NANDHAGOPAL N
@excelinstitutions.com
PROFESSOR,DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING
EXCEL ENGINEERING COLLEGE
RESEARCH INTERESTS
IMAGE PROCESSING ,EMBEDDED SYSTEM TECHNOLOGY,ELECTRON DEVICES
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
M. Kalimuthu, M. Ramya, S. Sreethar, and N. Nandhagopal
Informa UK Limited
M. Kalimuthu, S. Sreethar, Ramya Murugesan, and N. Nandhagopal
World Scientific Pub Co Pte Ltd
Automatic facial expression recognition (FER) is utilized in various applications like psychoanalysis, intelligent driving, robot manufacturing, etc. Numerous researchers have been looking for better techniques to improve the accuracy of FER. In fact, FER under laboratory conditions has almost achieved top accuracy. Besides, label deviations or errors caused by annotators’ subjectivity also make the FER task much tougher. Thus, more and more researchers begin to find new ways to handle with the FER problems. In this work, a new deep learning (DL) model called dense squeeze network with improved red deer optimization (DenseSNet_IRDO) is proposed for the recognition of facial emotions. The steps used for FER are pre-processing, fused deep feature extraction-selection and classification. Initially, the facial images are pre-processed using improved trilateral filter (ITF) for improving the quality of images. Next, the fusion of feature extraction and selection is performed using the DenseSNet. Here the extraction of deep features is done with the dense network and the relevant features are selected with the squeeze network. Finally, the last layer of squeeze network performs the classification of various facial emotions. Here, the loss in the classification is optimized using IRDO. This DenseSNet_IRDO architecture is more robust and avoids overfitting that occurs while training the small dataset. The datasets used in this work are CK[Formula: see text], JAFEE and FERFIN. The proposed FER classification using datasets CK[Formula: see text], JAFEE and FERFIN with DenseSNet_IRDO model achieved the accuracy of 99.91%, 99.90% and 99.89%, respectively. Thus, the proposed DenseSNet_IRDO classifier model obtained higher accuracy in the detection of FER than other methods.
A. Vasantharaj, S. Anbu Karuppusamy, N. Nandhagopal, and Ayyem Pillai Vasudevan Pillai
Elsevier BV
V. Nivedita and N. Nandhagopal
Springer Science and Business Media LLC
A. Soujanya and N. Nandhagopal
Computers, Materials and Continua (Tech Science Press)
A Vasantharaj, N Nandhagopal, S Anbu Karuppusamy, and Kamalraj Subramaniam
Elsevier BV
S. Sreethar, N. Nandhagopal, S. Anbu Karuppusamy, and M. Dharmalingam
Springer Science and Business Media LLC
P. Siva Satya Sreedhar and N. Nandhagopal
Computers, Materials and Continua (Tech Science Press)
S NAVANEETHAN and N NANDHAGOPAL
Springer Science and Business Media LLC
The pupil detection algorithm plays a key role in the non-contact tono-meter, auto ref-keratometry and optical coherence tomography in medical ophthalmology diagnostic equipment. A major challenge associated with pupil detection techniques is the use of conventional neural networks based on algorithms, integro-differential operator and circular hough transform, which leads to inefficient use of hardware resources in FPGA. To overcome this, using an average black pixel density technique, the proposed human eye pupil detection system is used to easily recognize and diagnose the human eye pupil area. Double threshold, logical OR, morphological closing and average black pixel density modules are involved in the proposed solution. To test the proposed method, the near infrared (NIR) iris databases are being used, namely: CASIA-IrisV4 and IIT Delhi and have achieved 98% percent accuracy, specificity, sensitivity. The proposed work was synthesized via Zynq XC7Z020 FPGA and the results are compared with previous approaches.
S. Sreethar, N. Nandhagopal, S. Anbu Karuppusamy, and M. Dharmalingam
Springer Science and Business Media LLC
G. Keerthana, P. Anandan, and N. Nandhagopal
Springer Science and Business Media LLC
V. Nivedita and N. Nandhagopal
Springer Science and Business Media LLC
The habit of using mobile devices increasing constantly, Considerably MANETs as the nodes are mobile, Trust management can help to improve the security in routing that guaranteed QoS provisioning in MANETs to achieve better deterministic behavior and appropriately the networks delivered the information in a better way and it can be well gain to exploit the network resources. Trust Calculation solves the problem of providing corresponding access control based on judging the quality of Sensor Nodes and their services and to analyze the route and alternate to route for efficient data transmission. This paper deals with the efficient approach based on multi-hop and relay dependent communication for enhancing the security. The improvement of QoS is based on Random Repeat Trust Computational Approach obtain a various trust evaluation Stages by estimating the direct and indirect trust degree to avoid the incorrect trust derivation problem and later than update the node trust of routing table as detection of malicious node subsequent to attain the trusted QoS routing of data transmission. Then it investigates the node location and distances among the nodes for data transmission to verify the false injection. To evaluate the trustworthy paths and nodes using to design and develop a trust based QoS routing integrated by Random Repeat Trust Computational Approach to improve QoS. Simulation results show that the progressing QOS and distrust worthy node detection of the proposed system more than 30% when compared to the existing system.
