threats to the SDN controller. This paper presents an approach to detecting DDoS attacks on SDN controllers using a detection model based on a parallel one-dimensional convolutional neural network-gated recurrent unit (1D CNN-GRU) architecture equipped with a hybrid feature selection approach. The suggested hybrid feature selection combines the filter feature methods (mutual information, chi-squared, and ANOVA) with a self-attention mechanism to effectively identify the most vital features for DDoS attack detection in SDN environments. The proposed detection model 1D CNN-GRU model leverages the strengths of both a convolutional neural network (CNN) and a gated recurrent unit (GRU). The CNN captures spatial and local patterns, while the GRU processes feature dimensions sequentially within each flow record. The proposed approach demonstrated high levels of accuracy (97.836%), precision (98.094%), and F1 measure (97.252%). Additionally, it exhibited a low false positive rate (FPR) of 1.228% and a short detection time of 1.192 s. The approach was also evaluated on an additional recent SDN-IoT DDoS dataset to reduce reliance on a single dataset.
Open AccessA Novel Hybrid Deep Learning Algorithm for DDoS Attack Detection on Software-Defined Networking
Software-defined networking (SDN) is an advanced network architecture designed to transform conventional networks to meet evolving requirements. Unlike traditional networks, SDN separates the control and data planes, making it efficient to monitor, configure, and optimize network resources. However, due to its centralized nature, SDN is prone to several attack vectors. Among them, distributed denial of service (DDoS) attacks are the most potent threats to the SDN controller. This paper presents an approach to detecting DDoS attacks on SDN controllers using a detection model based on a parallel one-dimensional convolutional neural network-gated recurrent unit (1D CNN-GRU) architecture equipped with a hybrid feature selection approach. The suggested hybrid feature selection combines the filter feature methods (mutual information, chi-squared, and ANOVA) with a self-attention mechanism to effectively identify the most vital features for DDoS attack detection in SDN environments. The proposed detection model 1D CNN-GRU model leverages the strengths of both a convolutional neural network (CNN) and a gated recurrent unit (GRU). The CNN captures spatial and local patterns, while the GRU processes feature dimensions sequentially within each flow record. The proposed approach demonstrated high levels of accuracy (97.836%), precision (98.094%), and F1 measure (97.252%). Additionally, it exhibited a low false positive rate (FPR) of 1.228% and a short detection time of 1.192 s. The approach was also evaluated on an additional recent SDN-IoT DDoS dataset to reduce reliance on a single dataset.
- Open Access
- 22 October 2026
- Published by MDPI AG in Preprints.org
A Mask-Guided, Explainable Deep-Learning Framework with Verifiable Attention for Benign–Malignant Classification of Breast Ultrasound Images
Background/Objectives: Breast cancer is one of the most common cancer types in women, and reading a breast ultrasound is still very much dependent on the individual radiologist, creating inter-observer variability. Deep-learning classifiers can achieve accuracy comparable to that of experts, but there are two challenges for their use in routine practice: they rarely explain their predictions, and a model can reach the right label for the wrong reason, by attending to tissue unrelated to the lesion. This study addresses both of these. Methods: On the BUS-BRA dataset, we used the lesion segmentation mask as an additional input channel to the network, along with the RGB image, so two backbones (EfficientNet-B3 and EfficientNet-B4) were directed towards the lesion from the first layer, while only the initial convolution layer was modified, keeping the ImageNet-pretrained weights unchanged. To ensure a fair evaluation, the data were split at the patient level rather than the image level, preventing leakage. Training was performed with focal loss, heavy augmentation and test-time augmentation, with the decision threshold set on the validation set. Grad-CAM, SHAP, and Occlusion Sensitivity explained each prediction, and were scored for fidelity, stability, and overlap with the true lesion. Results: During the controlled ablation, the mask channel improved accuracy by more than eight points over an region-of-interest (ROI)-only baseline. The average of the two backbone models yielded 86.7% accuracy and 0.911 AUC on the held-out test set, and when applied to the independent BUSI dataset, without any modification, with the threshold frozen, it achieved 90.4% accuracy and 0.972 AUC, with Grad-CAM maps landing on the annotated lesions. Conclusions: The accuracy of mask guidance using an imbalance-aware design is competitive, and the accuracy can be transferred to an external dataset, providing a credible basis for decision support that can be verified by the clinician with the image.
