Dr.L.M.I.LEO JOSEPH
@sru.edu.in
Assistant Professor / ECE
SR University ,warangal
EDUCATION
PhD VLSI
RESEARCH INTERESTS
MEDICAL IMAGING , SIGNAL PROCESSING , RECONFIGURABLE COMPUTING
Scopus Publications
Scopus Publications
R. Suganya, L.M.I. Leo Joseph, and Sreedhar Kollem
Elsevier BV
Manasa Koppula and Leo Joseph L.M.I
Seventh Sense Research Group Journals
- The Internet of Things has experienced explosive evolution as a ground-breaking phenomenon since its conception. The security sector has witnessed enormous growth in cyberattacks as a consequence of the increasing growth of IoT devices, which expanded the attack vector for hackers to carry out significantly more damaging vulnerabilities. A key component of assuring the cybersecurity of IoT is the identification of anomalies in network activity using an intrusion detection system. Conventional machine learning methods appear vain in the face of inconsistent network expertise and several attack tactics. Deep learning methods have proved their capability to recognize irregularities in a wide range of research fields accurately. An excellent substitute for conventional methods of anomaly detection and classification is Convolutional Neural Networks (CNN). In this research, a novel IDS-based improved CNN model for IoT networks has been developed. To solve the issue of overfitting and improve the sophistication of the classifier, various regularization techniques, including L1, L2, Dropout, and multi-regularization, have been deployed. The experimental findings demonstrate that, when contrasted to the other CNN2D models, the proposed method outperforms with an above 98% accuracy. The Detection Rate and False Discovery Rate of individual classes are above 0.9 and below 0.1, respectively.
L. M. I. Leo Joseph, V. Manonmani, S. Thulasi Prasad, V. Elamaran, Ganesan P, and G. Sajiv
IEEE
The woman safety is utmost priority at any situation. However, the woman harassment is common irrespective of places. It is necessary to improve the safety measures everywhere such as workplace, travel. There are number of methods proposed for woman safety. However, these methods have their own pros and cons. The work proposed a simple but smart light weight protection shoe for women safety. In this proposed system, we are trying to design a safety shoe which has smart features to defend and protect women in danger with a shock circuit placed right in front of the shoe. The shock circuit module helps to provide an electric shock to the assaulter. The proposed system uses an electric shock circuit, pulse sensor, 5V buzzer, USB Boost converter and NodeMCU board. The electric shock generator circuit produces a shockwave of 10mA which is enough to make a person to jump. The pulse sensor utilized to give the human heart rate. The USB Boost converter helps to charge small electronic gadgets. It also sends a notification to the concern people when the user is in danger. The unique feature of the suggested method is that the entire system triggered by a solitary press of a panic button. The proposed smart shoe designed will also act as a pulse monitoring device. It has a buzzer to alert nearby people and also used to recharge small electronic gadgets.
L.M.I.Leo Joseph, Edunuri Harini Reddy, Munigala Srinidhi, Pamu Venkata Saketh, Dharamsoth Mohan, and Ch. Rajendra Prasad
IEEE
The housing market involves consideration of numerous factors by prospective buyers that may change over time. As a result, the market for real estate prediction will constantly fluctuate. Machine Learning is used to analyze and predict real estate property prices. The XGBoost algorithm is used to achieve a high accuracy in forecasting the market values. The Primary Objective is to assist users in finding a suitable price according to their preferred geographical factors. The model provides accurate predictions for future real estate prices by analyzing the previous trend and market. This study sought to predict prices in Bengaluru City using the XGBoost Algorithm. The results of this study showed that the XGBoost Model achieved a prediction accuracy of 91.77%. Such an accurate valuation tool could benefit homeowners and prospective buyers when pricing properties, without requiring the help of a real estate agent.
R. Suganya, Leo Joseph, and Sreedhar Kollem
IEEE
Energy storage systems in electric vehicles come across boundaries interrelated to perilous parameters. There are challenging factors like charging infrastructure, constrained energy density which affects driving range, and battery degradation. The proposed system studies lithium-ion batteries' energy storage ability by considering three parameters: current, voltage, and temperature. The proposed model is simulated using MATLAB/ Simulink and studies the interplay of the considered parameters and is observed to be the energy-storing technique with their graphical analysis. The three-parameter outperforms the capacity of energy storage by its values that are not exceeded and limited to the ideal values which yields superior results, also essential for sustainable renewable energy sources, also for grid applications.
