Miguel-Ángel Luque-Nieto
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Miguel-Ángel Luque-Nieto has a total of 68 co-authors. Here are the top co-authors by number of shared publications:
- Pablo Otero
Institute of Oceanic Engineering Research, University of Malaga, 29010 Málaga, Spain
19 shared publications - Javier Poncela
Department of Communications Engineering, University of Malaga, Malaga, Spain
8 shared publications - Waheeduddin Hyder
Department of Ingeniería de Comunicaciones, University of Malaga, 29010 Malaga, Spain
7 shared publications - José-Miguel Moreno-Roldán
Department of Communication Engineering, University of Málaga, 29071 Málaga, Spain
6 shared publications - Alfonso Ariza
E.T.S. Telecommunications Engineering, University of Malaga, 29010 Malaga, Spain
4 shared publications - Jitander Kumar Pabani
Institute of Oceanic Engineering Research, University of Malaga, Mlaga, Spain
4 shared publications
Latest Publications
- Open Access
EXTREPSO: An Energy Estimator Based on Machine Learning with PSO Optimization for UASNs
Underwater acoustic sensor networks (UASNs) are severely affected by propagation delay, limited bandwidth, and constrained node energy. This work presents EXTREPSO, a machine learning framework that combines an Extra Trees Regressor with Particle Swarm Optimization (PSO) for offline hyperparameter tuning of a hop-level energy-efficiency predictor. The target is derived from Aqua-Sim NG trace fields using the payload size, the simulator Tx_power field, and observed hop delay. Aqua-Sim NG defines PT as transmitted signal power in watts, so the target has dimensional units of bit/J; however, because the observed hop delay includes propagation and protocol-related waiting, it is treated as an operational trace-derived metric rather than a state-resolved modem-energy measurement. In the original 80:20 benchmark, Extra Trees was the strongest baseline among seven classical ML and three DL models. A leakage-resistant sender-grouped outer five-fold evaluation with inner grouped validation was subsequently used to compare baseline Extra Trees, PSO-tuned Extra Trees, and an interpretable inverse-power baseline on untouched test folds. PSO tuning produced a modest average improvement over baseline Extra Trees (mean RMSE 2107.31 versus 2230.58), but the improvement was not consistent across folds; the inverse-power baseline achieved the lowest mean RMSE (1765.88). A separate five-run equal-budget PSO/GWO sensitivity check also found no meaningful superiority of PSO over GWO. These results support the trace-driven prediction and hyperparameter-tuning workflow while limiting claims of universal optimizer benefit or direct network-energy improvement.
Open AccessEXTREPSO: An Energy Estimator Based on Machine Learning with PSO Optimization for UASNs
Underwater acoustic sensor networks (UASNs) are severely affected by propagation delay, limited bandwidth, and constrained node energy. This work presents EXTREPSO, a machine learning framework that combines an Extra Trees Regressor with Particle Swarm Optimization (PSO) for offline hyperparameter tuning of a hop-level energy-efficiency predictor. The target is derived from Aqua-Sim NG trace fields using the payload size, the simulator Tx_power field, and observed hop delay. Aqua-Sim NG defines PT as transmitted signal power in watts, so the target has dimensional units of bit/J; however, because the observed hop delay includes propagation and protocol-related waiting, it is treated as an operational trace-derived metric rather than a state-resolved modem-energy measurement. In the original 80:20 benchmark, Extra Trees was the strongest baseline among seven classical ML and three DL models. A leakage-resistant sender-grouped outer five-fold evaluation with inner grouped validation was subsequently used to compare baseline Extra Trees, PSO-tuned Extra Trees, and an interpretable inverse-power baseline on untouched test folds. PSO tuning produced a modest average improvement over baseline Extra Trees (mean RMSE 2107.31 versus 2230.58), but the improvement was not consistent across folds; the inverse-power baseline achieved the lowest mean RMSE (1765.88). A separate five-run equal-budget PSO/GWO sensitivity check also found no meaningful superiority of PSO over GWO. These results support the trace-driven prediction and hyperparameter-tuning workflow while limiting claims of universal optimizer benefit or direct network-energy improvement.
