Deepak Dembla
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Deepak Dembla has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
Parul Tyagi
3 shared publications- Minakshi Kumar
JECRC University
2 shared publications - Vishal Goyal
Punjabi University
2 shared publications Akanksha Sharma
Department of MCA & IT JECRC University Jaipur, India
1 shared publication- Ashima Tiwari
JECRC University
1 shared publication Himanshu Yadav
JECRC University, Jaipur, Rajasthan India
1 shared publication
Latest Publications
- Open Access
- 1 January 2026
- Published by AOSIS in African Vision and Eye Health
- Vol. 85 (1) , a1192
Erratum: Artificial intelligence-driven glaucoma screening in ophthalmology: The GlaucoNet deep learning framework
- 1 January 2026
- Published by AOSIS in African Vision and Eye Health
- Vol. 85 (1) , a1192
Open AccessErratum: Artificial intelligence-driven glaucoma screening in ophthalmology: The GlaucoNet deep learning framework
- Open Access
- 1 January 2026
- Published by AOSIS in African Vision and Eye Health
- Vol. 85 (1) , 12
Artificial intelligence-driven glaucoma screening in ophthalmology: The GlaucoNet deep learning framework
Background: Glaucoma is a significant cause of permanent blindness in the world, and it is usually unnoticed at its initial stages. Deep learning allows scaling flexible quality screening of retinal fundus images to detect and implement effective intervention at the appropriate time. Aim: This study developed and evaluated GlaucoNet, an ensemble of MobileNetV2, InceptionV3, and ResNet50, for accurate glaucoma detection from retinal fundus images. Setting: Publicly available data sets used include Online Retinal fundus Image database for Glaucoma Analysis (325 normal and 325 glaucoma images), and Automatic Classification of Retinal Images for Medical Assessment (309 normal and 396 glaucoma images). Methods: To address the imbalance in classes, the process of data augmentation such as flipping, rotation, scaling and brightness was used. The data were separated into 5-fold cross-validation with patient-wise data separation training (70%), validation (10%) and testing (20%) to avoid data leakage. The early stopping that was done on validation loss, dropout (0.3) and L2 regularisation minimised overfitting in the branches of the convolutional neural network (CNN). Results: The accuracy of GlaucoNet was 97.0, sensitivity 97.32, and specificity 96.64, with the following 95% confidence intervals of 96.19–97.8, 96.39–98.68 and 95.59–96.63 respectively. DeLong testing showed that the area under the curve of GlaucoNet (0.9698) was significantly greater compared to MobileNetV2, InceptionV3 and ResNet50 (p < 0.05) and proved the statistical dominance of the ensemble method. Conclusion: GlaucoNet achieved high accuracy in glaucoma detection, supporting its clinical screening potential. Contribution: By leveraging the strengths of multiple CNN architectures, GlaucoNet offers a promising tool for enhancing early diagnosis and facilitating timely clinical intervention in ophthalmology.
- 1 January 2026
- Published by AOSIS in African Vision and Eye Health
- Vol. 85 (1) , 12
Open AccessArtificial intelligence-driven glaucoma screening in ophthalmology: The GlaucoNet deep learning framework
Background: Glaucoma is a significant cause of permanent blindness in the world, and it is usually unnoticed at its initial stages. Deep learning allows scaling flexible quality screening of retinal fundus images to detect and implement effective intervention at the appropriate time. Aim: This study developed and evaluated GlaucoNet, an ensemble of MobileNetV2, InceptionV3, and ResNet50, for accurate glaucoma detection from retinal fundus images. Setting: Publicly available data sets used include Online Retinal fundus Image database for Glaucoma Analysis (325 normal and 325 glaucoma images), and Automatic Classification of Retinal Images for Medical Assessment (309 normal and 396 glaucoma images). Methods: To address the imbalance in classes, the process of data augmentation such as flipping, rotation, scaling and brightness was used. The data were separated into 5-fold cross-validation with patient-wise data separation training (70%), validation (10%) and testing (20%) to avoid data leakage. The early stopping that was done on validation loss, dropout (0.3) and L2 regularisation minimised overfitting in the branches of the convolutional neural network (CNN). Results: The accuracy of GlaucoNet was 97.0, sensitivity 97.32, and specificity 96.64, with the following 95% confidence intervals of 96.19–97.8, 96.39–98.68 and 95.59–96.63 respectively. DeLong testing showed that the area under the curve of GlaucoNet (0.9698) was significantly greater compared to MobileNetV2, InceptionV3 and ResNet50 (p < 0.05) and proved the statistical dominance of the ensemble method. Conclusion: GlaucoNet achieved high accuracy in glaucoma detection, supporting its clinical screening potential. Contribution: By leveraging the strengths of multiple CNN architectures, GlaucoNet offers a promising tool for enhancing early diagnosis and facilitating timely clinical intervention in ophthalmology.
