Arunmozhi Sinouvassane
@mvit.edu.in
Professor ECE
Manakula Vinayagar Institute of Technology
RESEARCH INTERESTS
Wireless Communication
IOT and AI and ML
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
A. Baskaran, S. Arunmozhi, and S. Vishnu
IEEE
Streetlights provide illumination at night and provides safety on roads. Conventional monitoring of streetlights involves periodic human inspection which is time consuming, costly and sometimes unsafe. We proposed a remote monitoring and control of streetlights mounted on power transmission line posts using Power Line Carrie r Communication (PLCC) technology. PLC C allows transmission of data over existing power cables and does not require additional infrastructure. A PLCC modem is installed at each streetlight fixture which transmits status updates like lamp ON/OFF, voltage, current drawn etc. to a data concentrator unit via the low frequency power line network. The concentrator unit sends this information to a control center via a wireless or wired backhaul network. Such a solution ensures 24x7 monitoring of streetlights with minimum additional hardware investment. Mathematical models of different PLCC modulation schemes are developed and their performance is analyzed and compared through simulations. These Data were synchronized with SCADA to monitor the Streetlight. Practical implementation issues are also discussed.
A. Rama, Robertas Damaševičius, S. Arunmozhi, Mazin Abed Mohammed, Ragheed Hussam, and V. Rajinikanth
IEEE
Building health and structural integrity are important to ensure occupant safety and long-term durability. Monitoring and assessing building condition regularly ensures that potential problems are identified early. This research aims to develop and implement a deep-learning (DL) supported tool to examine the concrete surface images collected using digital camera. The data gathering forms the first stage and the chosen DL algorithm is used for feature extraction followed by the feature selection and then fusion is done and finally a fivefold cross verification schema is applied for a binary classification. This work considered the EfficientNetV2 (ENV2) model with variants like small, medium and large for the examination. The investigation is implemented using individual and fused deep-features and the performance of the developed tool is confirmed using the achieved accuracy. The outcome of this research confirms that the proposed tool produces a betterment of >96% accuracy with individual-features and 100% accuracy with fused-features. These results confirm that the proposed scheme provides significant result during the crack detection.
A. Rama, K.B. Sudeepa, S. Arunmozhi, Mazin Abed Mohammed, Aqeel Ali, and V. Rajinikanth
IEEE
Breast cancer is considered a severe illness in the female society, and if left untreated, it can be fatal. It is always desirable to detect the BC early utilizing a selected imaging strategy. Thermogram supported breast abnormality detection is one of the recent technique and this gives the necessary information in the form of the distributed thermal pattern. This research aims to implement the Convolutional-Neural-Network (CNN) based segmentation technique to extract breast region from the chosen thermogram. This scheme's multiple stages include: (i) data collecting and processing, (ii) implementation of CNN segmentation to extract the breast, (iii) comparing it to the binary-mask and computing performance metrics, and (iv) performance evaluation and verification of the chosen CNN techniques. Pre-trained CNN segmentations are used in this work to extract the necessary section from the thermogram, and the experimental results show that the VGG-UNet methodology helps to extract the essential region with an enhanced accuracy of 97.260.64% when compared to other CNN approaches.
A. Rama, N. Mythili, M.P. Rajakumar, S. Arunmozhi, Mazin Abed Mohammed, and V. Rajinikanth
IEEE
Deep-learning (DL) applications that are used real-time across various industries have gained a lot of traction and have become increasingly popular, especially when it comes to data-driven recommendation systems. This work aims to develop a DL scheme to support the music-recommendation system (MS) based on the music data. The various phases of this scheme includes; (i) data collection and signal-image conversion to get the necessary RGB scale images from the data, (ii) pre-trained DL based feature extraction, and (iii) deep-features based detection to recommend the appropriate music. This research considered the classic- (CL) and pop-music (PO) for the examination and the achieved results are evaluated to substantiate the performance of this arrangement. In this work, the signal-image conversion procedure is implemented to convert 1D signal to 2D image and then it is examined using proposed technique. The experimental outcome is separately presented for (i) spectrogram and (ii) synchro-extracting-transform and obtained results are presented. The experimental investigation is presented with MobileNet variants and this study authorizes that the implemented scheme achieved a better detection MobileNetV2 (>99%) compared to other schemes in this study.