- Open Access
- 22 October 2026
- Published by MDPI AG in Electronics
- Vol. 15 (21) , 4901
Hybrid Ant-Swarm Coverage and Routing Optimization for Wireless Sensor Networks
In Wireless Sensor Networks (WSNs), effective network coverage, energy-aware routing and reliable data delivery are major challenges. This paper addresses an intelligent optimized WSN using a two-phase hybrid network, thereby tackling these challenges. In the first phase, a proposed Improved Particle Swarm Optimization (IPSO) algorithm locates the collector node in the best position to achieve full coverage. In the second phase, the proposed Modified Elitist Ant Colony Optimization (MEACO) is used to determine the optimized multi-hop routing paths in terms of energy and delay. To evaluate the strength of the proposed framework, a patient network scenario is formulated where patients and collectors are placed to check baseline performance. In addition, a priority-based QoS Scheduling mechanism is introduced for data transfer, in which priority patients are monitored with the least end-to-end delay, while maintaining network lifetime. Reliable delivery of physiological information is challenged by packet loss and unstable wireless links. The proposed framework improves reliability through link quality evaluation, bounded retransmissions, and route recovery. The simulation results show that the proposed IPSO-MEACO approach improves network coverage by 13.2%, minimizes energy utilization by 14.7%, and drastically reduces end-to-end delay by 58.2% compared with classical algorithms. The results also affirm that the algorithm is suitable for full coverage with minimum end-to-end delay and energy consumption.
- 22 October 2026
- Published by MDPI AG in Electronics
- Vol. 15 (21) , 4901
Open AccessHybrid Ant-Swarm Coverage and Routing Optimization for Wireless Sensor Networks
In Wireless Sensor Networks (WSNs), effective network coverage, energy-aware routing and reliable data delivery are major challenges. This paper addresses an intelligent optimized WSN using a two-phase hybrid network, thereby tackling these challenges. In the first phase, a proposed Improved Particle Swarm Optimization (IPSO) algorithm locates the collector node in the best position to achieve full coverage. In the second phase, the proposed Modified Elitist Ant Colony Optimization (MEACO) is used to determine the optimized multi-hop routing paths in terms of energy and delay. To evaluate the strength of the proposed framework, a patient network scenario is formulated where patients and collectors are placed to check baseline performance. In addition, a priority-based QoS Scheduling mechanism is introduced for data transfer, in which priority patients are monitored with the least end-to-end delay, while maintaining network lifetime. Reliable delivery of physiological information is challenged by packet loss and unstable wireless links. The proposed framework improves reliability through link quality evaluation, bounded retransmissions, and route recovery. The simulation results show that the proposed IPSO-MEACO approach improves network coverage by 13.2%, minimizes energy utilization by 14.7%, and drastically reduces end-to-end delay by 58.2% compared with classical algorithms. The results also affirm that the algorithm is suitable for full coverage with minimum end-to-end delay and energy consumption.
- Open Access
- 9 June 2026
- Published by MDPI AG in Technologies
- Vol. 14 (6) , 344
Experimental Evaluation of a VANET Prototype Using ESP-NOW for Collision Avoidance: Latency, Packet Loss, and Statistical Performance in Urban Environments
Vehicle ad hoc networks (VANETs) can help prevent traffic accidents through wireless communication; however, most studies are based on simulations or static evaluations. This research paper presents the design, implementation, and experimental evaluation of a prototype early-warning system for vehicle proximity based on VANETs using ESP-NOW. The prototype utilizes five ESP32-CAM nodes equipped with MaxSonar sensors installed in vehicles and an RSU unit with a Raspberry Pi for vehicle-to-infrastructure (V2I) communication. Field tests were conducted in Quito, Ecuador, at speeds ranging from 10 to 70 km/h, measuring latency, packet loss, and received signal strength (RSSI). The results show average latencies of 9.9 ms at 10 km/h and 114.5 ms at 70 km/h, with packet loss rates of 2% and 60%, respectively. Statistical analysis reveals 95% confidence intervals for latency ranging from ±0.98 ms to ±6.90 ms, while obstacles introduce marginal attenuation (p = 0.051) with significant dispersion (σ = 5.85 dB). The Doppler shift is negligible (155.6 Hz), but the channel coherence time (2.7 ms) explains the observed degradation. Models were obtained that relate speed to latency (R2 = 0.994) and packet loss (R2 = 0.991). The prototype is viable for early collision warning at urban speeds (up to 60 km/h), outperforming human reaction time (1.5 s).