V. Nivedita and N. Nandhagopal
Computers, Materials and Continua (Tech Science Press)
Soujanya A
Institute of Advanced Scientific Research
V. Nivedita and N. Nandha Gopal
IEEE
With exponential growth of vast amount of data or traffic or information from various heterogeneous IOT devices stumbles the network as a result the response time of data delivery from cloud to end IOT devices turns to high latency, network traffic, data congestion, consume high bandwidth, huge power consumption curb the development of IOT specifically in the time sensitive applications. For instance health related services, and traffic light systems, etc. In order to cope up with mass connections, surplus data management, and resource allocation for billions of IOT devices the new paradigm has been emerged known as Fog Computing also called edge computing (fog means closer to ground). In fog computing resource allocation the main challenge is considered in this paper. The following challenges in resource allocation are i) IOT devices might be shortcomings in their capacity (eg: data, processor, memory, CPU, bandwidth, etc), ii) shortcoming in their network resources, iii) increased latency or response time to the centralized data server from IOT devices. By considering the following issues we have presented the model or architecture for IOT service delegation to mitigate the network latency and improve the Quality of Service(QOS) and proposed algorithm for optimizing in terms of capacity and big data distribution among fog and cloud computing.
S. Navaneethan, N. Nandhagopal, and V. Nivedita
American Scientific Publishers
Threshold based pupil detection algorithm was found tobe most efficient method to detect human eye. An implementation of a real-time system on an FPGA board to detect and track a human's eye is the main motive to obtain from proposed work. The Pupil detection algorithm involved thresholding and image filtering. The Pupil location was identified by computing the center value of the detected region. The proposed hardware architecture is designed using Verilog HDL and implemented on aAltera DE2 cyclone II FPGA for prototyping and logic utilizations are compared with Existing work. The overall setup included Cyclone II FPGA, a E2V camera, SDRAM and a VGA monitor. Experimental results proved the accuracy and effectiveness of the hardware realtime implementation as the algorithm was able to manage various types of input video frame. All calculation was performed in real time. Although the system can be furthered improved to obtain better results, overall the project was a success as it enabled any inputted eye to be accurately detected and tracked.
N. Nandhagopal, K. Rajiv Gandhi, and R. Sivasubramanian
Maxwell Scientific Publication Corp.
Our Goal is to increase the accuracy of brain tumor detection and classification and thereby replace conventional invasive and time consuming techniques. Here a new technique is proposed to classify the brain MRI images and to detect the brain tumor using probabilistic neural network. The proposed methodology comprises of three phases. 1) Discrete wavelet transform 2) Modified region growing algorithm and 3) Probabilistic neural network. Initially, the input is subjected to discrete wavelet transform. It is used to extract the wavelet coefficients from the MRI images. Then the texture features are extracted using modified region growing algorithm from the input MRI brain images, which are obtained from the database. The texture features taken in to consideration are correlation and contrast. Soon after, the extracted features are fed as the input to the Hybrid ANN-PNN to classify the brain MRI images. Based on the features extracted the tumor will be detected and will be classified as Benign and malignant tumor. The proposed methodology will be implemented in MATLAB 7.12 with different datasets. The performance will be analyzed with existing detection methods and we will prove our efficiency in terms of accuracy.
N. Nandha Gopal and M. Karnan
IEEE
Magnetic Resonance Imaging (MRI) is one of the best technologies currently being used for diagnosing brain tumor. Brain tumor is diagnosed at advanced stages with the help of the MRI image. Segmentation is an important process to extract suspicious region from complex medical images. Automatic detection of brain tumor through MRI can provide the valuable outlook and accuracy of earlier brain tumor detection. In this paper an intelligent system is designed to diagnose brain tumor through MRI using image processing clustering algorithms such as Fuzzy C Means along with intelligent optimization tools, such as Genetic Algorithm (GA), and Particle Swarm Optimization (PSO). The detection of tumor is performed in two phases: Preprocessing and Enhancement in the first phase and segmentation and classification in the second phase
M. Karnan and N. Nandha Gopal
IEEE
In this paper, a novel approach to MRI Brain Image segmentation based on the Hybrid Parallel Ant Colony Optimization (HPACO) with Fuzzy C-Means (FCM) Algorithm have been used to find out the optimum label that minimizes the Maximizing a Posterior (MAP) estimate to segment the image. There are M colonies, M-1 colonies treated as slaves and one colony for master. Each colonies visit all the pixels with out revisit. Initially, initialize the pheromone value for all the colonies. Posterior energy values or fitness values are computed by Markov Random Field. If this value is less than global minimum, the local minimum is assigned to global minimum. The pheromone of the Ant that generates the global minimum is updated. At the final iteration global minimum returns the optimum threshold value for select the initial clustering the FCM implementation in the brain Magnetic Resonance Image (MRI) segmentation.