R. Santhosh, S. Arunmozhi, and Nilanjan Tewari
IEEE
Computer algorithm supported data-analysis is one of the common practices to solve the chosen data-evaluation tasks. Recently, the computer algorithm assisted image-evaluation is emerged as one of the capable research field. The purpose of this research is to use leaf information to create a deep learning scheme to investigate rice plant disease (RD). This approach consists of three stages: (i) gathering and resizing leaf images; (ii) extracting deep features using selected DS; and (iii) using SoftMax based binary classification with 5-fold cross validation. In this work, 1000 photos from each class are examined, and the categorization result that is obtained is confirmed. This study considered the ResNet and ResNetV2 variants for the examination and the achieved result is separately verified for 50, 101 and 152 layered schemes. This investigation task confirms that the ResNet variants provided >91% accuracy and the ResNetV2 variants provided an accuracy of >94%. This demonstrates that the proposed method performs satisfactorily on the selected leaf data, and going forward, real-time data may be taken into consideration to validate the technique's effectiveness for RD detection.
A. Rama, M. P. Rajakumar, N. Mythili, S. Arunmozhi, Mazin Abed Mohammed, and V. Rajinikanth
IEEE
The lung is one of the prime organs, and any disease in the lung causes mild to severe breathing problems; untreated lung disease will lead to several complications. Tuberculosis (TB) is a lung ailment that needs premature recognition and handling. The primary objective is to employ the deep-learning (DL) based TB detection using chest $X$-rays. Various stages of the proposed scheme consist of (i) data collection and resizing, (ii) DL-supported feature extraction, (iii) binary classification and five-fold cross-validation, and (iv) comparison with earlier results and confirming the merit of the scheme. This research implements EfficientNet (EN) variants to classify the chosen $\\mathrm{X}$-rays into healthy/TB classes using the SoftMax classifier. The proposed scheme with EN_B2 (ENB2) has been successful in providing an accuracy of $96{\\% }$ as far as detection accuracy is considered when compared to other methods. The superiority of the suggested strategy is also confirmed by an analysis using the most recent technology, which confirms the worth of the proposed system on the chosen $\\mathrm{X}$-ray imagery.
N. Sri Madhava Raja, S. Arunmozhi, Hong Lin, Nilanjan Dey, and V. Rajinikanth
IGI Global
In recent years, a considerable number of approaches have been proposed by the researchers to evaluate infectious diseases by examining the digital images of peripheral blood cell (PBC) recorded using microscopes. In this chapter, a semi-automated approach is proposed by integrating the Shannon's entropy (SE) thresholding and DRLS-based segmentation procedure to extract the stained blood cell from digital PBC pictures. This work implements a two-step practice with cuckoo search (CS) and SE-based pre-processing and DRLS-based post-processing procedure to examine the PBC pictures. During the experimentation, the PBC pictures are adopted from the database leukocyte images for segmentation and classification (LISC). The proposed approach is implemented by considering the RGB scale and gray scale version of the PBC pictures, and the performance of the proposed approach is confirmed by computing the picture similarity and statistical measures computed with the extracted stained blood cell with the ground truth image.
Dhananjay Nigam, Shilp Nirajbhai Patel, P. M. Durai Raj Vincent, Kathiravan Srinivasan, and Sinouvassane Arunmozhi
Hindawi Limited
Secure identification is a critical system requirement for patients seeking health-related services. In the event of critical, aged, or disabled patients who require frequent health treatments, quick and easy identification is vital. Researchers describe the notion of the unprotected environment in this study, in which patients can receive health services from the hospital’s smart and intelligent surroundings without the use of explicit equipment. Patients would interact directly with the environment and be identified through it. We suggest a biometric-based authentication technique for the unprotected hospital environment that also safeguards the patient’s identity privacy. Furthermore, we demonstrate that this authentication technique is resistant to many well-known assaults, including insider attacks, replay attacks, and identity privacy. Doctors and other staff members showed enthusiastic responses after installing 2-factor authentications, as it makes their workflow efficient and makes things easier for patients. It also lets them focus on other factors rather than worrying about data security; hence, we need biometric authentication in intelligent and privacy-preserving healthcare systems. The paper deals with two-factor biometric authentication, and despite the added security, two-factor authentication adoption is said to be poor. It is due to a lack of awareness and difficulty to use and configure two-factor authentication (2FA) into a particular application by some individuals who struggle with the concept of authentication and its technology. Also, many 2FA methods in widespread use today have not been subjected to adequate usability testing. Research focuses on the point that there is still a large section of people unaware of the use of biometric systems to protect their online data. Researchers collected quantitative and qualitative data from 96 individuals during a two-week between-subjects usability survey of some common and rarely used 2FA approaches. The survey allowed the researcher to investigate which authentication methods are given higher priority and why, along with the relationship between different usage patterns and perceived usability, and identify user misconceptions and insecure habits to determine ease of use. It was observed that the biometric-based method was given the utmost preferability.