B. S. Sathish, T. Jerry Alexander, B. Padmavathy, L. M. I. Leo Joseph, P. Ganesan, G. Sajiv, and R. Murugesan
Springer Nature Singapore
P. Ganesan, L. M. I. Leo Joseph, V. G. Sivakumar, S. Thulasi Prasad, B. S. Sathish, and G. Sajiv
Springer Nature Singapore
P Ganesan, L.M.I. Leo Joseph, G. Sajiv, B Chandrasekaran, E. Manigandan, and V. Janakiraman
IEEE
The proposed work explains the comparative detailed study of multispectral image enhancement by means of contrast and decorrelation stretching. It is challenging to extract all necessary information from multispectral image which is a huge collection of information. Multispectral images are taken using different filters (sensors) that can recognize different spectral bands, giving us important details about the scene we’re looking at. Most of the data is intensified inside a narrow portion of the accessible dynamic range. So the true color composite image looks monotonous. In addition, in multispectral images the visible bands are highly correlated with each other. This is the reason they looks so dull. So it is obligatory to decorrelate the bands and enhance the contrast of the input data. In this study, the efficiency of the contrast and decorrelation stretching transformation on test images is compared. The experimental outcomes undoubtedly demonstrate that decorrelation stretching outperformed the contrast stretching in this context.
L. M. I. Leo Joseph, Nijaguna G. S, Janardhana D. R, Ramy Riad Al-Fatlawy, and D. Sharmila
IEEE
In Flying Ad Hoc Networks (FANET) consider of multiple Unmanned Aerial Vehicles (UAVs) with high mobility leads to quickly changing topologies, which are performed automatically onboard the system. To address these challenges, the proposed optimized LEACH protocol aims to increase the network cycle and the number of communicated packets while simultaneously reducing energy node discharge. Additionally, the Teaching Learning Based Optimization (TLBO) algorithm is involved to prevent rapid convergence, which may otherwise lead to getting easily trapped in a local optimum. The determination of the Cluster Head (CH) is performed statically and with fixed scalability to select the CH for transmitting to the Base Station (BS). The Teaching Learning based Improving Artificial Bee Colony Optimized (IABCO) method involves highly searching for high-quality food sources and ensuring sufficient population diversity through exchanging information and cooperation among the colony bees. Thus, the proposed algorithm combines the advantage of ABC strong global search and ability and TLBO rapid convergence. The proposed LEACH-Teaching learning based IABCO method demonstrates high performance compared to existing methods such as Moth Flame Optimization (MFO) and Improved Weighted and Location-Based Clustering (IWLC). Specifically, the IABCO method yields superior results, achieving an energy consumption of 0.1 J, Packet Delivery Ratio (PDR) of 0.99, and Network Lifetime of 15, outperforming existing methods in FANET.
Sreedhar Kollem, Ch. Rajendra Prasad, J. Ajayan, Sreejith S., LMI Leo Joseph, and Patteti Krishna
Bentham Science Publishers Ltd.
Background: In image processing, image segmentation is a more challenging task due to different shapes, locations, image intensities, etc. Brain tumors are one of the most common diseases in the world. So, the detection and segmentation of brain tumors are important in the medical field. Objective: The primary goal of this work is to use the proposed methodology to segment brain MRI images into tumor and non-tumor segments or pixels. Methods: In this work, we first selected the MRI medical images from the BraTS2020 database and transferred them to the contrast enhancement phase. Then, we applied thresholding for contrast enhancement to enhance the visibility of structures like blood arteries, tumors, or abnormalities. After the contrast enhancement process, the images were transformed into the image denoising phase. In this phase, a fourth-order partial differential equation was used for image denoising. After the image denoising process, these images were passed on to the segmentation phase. In this segmentation phase, we used an elephant herding algorithm for centroid optimization and then applied the multi-kernel fuzzy c-means clustering for image segmentation. Results: Peak signal-to-noise ratio, mean square error, sensitivity, specificity, and accuracy were used to assess the performance of the proposed methods. According to the findings, the proposed strategy produced better outcomes than the conventional methods. Conclusion: Our proposed methodology was reported to be a more effective technique than existing techniques.