- 22 October 2026
- Published by MDPI AG in Preprints.org
Open AccessA Mask-Guided, Explainable Deep-Learning Framework with Verifiable Attention for Benign–Malignant Classification of Breast Ultrasound Images
Background/Objectives: Breast cancer is one of the most common cancer types in women, and reading a breast ultrasound is still very much dependent on the individual radiologist, creating inter-observer variability. Deep-learning classifiers can achieve accuracy comparable to that of experts, but there are two challenges for their use in routine practice: they rarely explain their predictions, and a model can reach the right label for the wrong reason, by attending to tissue unrelated to the lesion. This study addresses both of these. Methods: On the BUS-BRA dataset, we used the lesion segmentation mask as an additional input channel to the network, along with the RGB image, so two backbones (EfficientNet-B3 and EfficientNet-B4) were directed towards the lesion from the first layer, while only the initial convolution layer was modified, keeping the ImageNet-pretrained weights unchanged. To ensure a fair evaluation, the data were split at the patient level rather than the image level, preventing leakage. Training was performed with focal loss, heavy augmentation and test-time augmentation, with the decision threshold set on the validation set. Grad-CAM, SHAP, and Occlusion Sensitivity explained each prediction, and were scored for fidelity, stability, and overlap with the true lesion. Results: During the controlled ablation, the mask channel improved accuracy by more than eight points over an region-of-interest (ROI)-only baseline. The average of the two backbone models yielded 86.7% accuracy and 0.911 AUC on the held-out test set, and when applied to the independent BUSI dataset, without any modification, with the threshold frozen, it achieved 90.4% accuracy and 0.972 AUC, with Grad-CAM maps landing on the annotated lesions. Conclusions: The accuracy of mask guidance using an imbalance-aware design is competitive, and the accuracy can be transferred to an external dataset, providing a credible basis for decision support that can be verified by the clinician with the image.
- Open Access
- 15 October 2026
- Published by MDPI AG in Preprints.org
An Explainable Genetic Algorithm-Optimized Weighted Soft Voting Framework for Heart Disease Prediction
Reliable and interpretable heart disease prediction across heterogeneous clinical cohorts remains a major barrier to clinical translation. This study presents an explainable, genetic-algorithm (GA)-optimized performance-weighted soft-voting (PWSV) framework for binary heart disease prediction using 920 records from the Cleveland, Hungary, Switzerland, and VA Long Beach cohorts. After data harmonization, leakage-aware preprocessing, and feature engineering, 29 encoded predictors were retained. GA independently optimized random forest, extra trees, radial basis function support vector machine, and gradient boosting using five-fold cross-validation confined to the development data. Each model’s mean cross-validated weighted F1-score was normalized into its ensemble weight, keeping both hyperparameter selection and weight estimation independent of the test data. Two complementary evaluation protocols were used: a stratified 80:20 development–test split to assess predictive performance and leave-one-dataset-out (LODO) validation to quantify cross-cohort generalizability. On the untouched test set, PWSV achieved 95.65% accuracy; disease-class precision, recall, and F1-score of 94.34%, 98.04%, and 96.15%, respectively; and a 98.11% area under the receiver operating characteristic curve, outperforming all optimized constituent classifiers. LODO accuracy ranged from 82.15% to 90.64%, revealing cohort-dependent performance variation that a random split alone would not capture. SHAP and LIME provided complementary global and patient-level explanations, and their aggregated feature rankings showed significant agreement (Spearman’s ρ = 0.867, p < 0.001). Overall, the framework integrates leakage-aware optimization, performance-driven probability fusion, source-aware validation, and concordant explanations within a reproducible pipeline. These findings support its potential for clinical decision support, while prospective multicenter validation remains necessary before deployment.
