S. Sabena
Research Subjects
S. Sabena has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Sai Ramesh Lakshmanan
Department of Information Science and Technology, CEG campus, Anna University, Chennai, India
4 shared publications - A. K. Jaithunbi
RMD Engineering College, Chennai, India
2 shared publications - Kamalanathan Selvakumar
Department of Computer Applications, NIT Trichy, Trichy, India
2 shared publications - C. Aruna Vinodhini
Anna University, Chennai
1 shared publication - C. Sureshkumar
Department of Information Technology, Vel Tech Rangarajan Dr. Sangunthala R&D Institute of Science and Technology, Avadi, Chennai, India
1 shared publication - G. Jaspher Willsie Kathrine
Karunya Institue of Technology and Sciences, India
1 shared publication
Latest Publications
- 15 January 2025
- Published by Springer Nature in Wireless Personal Communications
- Vol. 140 (1-2) , 467-482
Intelligent Intrusion Detection and Prevention System for IoT Using Game Theoretic Approach
- 15 January 2025
- Published by Springer Nature in Wireless Personal Communications
- Vol. 140 (1-2) , 467-482
Intelligent Intrusion Detection and Prevention System for IoT Using Game Theoretic Approach
- Open Access
- 3 October 2023
- Published by Springer Nature in Research Square
DREXUNET: A novel Deep CNN Algorithm for drusen and exudates classification on fundus images
Glaucoma is a deadly illness that affects individuals globally and can lead to loss of vision, particularly in those over 40 to 60 years old, following cataracts. Timely treatment can be facilitated through early detection of diabetic retinopathy, which requires significant investment in screening programs, particularly automated ones. Identifying exudates is crucial for prompt diagnosis of diabetic retinopathy. Deep neural networks, a promising machine learning approach, have demonstrated remarkable success in various computer vision issues. This paper presents a novel convolutional neural network, DREXUNET, to classify drusen, exudate and normal regions in eye fundus images. DREXUNET focuses on reducing the overfitting based on the input data set size, reducing the Batch loss and validation loss, and reducing the number of exponentially growing computation and time. The proposed model has been tested and validated in two different datasets of different sizes. The proposed model has been compared with pretrained models such as Alexnet, Googlenet, Densenet, VGG16 and Squeezenet. The outcome demonstrates that the drexunet model proposed surpasses pretrained models with a 97.50\% precision, 98.22\% recall, 98.54\% F1 score, 90.79\% sensitivity, 91.21\% specificity, and a 0.0779 loss.
- 3 October 2023
- Published by Springer Nature in Research Square
Open AccessDREXUNET: A novel Deep CNN Algorithm for drusen and exudates classification on fundus images
Glaucoma is a deadly illness that affects individuals globally and can lead to loss of vision, particularly in those over 40 to 60 years old, following cataracts. Timely treatment can be facilitated through early detection of diabetic retinopathy, which requires significant investment in screening programs, particularly automated ones. Identifying exudates is crucial for prompt diagnosis of diabetic retinopathy. Deep neural networks, a promising machine learning approach, have demonstrated remarkable success in various computer vision issues. This paper presents a novel convolutional neural network, DREXUNET, to classify drusen, exudate and normal regions in eye fundus images. DREXUNET focuses on reducing the overfitting based on the input data set size, reducing the Batch loss and validation loss, and reducing the number of exponentially growing computation and time. The proposed model has been tested and validated in two different datasets of different sizes. The proposed model has been compared with pretrained models such as Alexnet, Googlenet, Densenet, VGG16 and Squeezenet. The outcome demonstrates that the drexunet model proposed surpasses pretrained models with a 97.50\% precision, 98.22\% recall, 98.54\% F1 score, 90.79\% sensitivity, 91.21\% specificity, and a 0.0779 loss.
- 28 August 2023
- Published by Springer Nature
- p. 81-93
Collaborative CNN with Multiple Tuning for Automated Coral Reef Classification
- 28 August 2023
- Published by Springer Nature
- p. 81-93
Collaborative CNN with Multiple Tuning for Automated Coral Reef Classification
- 28 October 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 129 (1) , 271-284
Preservation of Data Integrity in Public Cloud Using Enhanced Vigenere Cipher Based Obfuscation
- 28 October 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 129 (1) , 271-284
Preservation of Data Integrity in Public Cloud Using Enhanced Vigenere Cipher Based Obfuscation
- 7 February 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 124 (4) , 3591-3602
Energy Efficient Service Selection from IoT Based on QoS Using HMM with KNN and XGBoost
- 7 February 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 124 (4) , 3591-3602
Energy Efficient Service Selection from IoT Based on QoS Using HMM with KNN and XGBoost
- 20 August 2021
- Published by Springer Nature in Wireless Personal Communications
- Vol. 121 (4) , 3281-3295
Trust Evaluation of Public Cloud Service Providers Using Genetic Algorithm with Intelligent Rules
- 20 August 2021
- Published by Springer Nature in Wireless Personal Communications
- Vol. 121 (4) , 3281-3295
Trust Evaluation of Public Cloud Service Providers Using Genetic Algorithm with Intelligent Rules
- 27 January 2020
- Published by Springer Nature in Wireless Personal Communications
- Vol. 112 (3) , 1517-1536
Fuzzy-Based Secure Authentication and Clustering Algorithm for Improving the Energy Efficiency in Wireless Sensor Networks
- 27 January 2020
- Published by Springer Nature in Wireless Personal Communications
- Vol. 112 (3) , 1517-1536
Fuzzy-Based Secure Authentication and Clustering Algorithm for Improving the Energy Efficiency in Wireless Sensor Networks