Sandip Bhattacharya, Mohammed Imran Hussain, John Ajayan, Shubham Tayal, Louis Maria Irudaya Leo Joseph, Sreedhar Kollem, Usha Desai, Syed Musthak Ahmed, and Ravichander Janapati
Wiley
S. Sreejith, L.M.I. Leo Joseph, Sreedhar Kollem, V.T. Vijumon, and J. Ajayan
Elsevier BV
Sreedhar Kollem, Katta Ramalinga Reddy, Ch. Rajendra Prasad, Avishek Chakraborty, J. Ajayan, S. Sreejith, Sandip Bhattacharya, L. M. I. Leo Joseph, and Ravichander Janapati
Wiley
J. Ajayan, P. Mohankumar, D. Nirmal, L.M.I. Leo Joseph, Sandip Bhattacharya, S. Sreejith, Sreedhar Kollem, Shashank Rebelli, Shubham Tayal, and B. Mounika
Elsevier BV
L. M. I. Leo Joseph, Gubbala Jahnavi Deepika, P S Dinesh, S. Vijayashaarathi, and N. Samanvita
IEEE
Wireless Sensor Network (WSN) is group of spatially distributed sensors which monitors environmental exchanges that data with each other over wireless channels. Energy aware in WSN are referred to as saving or balancing energy between nodes in WSN. Energy saving protocols are utilized to minimize energy consumed through WSN. Energy balancing protocols effectively distribute energy consumption throughout network. In this research, the Modified Chaotic Grey Wolf Optimization (MCGWO) algorithm is proposed for the effective energy aware in WSN. The optimal cluster head and route paths are selected by utilizing the proposed optimization algorithm. the developed MCGWO algorithm minimizes the nodes’ energy utilization while increasing the data transmission in WSN. The performance of proposed technique is estimated by utilizing performance measure of throughput, packet delivery ratio, delay and energy consumption. The proposed method attained high packet delivery ratio of 94.4%, 93.7%, 92.1% and 90.9% for 50, 100, 150 and 200 nodes which is comparatively higher than other existing algorithms like Cuckoo Search Algorithm (CSA), Ant Colony Optimization (ACO), Firefly Optimization Algorithm (FOA) and Grey Wolf Optimization (GWO).
B. Durgabhavani, Vijaya Bhaskar Reddy Muvva, L.M.I. Leo Joseph, Duvvapu Mahesh Babu, and Manohara H. T.
IEEE
Over the past decades, an Autonomous Vehicles (AVs) driving functions have obtained the more interests from both the industry and academia. The AVs needed a basic planner for all the practicable scenarios and the recent research develops named a planner through the integrated scenario interpretation. Nevertheless, it may remarkably enhance the planner difficulty indeed in some simple operations. Examining the safety of AVs is the significant challenge in relation to their distribution on the roads because of large number of achievable conditions happened in traffic. In this research, the adaptive path planning approach of AV established multi-light trained Reinforcement Learning (RL) is proposed, aimed to enhance the fuel cost as well as AV comfort. By utilizing a logical Deep Q-learning function, the training approach obtains the key setting data collected through the vehicles as inputs, eventually outcomes an acceleration that enhances a cumulative reward. The proposed method attains better results by using the performance metrices of average acceleration, distance and computational time values about 0.149, 0.98 and 0.49 respectively compared with Pertinent Boundary-based Unified Decision (PBUD), Adaptive Mode Predictive Control (MPC) and Dynamic Decoupling Lane-Changing (DDLC).
Ganesan P, M. Ravichandran, L.M.I. Leo Joseph, B.S. Sathish, M. Raiendra Prasad, and G. Sajiv
IEEE
The role of image segmentation is inevitable in image analysis. There are many algorithms and methods for image segmentation. However it is tricky to choose best possible clustering procedure. The proper selection of clustering method leads to the successful outcome in image analysis. In segmentation process, data is fragmented into some reasonable regions (clusters). Elements (image pixels) in a particular cluster should be very analogous w.r.t some attributes like lightness, texture or color. In the proposed method, test image is clustered based on chrominance (color) using four segmentation methods, viz., FCM, PCM, PFCM and MFCM. The outcome of segmentation methods are transformed into binary image using otsu threshold method. The binary image of FCM is taken as a reference image because it is standard and fundamental one. The binary images of other methods are compared with reference image. The comparison is done using some image quality metrics such as dice, jaccard. The optimal segmentation method is decided based on the analysis of these image quality metrics.
Ganesan P, M. Ravichandran, B.S. Sathish, L.M.I.Leo Joseph, G. Sajiv, and R. Murugesan
IEEE
To improve the resolution and quality of a low level image, single image super resolution (SISR) is a challenging endeavor in the turf of computer vision. It is essential to many applications, including image editing, security systems, medical imaging, and others. In order to increase the spatial quality of a low level image, SISR attempts to approximate the high frequency details that are missing from the image. The field was formerly based on manually created features and interpolation techniques, but current developments in deep learning have completely changed it. The suggested work compares the effectiveness of bicubic interpolation and neural network based very deep super resolution for single image super resolution image quality improvement. Utilizing blind and complete reference picture quality measures, the effectiveness of the two approaches is evaluated. The effectiveness of the very deep super resolution for single image is clearly demonstrated by the experimental results.