- 15 October 2026
- Published by MDPI AG in Preprints.org
Open AccessAn Explainable Genetic Algorithm-Optimized Weighted Soft Voting Framework for Heart Disease Prediction
Reliable and interpretable heart disease prediction across heterogeneous clinical cohorts remains a major barrier to clinical translation. This study presents an explainable, genetic-algorithm (GA)-optimized performance-weighted soft-voting (PWSV) framework for binary heart disease prediction using 920 records from the Cleveland, Hungary, Switzerland, and VA Long Beach cohorts. After data harmonization, leakage-aware preprocessing, and feature engineering, 29 encoded predictors were retained. GA independently optimized random forest, extra trees, radial basis function support vector machine, and gradient boosting using five-fold cross-validation confined to the development data. Each model’s mean cross-validated weighted F1-score was normalized into its ensemble weight, keeping both hyperparameter selection and weight estimation independent of the test data. Two complementary evaluation protocols were used: a stratified 80:20 development–test split to assess predictive performance and leave-one-dataset-out (LODO) validation to quantify cross-cohort generalizability. On the untouched test set, PWSV achieved 95.65% accuracy; disease-class precision, recall, and F1-score of 94.34%, 98.04%, and 96.15%, respectively; and a 98.11% area under the receiver operating characteristic curve, outperforming all optimized constituent classifiers. LODO accuracy ranged from 82.15% to 90.64%, revealing cohort-dependent performance variation that a random split alone would not capture. SHAP and LIME provided complementary global and patient-level explanations, and their aggregated feature rankings showed significant agreement (Spearman’s ρ = 0.867, p < 0.001). Overall, the framework integrates leakage-aware optimization, performance-driven probability fusion, source-aware validation, and concordant explanations within a reproducible pipeline. These findings support its potential for clinical decision support, while prospective multicenter validation remains necessary before deployment.
- Open Access
- 16 September 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (9) , 484
Understanding Differential Privacy in Decentralized Federated Learning: A Controlled Privacy–Utility Comparison
Centralized Federated Learning (FL) enables collaborative model training without sharing raw data. Differential privacy (DP) is widely used to protect sensitive information in FL; however, its behavior in decentralized environments remains poorly understood. This study empirically compares centralized FL and sequential Decentralized Federated Learning (DFL) under matched clipping and perturbation settings to examine model utility and privacy leakage. Both frameworks were evaluated under a common experimental setup with non-IID data, and each configuration was evaluated across five independent seeds. Utility was evaluated under matched experimental perturbation parameters, whereas formal client-level privacy accounting was applied to the perturbed round-end model releases, with each client’s complete dataset treated as the protected unit. Empirical leakage was evaluated separately using membership inference and gradient inversion attacks. Within the evaluated MNIST configuration, the results show that clipping, perturbation, and the learning procedure jointly influence the observed privacy–utility behavior. At and , the mean final accuracy was for FL and for sequential DFL. At , InvGrad reconstruction for FL produced an MSE of , PSNR of , and SSIM of , compared with , , and for sequential DFL, respectively. Future research will investigate topology-aware privacy mechanisms and adaptive noise allocation for decentralized systems.- 16 September 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (9) , 484
Open AccessUnderstanding Differential Privacy in Decentralized Federated Learning: A Controlled Privacy–Utility Comparison
Centralized Federated Learning (FL) enables collaborative model training without sharing raw data. Differential privacy (DP) is widely used to protect sensitive information in FL; however, its behavior in decentralized environments remains poorly understood. This study empirically compares centralized FL and sequential Decentralized Federated Learning (DFL) under matched clipping and perturbation settings to examine model utility and privacy leakage. Both frameworks were evaluated under a common experimental setup with non-IID data, and each configuration was evaluated across five independent seeds. Utility was evaluated under matched experimental perturbation parameters, whereas formal client-level privacy accounting was applied to the perturbed round-end model releases, with each client’s complete dataset treated as the protected unit. Empirical leakage was evaluated separately using membership inference and gradient inversion attacks. Within the evaluated MNIST configuration, the results show that clipping, perturbation, and the learning procedure jointly influence the observed privacy–utility behavior. At and , the mean final accuracy was for FL and for sequential DFL. At , InvGrad reconstruction for FL produced an MSE of , PSNR of , and SSIM of , compared with , , and for sequential DFL, respectively. Future research will investigate topology-aware privacy mechanisms and adaptive noise allocation for decentralized systems. - Open Access
- 6 August 2026
- Published by MDPI AG in Preprints.org
A Novel Hybrid Deep Learning Algorithm for DDoS Attack Detection on Software-Defined Networking
Software-defined networking (SDN) is an advanced network architecture to transform the conventional network to meet evolving requirements. Unlike traditional networks, SDN breaks down the control and data plane, making it efficient to monitor, configure, and optimize the network resources. However, due to its centralized nature, SDN is prone to several attack vectors. Among them, distributed denial of service (DDoS) attacks are the most potent threats to the SDN controller. This paper presents a novel approach to detect DDoS attacks on SDN controllers using a novel detection model based on a parallel one-dimensional convolutional neural network-gated recurrent unit (1D CNN-GRU) architecture equipped with a novel hybrid feature selection approach. The suggested hybrid feature selection combines the filter feature methods (mutual information, chi-squared, and ANOVA) with a self-attention mechanism to effectively detect the most vital features for DDoS attack detection in SDN environments. The proposed detection model 1D CNN-GRU leverages the strengths of both a convolutional neural network (CNN) and a gated recurrent unit (GRU). The CNN captures spatial and local patterns, while the GRU focuses on temporal dependencies within the data. The proposed approach showcased high levels of accuracy (97.882%), precision of (98.078%), and F1-measure of (97.276%). Additionally, it exhibited a low false positive rate (FPR) of 1.208% and a short detection time of 1.206 seconds.