- 31 March 2019
- Published by Springer Nature
- p. 103-112
A Novel Algorithm for Automatic Text Summarization System Using Lexical Chain
- 31 March 2019
- Published by Springer Nature
- p. 103-112
A Novel Algorithm for Automatic Text Summarization System Using Lexical Chain
- 10 May 2018
- Published by Springer Nature in Wireless Personal Communications
- Vol. 102 (1) , 41-60
Advanced Secured Routing Algorithm of Vehicular Ad-Hoc Network
- 10 May 2018
- Published by Springer Nature in Wireless Personal Communications
- Vol. 102 (1) , 41-60
Advanced Secured Routing Algorithm of Vehicular Ad-Hoc Network
- 29 September 2017
- Published by Springer Nature
- p. 601-610
Implementation of Modified K-means Approach for Privacy Preserving in Data Mining
- 29 September 2017
- Published by Springer Nature
- p. 601-610
Implementation of Modified K-means Approach for Privacy Preserving in Data Mining
- 1 September 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 759-765
Implementation of advanced authentication system using opencv by capturing motion images
Authentication means preserving the systems and information from unwanted unauthorized access, usage, disclosure, disruption, integration, inspection, modification or destruction of the information. The studies have shown that the text passwords are more vulnerable to shoulder surfing, brute force, and dictionary attacks. The password with images are hard to guess and will provide a more secure way to access the recourse, with much more security but they are still vulnerable to shoulder surfing attacks. In proposed system the video is used as a password. OpenCv (open source computer vision) is used, which is an open source to capture video and will extract the frames out of the videos and these frames are stored in an encrypted form using AES(advanced encryption standard) and in turn to authenticate the user JAMA(numerical linear algebra library) is used. Java is used for implementation which is platform independent. The proposed method can be used where the password protection is required like in banking, educational systems, social accounts, Gmail, email, desktops and in every field where the information is required to be protected.
- 1 September 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 759-765
Implementation of advanced authentication system using opencv by capturing motion images
Authentication means preserving the systems and information from unwanted unauthorized access, usage, disclosure, disruption, integration, inspection, modification or destruction of the information. The studies have shown that the text passwords are more vulnerable to shoulder surfing, brute force, and dictionary attacks. The password with images are hard to guess and will provide a more secure way to access the recourse, with much more security but they are still vulnerable to shoulder surfing attacks. In proposed system the video is used as a password. OpenCv (open source computer vision) is used, which is an open source to capture video and will extract the frames out of the videos and these frames are stored in an encrypted form using AES(advanced encryption standard) and in turn to authenticate the user JAMA(numerical linear algebra library) is used. Java is used for implementation which is platform independent. The proposed method can be used where the password protection is required like in banking, educational systems, social accounts, Gmail, email, desktops and in every field where the information is required to be protected.