M. Ravichandran, T. Jerry Alexander, B.S. Sathish, L.M.I. Leo Joseph, Ganesan P, and G. Sajiv
IEEE
The proposed work elucidates the role of metrics to decide the superiority of noisy and compressed images. The metrics approximate the superiority of degraded image (noisy or compressed) with or without the reference images. The computed numerical value based on this comparison finalizes the excellence of the image under test. The degradation in any stage of image processing (acquisition, transmission and reception) deteriorates the eminence of an image. It is challenging to understand the source and reason for low quality of the image. The suggested work analyzes the quality of salt and pepper noisy image and compressed image using blind and complete reference metrics. The analysis using full reference metrics is an unpretentious process. However, they do not related to human perception. In contrast, blind quality metrics demonstrated its hegemony in this facet.
Sandip Bhattacharya, L. M. I. Leo Joseph, Sheshikala Martha, Ch. Rajendra Prasad, Syed Musthak Ahmed, Subhajit Das, Debaprasad Das, and P. Anuradha
CRC Press
L.M.I. Leo Joseph, J. Ajayan, Sandip Bhattacharya, and Sreedhar Kollem
Wiley
Koppula Manasa and L. M. I. Leo Joseph
Springer Nature Singapore
Manasa Koppula and Leo Joseph L. M. I
IEEE
The Internet of Things (IoT) brings together more devices that can communicate with one another while requiring little user input. IoT is one of the computer disciplines that is expanding rapidly, but the fact is that with the increasingly intimidating Internet world, IoT is susceptible to different kinds of cyberattacks. Practical defenses against this, including network anomaly detection, must be built to secure IoT networks. Attacks cannot be completely prevented forever, but practical defense depends on the ability to identify an attack as soon as possible. IoT systems cannot be protected by conventional high-end security solutions because IoT devices have a limited amount of storage and processing capability. This suggests the need for the creation of smart network-based solutions for cyberattacks, such as Machine Learning (ML). Although the application of ML methods in detecting attacks has numerous studies in recent years, attack detection in IoT networks has received less attention. The major goal of this study is to create and evaluate a hybrid ensemble algorithm called LNKDSEA (Logistic regression, Naïve Bayes, K-nearest neighbor, Decision tree, and Support vector machine-based Ensemble Algorithm). The proposed approach can efficiently identify IoT network attacks including DDoS, information gathering, Malware, Injection attacks, and Man-in-The-Middle- Attack. The edge-IIoTset dataset is used to evaluate the proposed model. During the implementation stage, the proposed technique is evaluated by employing binary and multi-class (6 and 15 Class) classifications of cyberattacks, and high performance is accomplished.
P Ganesan, B. S. Sathish, L. M. I. Leo Joseph, R. Murugesan, G Sajiv, A. Akilandeswari, and M. Gnanaprakash
IEEE
In the proposed article, an innovative approach for the segmentation of region of interest (ROI) based on fuzzy and neural network (SOM-self organizing map) is investigated. In image analysis, clustering is an unique imperative practice wherein the complete data is separated as group of pixels and/or regions using its features. A satellite image is huge collection information. Many approaches and algorithms suggested for the dissection of satellite imagesto extract the indispensable and hidden information. In this work, fuzzy based FCM, PCM, PFCM, MFCM and SOM is applied to the satellite images to retrieve the information. The proposed method transforms the image in RG to HSL model prior to clustering. HSL model has some distinct characteristics such as perceptual, intuitive, polar coordinated, approximate the vision. The segmentation result is validated with image quality measures.
P Ganesan, B. S. Sathish, L. M. I. Leo Joseph, G. Sajiv, R. Murugesan, A. Akilandeswari, and S. Gomathi
IEEE
In human, skin tone is significantly varying from one extend (darkest) to other (lightest) due to the difference in the amount of pigmentation (melanin). Even though skin color detection and segmentation is the challenging and complex task which finds applications in face recognition, human tracking, video conferencing, content based image retrieval. The variation in the pigmentation is the consequence of the exposure to the amount of solar radiation and heredity. The segmentation is the clustering process wherein the whole image is clustered into small groups according to the color or texture characteristics. The attainment of image analysis is exclusively based on the upshot of the segmentation. The proposed segmentation procedure is provocative due to the attributes like illumination, the characteristics of the image capturing device, ethnicity and individual characteristics. In skin color segmentation, pixels clustered as the skinny or non-skin pixels. The proposed method utilized HSV color model as a tool for the skin tone detection and segmentation. The outcome of the proposed system is weighed against with RGB color model. Experimental results clearly illustrated that the proposed method has a very good efficiency to segment skin color pixels.