- 6 August 2026
- Published by MDPI AG in Preprints.org
Open AccessA Novel Hybrid Deep Learning Algorithm for DDoS Attack Detection on Software-Defined Networking
Software-defined networking (SDN) is an advanced network architecture to transform the conventional network to meet evolving requirements. Unlike traditional networks, SDN breaks down the control and data plane, making it efficient to monitor, configure, and optimize the network resources. However, due to its centralized nature, SDN is prone to several attack vectors. Among them, distributed denial of service (DDoS) attacks are the most potent threats to the SDN controller. This paper presents a novel approach to detect DDoS attacks on SDN controllers using a novel detection model based on a parallel one-dimensional convolutional neural network-gated recurrent unit (1D CNN-GRU) architecture equipped with a novel hybrid feature selection approach. The suggested hybrid feature selection combines the filter feature methods (mutual information, chi-squared, and ANOVA) with a self-attention mechanism to effectively detect the most vital features for DDoS attack detection in SDN environments. The proposed detection model 1D CNN-GRU leverages the strengths of both a convolutional neural network (CNN) and a gated recurrent unit (GRU). The CNN captures spatial and local patterns, while the GRU focuses on temporal dependencies within the data. The proposed approach showcased high levels of accuracy (97.882%), precision of (98.078%), and F1-measure of (97.276%). Additionally, it exhibited a low false positive rate (FPR) of 1.208% and a short detection time of 1.206 seconds.
- Open Access
- 20 June 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (12) , 6224
Beta Normalization Aggregation-Based Ensemble Learning for Lung Cancer Classification: Evaluation on CT and Histopathological Images
The early and accurate detection of lung cancer (LC) is one of the primary challenges in the clinical diagnostics process, which plays a vital role in the treatment of the disease. Although various deep learning (DL) techniques have been presented, the existing DL methods are mainly focused on single-modal images, either computed tomography (CT) or histopathological images, which are associated with poor generalization, diversity, and applicability. To mitigate the existing issues, the present work aims to develop a modality-independent ensemble DL framework that is independently evaluated on CT and histopathological image datasets for LC classification. In this work, the proposed framework was developed using the Beta Normalization Aggregation (BNA) technique, where the performance of three state-of-the-art pre-trained convolutional neural network (CNN) architectures was compared on two distinct imaging modalities images. Based on the comparative analysis of the performance metrics, Xception, DenseNet121, and MobileNetV2, are chosen to develop the Ensemble model. Predictions generated by the selected CNN models are aggregated using the proposed BNA strategy to improve classification robustness, which improves the confidence of the prediction results and discriminative capabilities. The experiments using public data sets have confirmed the excellent performance of the model. On the CT dataset, the proposed BNA Ensemble achieved a testing accuracy of 97.45%, with a precision of 97.88%, recall of 97.45%, F1-score of 97.45%, and an AUC of 0.9986. On the histopathological dataset, the framework achieved an accuracy of 99.80%, with precision, recall, and F1-score all reaching 99.80%, and an AUC of 1.0000. These results demonstrate the effectiveness, robustness, and generalizability of the proposed BNA framework. The analysis of the results using t-SNE plots, confusion matrices, ROC curves, and confidence distributions provided additional insights into feature separability, classification performance, and prediction confidence of the proposed framework.