R. Sofia, R. Valli, and S. Arunmozhi
IEEE
The state of trauma continues to be a contentious issue and inevitably causes a concern in the sense it disrupts the routine life of an individual. The increasing cases of trauma in recent times owing to a number of reasons become even more serious and can further hamper the life. It invites measures to quickly recognize the state and serve as a diagnostic platform where remedies can be initiated. So this paper has been done in motivation of identifying the trauma patients using the FFNN, RNN and ANFIS.
S. Arunmozhi, V. Rajinikanth, and M.P. Rajakumar
IEEE
Pneumonia is one of the communicable illnesses in humans which generally affect lungs. The untreated pneumonia will cause severe problems in elderly people (age>65 years) and children (age<5 years) and hence the early detection and treatment is commonly preferred to recover the infected patients from the disease. The infection is lungs are commonly diagnosed with chest radiographs (X-ray) due to its clinical significance. This work implements the deep-learning (DL) scheme to detect the pneumonia. The disease detection performance of the DL scheme is confirmed using a binary classification achieved with SoftMax classifier unit. During this assessment 2000 (1000 healthy and 1000 pneumonia) images are considered for the appraisal and the necessary performance measures are computed to confirm the performance. The experimental outcome of AlexNet offered an accuracy of >98% on the considered image database.
Anukirthika T. S., Dellecta Jessy Rashmi R, N. Sri Madhavaraja, S. Arunmozhi, and K. Suresh Manic
IEEE
Hypertension or high blood pressure is sometimes called a "silent killer", since it has no warning signs, yet it can lead to life-threatening conditions. The good news is prevention and treatment for hypertension can be done with early diagnosis. So, there is a need for real time monitoring of blood pressure that can be done using Internet of Things (IoT) assisted health monitoring system. The proposed system collects the user’s health parameters using IoT sensors and identifies the stage of hypertension. The vital point of this system is to continuously generate emergency alerts of blood pressure fluctuation and dangerous changes in any other health parameters to hypertensive users on their mobile phones.
S. Arunmozhi, Aditya Prabhakara Kamath, and Venkatesan Rajinikanth
IEEE
The lung infection in human causes various respiratory problems, which affects the oxygen supply in blood stream. Tuberculosis (TB) is a severe lung disease in humans and the uncontrolled TB causes various respiratory problems, including death. TB is also a communicable disease and appropriate diagnosis and treatment will reduce the severity of the disease. This research aims to implement a novel disease diagnosis procedure to detect the TB infection in chest X-ray with better accuracy. This research employs the concatenation of deep-features (DF) with handcrafted-features (HF) to improve the diagnostic accuracy. In this work, the VGG16 is employed to extort the DF and the HF is obtained using the discrete-wavelet-transform (DWT) approach. The optimal values of DF and HF are arranged as per their rank and then a serial feature concatenation is employed to get a new 1D feature (DF+HF). This feature is then considered to train and validate the performance of considered classifiers using a 5-fold cross validation and it offered an accuracy of >97% with the Fine-Tree classifier.
Ajanthwin Prabagar, N. Sri Madhavaraja, S. Arunmozhi, and K. Suresh Manic
IEEE
With the increasing vehicle population in urban areas, it is hard to find a suitable parking lot to park the vehicle. Also, traditional parking systems involve human labour to monitor parking of vehicles in the specified slot. This can be overcome by using computer vision to identify the available slots and notify the drivers about the availability of slots, as well as improve the security by monitoring the entry and exit of the vehicle with their number plate. This system uses image processing to identify unoccupied slots and provide users a hassle-free experience. The availability of slots is processed and suggestions are provided accordingly. This reduces enormous amount of time spent in search of a parking slot. The number plate of the vehicles parked are noted along with their entry and exit timings.
R. SambathKumar, S. Gowshameed, and S. Arunmozhi
IEEE
(WSN) is used for determining the Indoor Positioning of objects and persons in recent years. WSN has been implemented in indoor positioning applications such as real-time tracking of humans/objects, patient monitoring in health care, navigation, warehouses for inventory monitoring, shopping malls, etc. But one of the problems while implementing WSN an Indoor positioning system is to ensure more coverage large number of sensors must be deployed which increases the installation cost. So, in this paper, MATLAB GUI named Sensor Network Localization Explorer to analyze the impact of node density on indoor aligning localization schemes. Later Kalman filter with the indoor positioning system to increase the reliability and reduce localization error of the system is introduced with a lesser number of nodes.