- 9 June 2026
- Published by MDPI AG in Technologies
- Vol. 14 (6) , 344
Open AccessExperimental Evaluation of a VANET Prototype Using ESP-NOW for Collision Avoidance: Latency, Packet Loss, and Statistical Performance in Urban Environments
Vehicle ad hoc networks (VANETs) can help prevent traffic accidents through wireless communication; however, most studies are based on simulations or static evaluations. This research paper presents the design, implementation, and experimental evaluation of a prototype early-warning system for vehicle proximity based on VANETs using ESP-NOW. The prototype utilizes five ESP32-CAM nodes equipped with MaxSonar sensors installed in vehicles and an RSU unit with a Raspberry Pi for vehicle-to-infrastructure (V2I) communication. Field tests were conducted in Quito, Ecuador, at speeds ranging from 10 to 70 km/h, measuring latency, packet loss, and received signal strength (RSSI). The results show average latencies of 9.9 ms at 10 km/h and 114.5 ms at 70 km/h, with packet loss rates of 2% and 60%, respectively. Statistical analysis reveals 95% confidence intervals for latency ranging from ±0.98 ms to ±6.90 ms, while obstacles introduce marginal attenuation (p = 0.051) with significant dispersion (σ = 5.85 dB). The Doppler shift is negligible (155.6 Hz), but the channel coherence time (2.7 ms) explains the observed degradation. Models were obtained that relate speed to latency (R2 = 0.994) and packet loss (R2 = 0.991). The prototype is viable for early collision warning at urban speeds (up to 60 km/h), outperforming human reaction time (1.5 s).
- Open Access
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. PP (15) , 1
U-Tangle: A Lightweight DAG-Based Architecture for Secure and Scalable Underwater Acoustic IoT Networks
The Internet of Underwater Things (IoUT) requires robust security architectures capable of operating under severe constraints: high latency, low bandwidth, and limited energy. Traditional blockchain-based solutions are often infeasible due to their heavy synchronization and mining requirements. In this work, we propose U-Tangle, a lightweight, Directed Acyclic Graph (DAG)-based security architecture inspired by Internet of Things Applications (IOTA) but optimized for Underwater Acoustic Sensor Networks (UASNs). Unlike existing blockchain schemes, U-Tangle eliminates mining and linear block generation, utilizing a collaborative validation mechanism where each transaction confirms previous ones. We integrate lightweight cryptography (ED25519, X25519, Ascon) to ensure confidentiality and integrity with minimal overhead. Extensive Python-based simulations show that U-Tangle achieves sub-second end-to-end authentication latency (316.04 msec with 95% CI [310.56, 321.52] over 30 runs in Scenario S1 with N=200) while reducing energy consumption by avoiding continuous ledger synchronization. Comparative insights against representative block-based approaches are discussed based on published system-level characteristics. Our security analysis confirms resilience against replay, Sybil, and Man-in-the-Middle attacks without relying on online centralized authorities. This work provides a scalable, energy-efficient alternative for securing critical underwater infrastructure.
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. PP (15) , 1
Open AccessU-Tangle: A Lightweight DAG-Based Architecture for Secure and Scalable Underwater Acoustic IoT Networks
The Internet of Underwater Things (IoUT) requires robust security architectures capable of operating under severe constraints: high latency, low bandwidth, and limited energy. Traditional blockchain-based solutions are often infeasible due to their heavy synchronization and mining requirements. In this work, we propose U-Tangle, a lightweight, Directed Acyclic Graph (DAG)-based security architecture inspired by Internet of Things Applications (IOTA) but optimized for Underwater Acoustic Sensor Networks (UASNs). Unlike existing blockchain schemes, U-Tangle eliminates mining and linear block generation, utilizing a collaborative validation mechanism where each transaction confirms previous ones. We integrate lightweight cryptography (ED25519, X25519, Ascon) to ensure confidentiality and integrity with minimal overhead. Extensive Python-based simulations show that U-Tangle achieves sub-second end-to-end authentication latency (316.04 msec with 95% CI [310.56, 321.52] over 30 runs in Scenario S1 with N=200) while reducing energy consumption by avoiding continuous ledger synchronization. Comparative insights against representative block-based approaches are discussed based on published system-level characteristics. Our security analysis confirms resilience against replay, Sybil, and Man-in-the-Middle attacks without relying on online centralized authorities. This work provides a scalable, energy-efficient alternative for securing critical underwater infrastructure.