- 20 June 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (12) , 6224
Open AccessBeta Normalization Aggregation-Based Ensemble Learning for Lung Cancer Classification: Evaluation on CT and Histopathological Images
The early and accurate detection of lung cancer (LC) is one of the primary challenges in the clinical diagnostics process, which plays a vital role in the treatment of the disease. Although various deep learning (DL) techniques have been presented, the existing DL methods are mainly focused on single-modal images, either computed tomography (CT) or histopathological images, which are associated with poor generalization, diversity, and applicability. To mitigate the existing issues, the present work aims to develop a modality-independent ensemble DL framework that is independently evaluated on CT and histopathological image datasets for LC classification. In this work, the proposed framework was developed using the Beta Normalization Aggregation (BNA) technique, where the performance of three state-of-the-art pre-trained convolutional neural network (CNN) architectures was compared on two distinct imaging modalities images. Based on the comparative analysis of the performance metrics, Xception, DenseNet121, and MobileNetV2, are chosen to develop the Ensemble model. Predictions generated by the selected CNN models are aggregated using the proposed BNA strategy to improve classification robustness, which improves the confidence of the prediction results and discriminative capabilities. The experiments using public data sets have confirmed the excellent performance of the model. On the CT dataset, the proposed BNA Ensemble achieved a testing accuracy of 97.45%, with a precision of 97.88%, recall of 97.45%, F1-score of 97.45%, and an AUC of 0.9986. On the histopathological dataset, the framework achieved an accuracy of 99.80%, with precision, recall, and F1-score all reaching 99.80%, and an AUC of 1.0000. These results demonstrate the effectiveness, robustness, and generalizability of the proposed BNA framework. The analysis of the results using t-SNE plots, confusion matrices, ROC curves, and confidence distributions provided additional insights into feature separability, classification performance, and prediction confidence of the proposed framework.
- Top Cited PapersOpen Access
A Deep Learning Framework for Ultrasound Image Quality Assessment and Automated Nuchal Translucency Measurement to Improve First-Trimester Chromosomal Abnormality Screening
Background: First-trimester prenatal screening is a fundamental component of modern obstetric care, offering early insights into fetal health and development. A key focus of this screening is the detection of chromosomal abnormalities, such as Trisomy 21 (Down syndrome), which can have significant implications for pregnancy management and parental counseling. Over the years, various non-invasive methods have been developed, with ultrasound-based assessments becoming a cornerstone of early evaluation. Among these, the measurement of Nuchal Translucency (NT) has emerged as a critical marker. This sonographic measurement, typically performed between 11- and 13-weeks 6+ days of gestation, quantifies the fluid-filled space at the back of the fetal neck. An increased NT measurement is a well-established indicator of a higher risk for aneuploidies and other congenital conditions, including heart defects. The Fetal Medicine Foundation has established standardized criteria for this measurement to ensure its reliability and widespread adoption in clinical practice. Methods: We utilized two datasets comprising 2425 ultrasound images from Shenzhen People’s Hospital China and the National Hospital of Obstetrics and Gynecology Vietnam. The methodology employs a two-stage Deep Learning framework: first, a DenseNet121 model assesses image quality to filter non-standard planes; second, a novel DenseNet-based segmentation delineates the NT region for automated measurement. Results: The quality assessment module achieved 94% accuracy in distinguishing standard from non-standard planes. For segmentation, the proposed model achieved a Dice coefficient of 0.897 and an overall accuracy of 98.9%, outperforming the standard U-Net architecture. Clinically, 55.47% of automated measurements deviated by less than 1 mm from expert annotations, and the system demonstrated > 90% sensitivity and specificity for identifying high-risk cases (NT ≥ 2.5 mm). Conclusions: The proposed framework successfully integrates quality assurance with automated measurement, offering a robust decision-support tool to reduce variability and improve screening accuracy in prenatal care.