- Open Access
- 1 July 2017
- Published by Elsevier in Egyptian Informatics Journal
- Vol. 18 (2) , 133-139
Performance analysis and implementation of proposed mechanism for detection and prevention of security attacks in routing protocols of vehicular ad-hoc network (VANET)
- 1 July 2017
- Published by Elsevier in Egyptian Informatics Journal
- Vol. 18 (2) , 133-139
Open AccessPerformance analysis and implementation of proposed mechanism for detection and prevention of security attacks in routing protocols of vehicular ad-hoc network (VANET)
- Open Access
- 1 January 2016
- Published by Association for Computing Machinery (ACM)
- p. 1-5
A Novel Technique of implementing Bidirectional Point Coordination Function for Voice Traffic in WLAN
WLANs have substituted wired networks and have formed an uprising in the area of communication. The Point Coordination Function (PCF) of the IEEE 802.11 protocol offersto sustain to real time traffic for Wireless Local Area Network (WLANs). However, PCF being a central polling protocol, a little bandwidth is exhausted on null packets and polling operating cost. To supplyimproved channel operation and to decreaseconsumption of bandwidth, an enhanced bidirectional broadcast of setextent known as Bi-Directional Point coordination Function (BD-PCF) is included into PCF. Based on this policy, wireless Access Points (APs) can approximatesuitablelength of the contention free period by conceding any received packet equal to the received packet in size. But as only one packet is transported during communication, the constraints of quality of service needed an enhancement. So a novel procedure, is projected in which two packets are conveyed in same interval with the AP, can resolve its wake-up timer and trigger sleep mode for the remaining of the Contention Free Period (CFP). Extensive computer based models of the new proposed method is established for the upgrading in terms of throughput and delay in voice influx.
- 1 January 2016
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Open AccessA Novel Technique of implementing Bidirectional Point Coordination Function for Voice Traffic in WLAN
WLANs have substituted wired networks and have formed an uprising in the area of communication. The Point Coordination Function (PCF) of the IEEE 802.11 protocol offersto sustain to real time traffic for Wireless Local Area Network (WLANs). However, PCF being a central polling protocol, a little bandwidth is exhausted on null packets and polling operating cost. To supplyimproved channel operation and to decreaseconsumption of bandwidth, an enhanced bidirectional broadcast of setextent known as Bi-Directional Point coordination Function (BD-PCF) is included into PCF. Based on this policy, wireless Access Points (APs) can approximatesuitablelength of the contention free period by conceding any received packet equal to the received packet in size. But as only one packet is transported during communication, the constraints of quality of service needed an enhancement. So a novel procedure, is projected in which two packets are conveyed in same interval with the AP, can resolve its wake-up timer and trigger sleep mode for the remaining of the Contention Free Period (CFP). Extensive computer based models of the new proposed method is established for the upgrading in terms of throughput and delay in voice influx.
- 1 April 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 273-277
Performance Analysis and Quality-of-Service Monitoring of Protected and Unprotected TCP Networks Using NCTUns Simulator
The research trends in computer and wireless communication networks are growingly facilitated with the emergence of various Network Simulation tools. Network simulation software enables users to model network behavior demonstrating the packet transmission behavior and interaction amongst the nodes. These models are further used to analyze and predict the behavior of a computer network on various devices such as links and applications. Network Simulators also permit these network topologies to be studied and tested on several popular protocols such as WLAN, Wi-Max, TCP, WSN, cognitive radio. This paper analyses the performance and quality-of service of packet flow in TCP networks in VANET using novel network simulator and emulator known as NCTUns. The investigation and study of various VANET topologies is facilitated by NCTUns architecture, simulation methodology, functional design and execution steps. NCTUns offers a series of advantages over other contemporary network simulators where the outcomes of diverse quality of service (QoS) and resource management and schemes implemented on higher-layer protocols can be easily studied. Most importantly, NCTUns allows testing of time-consuming, complex or extravagant topologies without emulating them using hardware. This paper reviews Diffserv model used for monitoring packet flows in TCP architecture and gives an insight into the analysis of such networks.