- Top Cited PapersOpen Access
- 31 January 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Identification of multiple ocular diseases using a hybrid quantum convolutional neural network with fundus images
Ocular diseases remain a major cause of vision impairment globally, making early and accurate diagnosis essential. This study presents a novel diagnostic model for identifying seven common ocular conditions age-related macular degeneration, glaucoma, hypertension, diabetic retinopathy, myopia, cataracts, and other pathologies using clinical fundus images. To improve image quality, Anisotropic Diffusion Filtering and Wavelet Transform are applied for hue and contrast enhancement. Data imbalance is addressed through targeted augmentation techniques. The core of the model is a hybrid Quantum Convolutional Neural Network (QCNN), which integrates quantum convolutional pooling into a classical CNN architecture to boost feature extraction and classification. Evaluated on the OIA-ODIR dataset, the proposed model outperformed benchmarks such as Fundus-DeepNet, Inception-v4, VGG16 with SGD, and ResNet-101. It achieved 94% classification accuracy, along with substantial gains in precision, recall, and F1-score. These results confirm the model’s effectiveness and its potential for supporting early, multi-disease ocular diagnosis in clinical settings.
- 31 January 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Top Cited PapersOpen AccessIdentification of multiple ocular diseases using a hybrid quantum convolutional neural network with fundus images
Ocular diseases remain a major cause of vision impairment globally, making early and accurate diagnosis essential. This study presents a novel diagnostic model for identifying seven common ocular conditions age-related macular degeneration, glaucoma, hypertension, diabetic retinopathy, myopia, cataracts, and other pathologies using clinical fundus images. To improve image quality, Anisotropic Diffusion Filtering and Wavelet Transform are applied for hue and contrast enhancement. Data imbalance is addressed through targeted augmentation techniques. The core of the model is a hybrid Quantum Convolutional Neural Network (QCNN), which integrates quantum convolutional pooling into a classical CNN architecture to boost feature extraction and classification. Evaluated on the OIA-ODIR dataset, the proposed model outperformed benchmarks such as Fundus-DeepNet, Inception-v4, VGG16 with SGD, and ResNet-101. It achieved 94% classification accuracy, along with substantial gains in precision, recall, and F1-score. These results confirm the model’s effectiveness and its potential for supporting early, multi-disease ocular diagnosis in clinical settings.
- Open Access
- 20 November 2025
- Published by MDPI AG in Electronics
- Vol. 14 (22) , 4543
Range Extension for Underwater Communication via Magnetic Induction Using Parametric Analysis of MI Coils in IoUT Networks
This paper discusses the method for extending the range of Magnetic Induction (MI) and its application in underwater networks for the Internet of Underwater Things (IoUT). In underwater communication, this technology would provide a wider frequency band than acoustic systems, shorter propagation delay, and increased conductivity, with the added benefit of underwater wireless power transfer. As a use case, we consider a system that allows energy to be transferred from one circuit to another without cables, as in an aerial environment. In this work, transmit and receive coils for underwater environments are designed and analyzed using ANSYS Maxwell v16.0 software. The results show an improvement in terms of underwater magnetic field propagation. We have conducted underwater experiments by applying a frequency range up to 100 kHz and 12 Volts with varied current, achieving a distance up to 80% greater than in air, as determined by parametric analysis. With an improved bit error rate, a delay of less than 2 microseconds, a packet delivery ratio near 100%, and a packet loss ratio less than 10%, the results show an improvement in magnetic field propagation underwater. This demonstrates that it is possible to conduct future research into other underwater applications by implementing MI for underwater communication.