Top Cited PapersOpen AccessA Deep Learning Framework for Ultrasound Image Quality Assessment and Automated Nuchal Translucency Measurement to Improve First-Trimester Chromosomal Abnormality Screening
Background: First-trimester prenatal screening is a fundamental component of modern obstetric care, offering early insights into fetal health and development. A key focus of this screening is the detection of chromosomal abnormalities, such as Trisomy 21 (Down syndrome), which can have significant implications for pregnancy management and parental counseling. Over the years, various non-invasive methods have been developed, with ultrasound-based assessments becoming a cornerstone of early evaluation. Among these, the measurement of Nuchal Translucency (NT) has emerged as a critical marker. This sonographic measurement, typically performed between 11- and 13-weeks 6+ days of gestation, quantifies the fluid-filled space at the back of the fetal neck. An increased NT measurement is a well-established indicator of a higher risk for aneuploidies and other congenital conditions, including heart defects. The Fetal Medicine Foundation has established standardized criteria for this measurement to ensure its reliability and widespread adoption in clinical practice. Methods: We utilized two datasets comprising 2425 ultrasound images from Shenzhen People’s Hospital China and the National Hospital of Obstetrics and Gynecology Vietnam. The methodology employs a two-stage Deep Learning framework: first, a DenseNet121 model assesses image quality to filter non-standard planes; second, a novel DenseNet-based segmentation delineates the NT region for automated measurement. Results: The quality assessment module achieved 94% accuracy in distinguishing standard from non-standard planes. For segmentation, the proposed model achieved a Dice coefficient of 0.897 and an overall accuracy of 98.9%, outperforming the standard U-Net architecture. Clinically, 55.47% of automated measurements deviated by less than 1 mm from expert annotations, and the system demonstrated > 90% sensitivity and specificity for identifying high-risk cases (NT ≥ 2.5 mm). Conclusions: The proposed framework successfully integrates quality assurance with automated measurement, offering a robust decision-support tool to reduce variability and improve screening accuracy in prenatal care.
- Open Access
Machine Learning and Deep Learning in Lung Cancer Diagnostics: A Systematic Review of Technical Breakthroughs, Clinical Barriers, and Ethical Imperatives
The use of machine learning (ML) and deep learning (DL) in lung cancer detection and classification offers great promise for improving early diagnosis and reducing death rates. Despite major advances in research, there is still a significant gap between successful model development and clinical use. This review identifies the main obstacles preventing ML/DL tools from being adopted in real healthcare settings and suggests practical advice to tackle them. Using PRISMA guidelines, we examined over 100 studies published between 2022 and 2024, focusing on technical accuracy, clinical relevance, and ethical aspects. Most of the reviewed studies rely on computed tomography (CT) imaging, reflecting its dominant role in current lung cancer screening workflows. While many models achieve high performance on public datasets (e.g., >95% sensitivity on LUNA16), they often perform poorly on real clinical data due to issues like domain shift and bias, especially toward underrepresented groups. Promising solutions include federated learning for data privacy, synthetic data to support rare subtypes, and explainable AI to build trust. We also present a checklist to guide the development of clinically applicable tools, emphasizing generalizability, transparency, and workflow integration. The study recommends early collaboration between developers, clinicians, and policymakers to ensure practical adoption. Ultimately, for ML/DL solutions to gain clinical acceptance, they must be designed with healthcare professionals from the beginning.
Open AccessMachine Learning and Deep Learning in Lung Cancer Diagnostics: A Systematic Review of Technical Breakthroughs, Clinical Barriers, and Ethical Imperatives
The use of machine learning (ML) and deep learning (DL) in lung cancer detection and classification offers great promise for improving early diagnosis and reducing death rates. Despite major advances in research, there is still a significant gap between successful model development and clinical use. This review identifies the main obstacles preventing ML/DL tools from being adopted in real healthcare settings and suggests practical advice to tackle them. Using PRISMA guidelines, we examined over 100 studies published between 2022 and 2024, focusing on technical accuracy, clinical relevance, and ethical aspects. Most of the reviewed studies rely on computed tomography (CT) imaging, reflecting its dominant role in current lung cancer screening workflows. While many models achieve high performance on public datasets (e.g., >95% sensitivity on LUNA16), they often perform poorly on real clinical data due to issues like domain shift and bias, especially toward underrepresented groups. Promising solutions include federated learning for data privacy, synthetic data to support rare subtypes, and explainable AI to build trust. We also present a checklist to guide the development of clinically applicable tools, emphasizing generalizability, transparency, and workflow integration. The study recommends early collaboration between developers, clinicians, and policymakers to ensure practical adoption. Ultimately, for ML/DL solutions to gain clinical acceptance, they must be designed with healthcare professionals from the beginning.