- 1 April 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 273-277
Performance Analysis and Quality-of-Service Monitoring of Protected and Unprotected TCP Networks Using NCTUns Simulator
The research trends in computer and wireless communication networks are growingly facilitated with the emergence of various Network Simulation tools. Network simulation software enables users to model network behavior demonstrating the packet transmission behavior and interaction amongst the nodes. These models are further used to analyze and predict the behavior of a computer network on various devices such as links and applications. Network Simulators also permit these network topologies to be studied and tested on several popular protocols such as WLAN, Wi-Max, TCP, WSN, cognitive radio. This paper analyses the performance and quality-of service of packet flow in TCP networks in VANET using novel network simulator and emulator known as NCTUns. The investigation and study of various VANET topologies is facilitated by NCTUns architecture, simulation methodology, functional design and execution steps. NCTUns offers a series of advantages over other contemporary network simulators where the outcomes of diverse quality of service (QoS) and resource management and schemes implemented on higher-layer protocols can be easily studied. Most importantly, NCTUns allows testing of time-consuming, complex or extravagant topologies without emulating them using hardware. This paper reviews Diffserv model used for monitoring packet flows in TCP architecture and gives an insight into the analysis of such networks.
- 1 September 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2084-2090
Investigating the security threats in Vehicular ad hoc Networks (VANETs): Towards security engineering for safer on-road transportation
The state-of-the-art improvements in cellular communication and ubiquitous availability of internet have led to significant breakthroughs in intelligent transportation systems where connectivity, autonomous driving and infotainment play a pivotal role in the enhanced driving experience. The Vehicle ad hoc Networks (VANET) have emerged as a distinguished branch of wireless communication pertaining to transportation systems. VANET is intended to dispense on-road vehicle safety and to boost the comfort experienced by drivers, passengers and other commuters. Whereas VANET offers exciting applications and explores unfamiliar dimensions in transportation, concerns regarding VANET security also continue to intensify. Security of vehicular networks, the authenticity and integrity of data dissemination remains a concern of utmost significance in VANET deployment. VANET architecture, by virtue of an abundance of networked vehicles, is susceptible to illegal use, unauthorized access, protocol tunneling, eavesdropping, and denial-of-service as the vehicles are unknowingly exposed to illegitimate information from unidentified adversaries. This paper investigates the security aspects of VANET and the attacks and vulnerabilities the VANET architecture is prone to. The study of security features and flaws is expected to lead to developed broadcasting and routing services, adding to the quality-of-service. Due to mobility of vehicles, large scale networks, rapidly restructuring nodes and frequently changing topological structure; a fundamental requirement of VANETs is to ensure safe transmission of the time critical data. The paper examines various security threats in VANETs, analyses how they are implemented and their impact on the VANET security architecture. A few gaps in the VANET security frameworks have also been highlighted which can be worked upon in the future.
- 1 September 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2084-2090
Investigating the security threats in Vehicular ad hoc Networks (VANETs): Towards security engineering for safer on-road transportation
The state-of-the-art improvements in cellular communication and ubiquitous availability of internet have led to significant breakthroughs in intelligent transportation systems where connectivity, autonomous driving and infotainment play a pivotal role in the enhanced driving experience. The Vehicle ad hoc Networks (VANET) have emerged as a distinguished branch of wireless communication pertaining to transportation systems. VANET is intended to dispense on-road vehicle safety and to boost the comfort experienced by drivers, passengers and other commuters. Whereas VANET offers exciting applications and explores unfamiliar dimensions in transportation, concerns regarding VANET security also continue to intensify. Security of vehicular networks, the authenticity and integrity of data dissemination remains a concern of utmost significance in VANET deployment. VANET architecture, by virtue of an abundance of networked vehicles, is susceptible to illegal use, unauthorized access, protocol tunneling, eavesdropping, and denial-of-service as the vehicles are unknowingly exposed to illegitimate information from unidentified adversaries. This paper investigates the security aspects of VANET and the attacks and vulnerabilities the VANET architecture is prone to. The study of security features and flaws is expected to lead to developed broadcasting and routing services, adding to the quality-of-service. Due to mobility of vehicles, large scale networks, rapidly restructuring nodes and frequently changing topological structure; a fundamental requirement of VANETs is to ensure safe transmission of the time critical data. The paper examines various security threats in VANETs, analyses how they are implemented and their impact on the VANET security architecture. A few gaps in the VANET security frameworks have also been highlighted which can be worked upon in the future.