- 20 November 2025
- Published by MDPI AG in Electronics
- Vol. 14 (22) , 4543
Open AccessRange Extension for Underwater Communication via Magnetic Induction Using Parametric Analysis of MI Coils in IoUT Networks
This paper discusses the method for extending the range of Magnetic Induction (MI) and its application in underwater networks for the Internet of Underwater Things (IoUT). In underwater communication, this technology would provide a wider frequency band than acoustic systems, shorter propagation delay, and increased conductivity, with the added benefit of underwater wireless power transfer. As a use case, we consider a system that allows energy to be transferred from one circuit to another without cables, as in an aerial environment. In this work, transmit and receive coils for underwater environments are designed and analyzed using ANSYS Maxwell v16.0 software. The results show an improvement in terms of underwater magnetic field propagation. We have conducted underwater experiments by applying a frequency range up to 100 kHz and 12 Volts with varied current, achieving a distance up to 80% greater than in air, as determined by parametric analysis. With an improved bit error rate, a delay of less than 2 microseconds, a packet delivery ratio near 100%, and a packet loss ratio less than 10%, the results show an improvement in magnetic field propagation underwater. This demonstrates that it is possible to conduct future research into other underwater applications by implementing MI for underwater communication.
- Open Access
- 1 September 2025
- Published by Universidade da Coruna in Jornadas de Automática
- No. 46
Robótica Colaborativa Marina: Explorando Escenarios con NauSim y YOLO
Este artículo presenta los resultados de las pruebas con NauSim, un simulador de código abierto para drones acuáticos, junto con el sistema de visión por computador YOLO en el contexto de la robótica colaborativa para el control y posicionamiento de vehículos de superficie no tripulados (USV). Se desarrollan algoritmos de control para tareas como formaciones, seguimiento de rutas y evasión de obstáculos, que se prueban en simulación y en escenarios reales. Los resultados muestran un notable rendimiento general del sistema y buena correspondencia entre la simulación y la realidad, apoyando la viabilidad del sistema propuesto.
- 1 September 2025
- Published by Universidade da Coruna in Jornadas de Automática
- No. 46
Open AccessRobótica Colaborativa Marina: Explorando Escenarios con NauSim y YOLO
Este artículo presenta los resultados de las pruebas con NauSim, un simulador de código abierto para drones acuáticos, junto con el sistema de visión por computador YOLO en el contexto de la robótica colaborativa para el control y posicionamiento de vehículos de superficie no tripulados (USV). Se desarrollan algoritmos de control para tareas como formaciones, seguimiento de rutas y evasión de obstáculos, que se prueban en simulación y en escenarios reales. Los resultados muestran un notable rendimiento general del sistema y buena correspondencia entre la simulación y la realidad, apoyando la viabilidad del sistema propuesto.
- Open Access
- 14 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 38655-38670
Smart Packet Delivery in Mobile Underwater Sensors Networks (M-CTSP)
This research is focused on leveraging the Travelling Salesman Problem (TSP) to establish the optimal route of getting from a source node to a destination node in an underwater dynamic location. In this context, all nodes are mobile and slightly adjust their positions due to water currents. Clustering is introduced (CTSP) to identify the best possible route, where cluster heads are chosen using the known criterium of center of mass of the cluster as well as residual energy. Packets travel from source to destination (nodes) in a multiloop process, visiting cluster heads nodes, with minimum energy consumption. The intention behind using TSP in this approach is thus to efficiently handle the dynamic movement of underwater nodes by conserving energy and ensuring delivery of packets can be relied upon. Additionally, to the movement-related challenges, this method also enhances network lifetime by providing a systematic and energy-efficient way of transmitting information through cluster heads. Simulations using the novel M-CTSP into the routing algorithm yields to promising results the M-CTSP protocol enhances routing execution with the use of dynamic clustering and traveling salesman optimization. This results in a 40% rise in energy use reduction, 15% improvement in packet delivery with a 30% increase in network lifetime when compared to normal clustering-based protocols: LEACH and EECBP-FOA. Experimental results prove that M-CTSP contributes extensively in terms of node mobility packet delivery and energy conservation in dynamic UWSN environments.