- Open Access
- 16 December 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (24) , 3216
A Lightweight Hybrid Deep Learning Model for Tuberculosis Detection from Chest X-Rays
Background/Objectives: Tuberculosis remains a significant global health problem, particularly in resource-limited environments. Its mortality and spread can be considerably decreased by early and precise detection via chest X-ray imaging. This study introduces a novel approach based on hybrid deep learning for Tuberculosis detection from chest X-ray images. Methods: The introduced approach combines GhostNet, a lightweight convolutional neural network tuned for computational efficiency, and MobileViT, a transformer-based model that can capture both local spatial patterns and global contextual dependencies. Through such integration, the model attains a balanced trade-off between classification accuracy and computational efficiency. The architecture employs feature fusion, where spatial features from GhostNet and contextual representations from MobileViT are globally pooled and concatenated, which allows the model to learn discriminative and robust feature representations. Results: The suggested model was assessed on two publicly available chest X-ray datasets and contrasted against several cutting-edge convolutional neural network architectures. Findings showed that the introduced hybrid model surpasses individual baselines, attaining 99.52% accuracy on dataset 1 and 99.17% on dataset 2, while keeping low computational cost (7.73M parameters, 282.11M Floating Point Operations). Conclusions: These outcomes verify the efficacy of feature-level fusion between a convolutional neural network and transformer branches, allowing robust tuberculosis detection with low inference overhead. The model is ideal for clinical deployment and resource-constrained contexts due to its high accuracy and lightweight design.
- 16 December 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (24) , 3216
Open AccessA Lightweight Hybrid Deep Learning Model for Tuberculosis Detection from Chest X-Rays
Background/Objectives: Tuberculosis remains a significant global health problem, particularly in resource-limited environments. Its mortality and spread can be considerably decreased by early and precise detection via chest X-ray imaging. This study introduces a novel approach based on hybrid deep learning for Tuberculosis detection from chest X-ray images. Methods: The introduced approach combines GhostNet, a lightweight convolutional neural network tuned for computational efficiency, and MobileViT, a transformer-based model that can capture both local spatial patterns and global contextual dependencies. Through such integration, the model attains a balanced trade-off between classification accuracy and computational efficiency. The architecture employs feature fusion, where spatial features from GhostNet and contextual representations from MobileViT are globally pooled and concatenated, which allows the model to learn discriminative and robust feature representations. Results: The suggested model was assessed on two publicly available chest X-ray datasets and contrasted against several cutting-edge convolutional neural network architectures. Findings showed that the introduced hybrid model surpasses individual baselines, attaining 99.52% accuracy on dataset 1 and 99.17% on dataset 2, while keeping low computational cost (7.73M parameters, 282.11M Floating Point Operations). Conclusions: These outcomes verify the efficacy of feature-level fusion between a convolutional neural network and transformer branches, allowing robust tuberculosis detection with low inference overhead. The model is ideal for clinical deployment and resource-constrained contexts due to its high accuracy and lightweight design.
- Open Access
Advanced Deep Learning Techniques for Accurate Lung Cancer Detection and Classification
Lung cancer (LC) ranks among the most frequently diagnosed cancers and is one of the most common causes of death for men and women worldwide. Computed Tomography (CT) images are the most preferred diagnosis method because of their low cost and their faster processing times. Many researchers have proposed various ways of identifying lung cancer using CT images. However, such techniques suffer from significant false positives, leading to low accuracy. The fundamental reason results from employing a small and imbalanced dataset. This paper introduces an innovative approach for LC detection and classification from CT images based on the DenseNet201 model. Our approach comprises several advanced methods such as Focal Loss, data augmentation, and regularization to overcome the imbalanced data issue and overfitting challenge. The findings show the appropriateness of the proposal, attaining a promising performance of 98.95% accuracy.
Open AccessAdvanced Deep Learning Techniques for Accurate Lung Cancer Detection and Classification
Lung cancer (LC) ranks among the most frequently diagnosed cancers and is one of the most common causes of death for men and women worldwide. Computed Tomography (CT) images are the most preferred diagnosis method because of their low cost and their faster processing times. Many researchers have proposed various ways of identifying lung cancer using CT images. However, such techniques suffer from significant false positives, leading to low accuracy. The fundamental reason results from employing a small and imbalanced dataset. This paper introduces an innovative approach for LC detection and classification from CT images based on the DenseNet201 model. Our approach comprises several advanced methods such as Focal Loss, data augmentation, and regularization to overcome the imbalanced data issue and overfitting challenge. The findings show the appropriateness of the proposal, attaining a promising performance of 98.95% accuracy.