V. Rajinikanth, R. Sivakumar, D. Jude Hemanth, Seifedine Kadry, J. R. Mohanty, S. Arunmozhi, N. Sri Madhava Raja, and Nguyen Gia Nhu
Springer Science and Business Media LLC
D. Jude Hemanth, V. Rajinikanth, Vaddi Seshagiri Rao, Samaresh Mishra, Naeem M. S. Hannon, R. Vijayarajan, and S. Arunmozhi
Springer Science and Business Media LLC
Yongdong Wang, Liangliang Zhao, S. Arunmozhi, and N. Sri Madhava Raja
Elsevier BV
P. Resmi, R. Reshika, N. Sri Madhava Raja, S. Arunmozhi, and Vaddi Seshagiri Rao
Springer Singapore
S. Manasi, M. Ramyaa, N. Sri Madhava Raja, S. Arunmozhi, and Suresh Chandra Satapathy
Springer Singapore
Jenif D Souza W.S. and Parvathavarthini B.
IEEE
Intrusion detection has a prominent part for ensuring the information security. Machine learning approaches are followed to detect intrusion or anomaly of a network. The network traffic produce large amount of data, the Analyzing and monitoring the data is the biggest challenge here. To overcome that feature elimination or selection is done before classification. The dataset has some features which are irrelevant which makes the detection process slower and degrades the system performance. In order to improve the performance, this system identifies the features which are irrelevant and eliminated it. The feature selection is achieved by using Recursive Feature elimination method. For the selected feature classification is done by using classification model. The proposed system use KDD CUP 99 dataset. In this system four classifier models such as LDA, SVMr, Random forest and Adaboost are used, among that the Adaboost gives 99.75 % sensitivity and 95.69 % specificity which are higher when compared to other classifier. Using this system unknown future attacks can also be detected.
B. Nirupriya, P. Atilakshmy, G. Jayashree, P. Deepa, and S. Arunmozhi
IEEE
Lung is the one of the vital internal organs responsible to supply the oxygen to other body parts throughout the life span. In humans, the abnormalities in lung arise due to various reasons and the timely screening of the lung abnormality will help to regulate/cure the disease. The common infections in lung are due to tuberculosis, pneumonia and cancer. In which, the Lung-Cancer (LC) is very cruel and untreated LC will lead to death. The LC is normally seen as a large Lung-Nodule (LN) and in clinical diagnosis, the detection and categorization of the LN plays a major role in disease detection and treatment planning process. The proposed work aims to implement a methodology to extract the LN from the Computed-Tomography (CT) image with a considerable accuracy. This work implements Social-Group-Optimization (SGO) and Kapur's threshold (SGO+KE) to enhance the CT image. Later the Active-Contour (AC) and Watershed-Segmentation (WS) is executed to extract the LN. The merit of the proposed work is confirmed based on the performance measures attained with the proposed tool.
P. Monica, K. Priyanga, S. Keerthana, N. Sri Madhava Raja, and S. Arunmozhi
IEEE
Recently, a considerable number of research works are proposed to examine the abnormality in the medical images using semi-automated and automated techniques. The final goal of the implemented technique is to provide a perfect disease evaluation procedure, which helps to assist the doctor in the decision making process. This research aims to develop a semiautomated technique to examine the cancerous section from the Skin-Melanoma-Picture (SMP) of the benchmark PH2 database. This work implements Firefly-Algorithm (FA) assisted Shannon's thresholding to enhance the picture and Level-Set based segmentation to extract the melanoma segment. The PH2 dataset is associated with 200 test images along with the ground-truth (GT) provided by a skin expert. This work also implements a comparison of GT with extracted melanoma and computes the various performance measures essential to confirm the superiority of proposed technique. Proposed work is executed on 200 numbers of RGB class images and the average results attained is considered to confirm the superiority of the methodology.
A. Bakiya, K. Kamalanand, S. Arunmozhi, and V. Rajinikanth
IEEE
Pathological variation in biological soft tissues are commonly interrelated with changes in their mechanical as well as electrical and properties, which helps to distinguish abnormalities. The interrelation between the dielectric and viscoelastic properties is not well established in the biological soft tissue analysis. In this work, an effort has been made to develop a mathematical model to interrelate the dielectric properties and viscoelastic properties of the soft tissues, in frequency domain. The proposed mathematical models have been derived using standard rheological model namely Zener model and dielectric model known as the Debye model. This work is highly useful for predicting the viscoelastic characteristics of the soft tissues using measurements of dielectric quantities as a function of frequency.