- 14 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 38655-38670
Open AccessSmart Packet Delivery in Mobile Underwater Sensors Networks (M-CTSP)
This research is focused on leveraging the Travelling Salesman Problem (TSP) to establish the optimal route of getting from a source node to a destination node in an underwater dynamic location. In this context, all nodes are mobile and slightly adjust their positions due to water currents. Clustering is introduced (CTSP) to identify the best possible route, where cluster heads are chosen using the known criterium of center of mass of the cluster as well as residual energy. Packets travel from source to destination (nodes) in a multiloop process, visiting cluster heads nodes, with minimum energy consumption. The intention behind using TSP in this approach is thus to efficiently handle the dynamic movement of underwater nodes by conserving energy and ensuring delivery of packets can be relied upon. Additionally, to the movement-related challenges, this method also enhances network lifetime by providing a systematic and energy-efficient way of transmitting information through cluster heads. Simulations using the novel M-CTSP into the routing algorithm yields to promising results the M-CTSP protocol enhances routing execution with the use of dynamic clustering and traveling salesman optimization. This results in a 40% rise in energy use reduction, 15% improvement in packet delivery with a 30% increase in network lifetime when compared to normal clustering-based protocols: LEACH and EECBP-FOA. Experimental results prove that M-CTSP contributes extensively in terms of node mobility packet delivery and energy conservation in dynamic UWSN environments.
- Open Access
- 1 January 2025
- Published by Walter de Gruyter GmbH in Journal of Intelligent Systems
- Vol. 34 (1)
A fast and fully automated system for segmenting retinal blood vessels in fundus images
The problem of segmenting retinal blood vessels in fundus images arises from the challenges of accurately detecting and delineating vessels due to their complex structures, varying sizes, and overlapping features. Manual segmentation is time-consuming and prone to human error, leading to inconsistent results. Additionally, existing automated methods often struggle with low-quality images or variations in illumination, hindering their effectiveness. Therefore, there is a pressing need for an efficient and accurate automated system to improve segmentation outcomes for better diagnosis of retinal diseases. This study proposes a fully automated model for blood vessel segmentation in retinal fundus images, addressing key challenges such as poor image quality, weak vessel detection, and inhomogeneity in contrast. Macular degeneration and diabetic retinopathy are major causes of vision impairment, making accurate retinal analysis crucial. The proposed model enhances image quality through a novel pre-processing pipeline that includes logarithmic contrast enhancement, noise reduction using an improved complex wavelet transform with shrinkage, and anisotropic diffusion filtering for edge enhancement. The segmentation method combines morphological operations with an optimized Canny edge detector, effectively identifying and segmenting blood vessels. This approach aims to improve the accuracy and efficiency of retinal image analysis, overcoming the limitations of manual segmentation and complex vascular structures. The results obtained from the DRIVE dataset achieved high values for accuracy (Acc, 99%), sensitivity (Sen, 95.83%), specificity (Spe, 98.62%), positive predictive value (PPV, 91.34%), and negative predictive value (NPV, 94%). In addition, the results obtained using the high-resolution fundus dataset were equally satisfactory, achieving an Acc., Sen., Spe., PPV, and NPV of 99.11, 97.97, 98.97, 97.98, and 100%, respectively. These results outperform the gold standard and state-of-the-art schemes to date. The proposed approach increases the performance and reliability of the process of vessel detection in fundus images.
- 1 January 2025
- Published by Walter de Gruyter GmbH in Journal of Intelligent Systems
- Vol. 34 (1)
Open AccessA fast and fully automated system for segmenting retinal blood vessels in fundus images
The problem of segmenting retinal blood vessels in fundus images arises from the challenges of accurately detecting and delineating vessels due to their complex structures, varying sizes, and overlapping features. Manual segmentation is time-consuming and prone to human error, leading to inconsistent results. Additionally, existing automated methods often struggle with low-quality images or variations in illumination, hindering their effectiveness. Therefore, there is a pressing need for an efficient and accurate automated system to improve segmentation outcomes for better diagnosis of retinal diseases. This study proposes a fully automated model for blood vessel segmentation in retinal fundus images, addressing key challenges such as poor image quality, weak vessel detection, and inhomogeneity in contrast. Macular degeneration and diabetic retinopathy are major causes of vision impairment, making accurate retinal analysis crucial. The proposed model enhances image quality through a novel pre-processing pipeline that includes logarithmic contrast enhancement, noise reduction using an improved complex wavelet transform with shrinkage, and anisotropic diffusion filtering for edge enhancement. The segmentation method combines morphological operations with an optimized Canny edge detector, effectively identifying and segmenting blood vessels. This approach aims to improve the accuracy and efficiency of retinal image analysis, overcoming the limitations of manual segmentation and complex vascular structures. The results obtained from the DRIVE dataset achieved high values for accuracy (Acc, 99%), sensitivity (Sen, 95.83%), specificity (Spe, 98.62%), positive predictive value (PPV, 91.34%), and negative predictive value (NPV, 94%). In addition, the results obtained using the high-resolution fundus dataset were equally satisfactory, achieving an Acc., Sen., Spe., PPV, and NPV of 99.11, 97.97, 98.97, 97.98, and 100%, respectively. These results outperform the gold standard and state-of-the-art schemes to date. The proposed approach increases the performance and reliability of the process of vessel detection in fundus images.
- Open Access
- 19 December 2024
- Published by MDPI AG in Journal of Imaging
- Vol. 10 (12) , 328
X-Ray Image-Based Real-Time COVID-19 Diagnosis Using Deep Neural Networks (CXR-DNNs)
On 11 February 2020, the prevalent outbreak of COVID-19, a coronavirus illness, was declared a global pandemic. Since then, nearly seven million people have died and over 765 million confirmed cases of COVID-19 have been reported. The goal of this study is to develop a diagnostic tool for detecting COVID-19 infections more efficiently. Currently, the most widely used method is Reverse Transcription Polymerase Chain Reaction (RT-PCR), a clinical technique for infection identification. However, RT-PCR is expensive, has limited sensitivity, and requires specialized medical expertise. One of the major challenges in the rapid diagnosis of COVID-19 is the need for reliable imaging, particularly X-ray imaging. This work takes advantage of artificial intelligence (AI) techniques to enhance diagnostic accuracy by automating the detection of COVID-19 infections from chest X-ray (CXR) images. We obtained and analyzed CXR images from the Kaggle public database (4035 images in total), including cases of COVID-19, viral pneumonia, pulmonary opacity, and healthy controls. By integrating advanced techniques with transfer learning from pre-trained convolutional neural networks (CNNs), specifically InceptionV3, ResNet50, and Xception, we achieved an accuracy of 95%, significantly higher than the 85.5% achieved with ResNet50 alone. Additionally, our proposed method, CXR-DNNs, can accurately distinguish between three different types of chest X-ray images for the first time. This computer-assisted diagnostic tool has the potential to significantly enhance the speed and accuracy of COVID-19 diagnoses.
- 19 December 2024
- Published by MDPI AG in Journal of Imaging
- Vol. 10 (12) , 328
Open AccessX-Ray Image-Based Real-Time COVID-19 Diagnosis Using Deep Neural Networks (CXR-DNNs)
On 11 February 2020, the prevalent outbreak of COVID-19, a coronavirus illness, was declared a global pandemic. Since then, nearly seven million people have died and over 765 million confirmed cases of COVID-19 have been reported. The goal of this study is to develop a diagnostic tool for detecting COVID-19 infections more efficiently. Currently, the most widely used method is Reverse Transcription Polymerase Chain Reaction (RT-PCR), a clinical technique for infection identification. However, RT-PCR is expensive, has limited sensitivity, and requires specialized medical expertise. One of the major challenges in the rapid diagnosis of COVID-19 is the need for reliable imaging, particularly X-ray imaging. This work takes advantage of artificial intelligence (AI) techniques to enhance diagnostic accuracy by automating the detection of COVID-19 infections from chest X-ray (CXR) images. We obtained and analyzed CXR images from the Kaggle public database (4035 images in total), including cases of COVID-19, viral pneumonia, pulmonary opacity, and healthy controls. By integrating advanced techniques with transfer learning from pre-trained convolutional neural networks (CNNs), specifically InceptionV3, ResNet50, and Xception, we achieved an accuracy of 95%, significantly higher than the 85.5% achieved with ResNet50 alone. Additionally, our proposed method, CXR-DNNs, can accurately distinguish between three different types of chest X-ray images for the first time. This computer-assisted diagnostic tool has the potential to significantly enhance the speed and accuracy of COVID-19 diagnoses.