An Early Detection of Multi-Class Skin Lesion Using Convolutional Neural Network Architecture and Kernel Soft Plus Extreme Learning Mohana Prasad Mendu, Ashraf Shaqadan, Sk. Rabbani Basha, V Senthil Kumar, Preeti Khurana, L Deepak IEEE International Conference on Electronic Systems and Intelligent Computing Icesic 2026 Proceedings, 2026 A CNN or Convolutional Neural Network is a deep learning model that uses three components’ layers: Convolutional layers are composed of multiple filters for the extraction of spatial features in the images. Pooling layers can be used for reducing overall size (dimensions) of the data set. Fully connected layers are composed of multiple neurons that can perform analysis for classification or prediction. The authors present a hybrid approach, MSLCKE (Multi-Stage Learning Convolutional Network with Kernel-based Extreme Learning), that helps develop intelligent models for skin cancer detection. The goal of their model is to improve accuracy in diagnosing skin cancer while decreasing time required to process patient’s data. The authors develop a new standard for skin cancer detection using HAM10000 dataset and enhance dermatoscopic images and therefore make use of a CNN Model based on GoogleNet & ResNet to acquire more features from an image of their developing system. To expedite training, an Improved Extreme Learning Machine (IELM) classifier enhances accuracy. Compared to CNN-only methods, MSLCKE shows superior accuracy across diagnostic categories with lower error rates. Key parameter analysis includes accuracy, recall, precision, computational time, and complexity.
Secured Home Automation System Through IEEE 802.15.4-Based Efficient Internet of Things Communications S. Hariharan, G P Susanna Wesley, Mohana Prasad Mendu, Bharani B R, T Vijetha, Ramya M Proceedings Iceconf 2025 2025 2nd International Conference on Artificial Intelligence and Knowledge Discovery in Concurrent Engineering, 2025 The emergence of smart home energy systems has created challenges related to maintaining secure, efficient, and intelligent communication among connected devices. The Secured Home Automation System Through IEEE 802.15.4-Based Efficient Internet of Things Communications (SHIEIC) model has been developed to address the security, energy management, and communication efficiency challenges of smart home energy systems. This model incorporates networking and deep learning techniques, employs the 6LoWPAN protocol for low-power communication, and uses Hyperledger Fabric for device registration, including a multi-parameter trust evaluation for secure onboarding. It features a smart communication model using HomeIO for voice control, PSO for resource management and routing, and a deep learning-based MQTT attack detection module for real-time threat analysis.
Deep Learning For Dark Web Text Analysis: A Convolutional Approach To Content Categorization Sasidhar Reddy Gaddam, Pathan HussainBasha, Mohana Prasad Mendu, Purimitla Ramalingamma, Buddaraju Revathi, V T Ram Pavan Kumar M Proceedings of the International Conference on Research in Computational Intelligence and Communication Networks Icrcicn, 2025 The rapid expansion of dark web platforms has intensified the need for efficient and automated text analysis systems capable of identifying illicit and high-risk activities. Traditional deep learning models for dark web content classification often suffer from high computational complexity, limited robustness to adversarial manipulation, and reduced effectiveness in multilingual environments. This paper proposes an Enhanced Text Convolutional Neural Network (ET-CNN) framework designed to address these challenges. The proposed model integrates adaptive word embeddings, adversarial training based on gradient perturbation, and cross-lingual embedding alignment to improve classification reliability and generalization. Experiments conducted on a large-scale dataset consisting of 100,000 labeled dark web forum posts demonstrate that the ETCNN achieves an accuracy of 94.2% and an F1-score of 92.5%, while maintaining significantly lower computational overhead compared to transformer-based architectures. The results indicate that the proposed approach provides an effective balance between performance, robustness, and efficiency, making it suitable for real-time dark web monitoring and analysis.
Automatic Weed Detection and Removal Using AI K. Kishore Babu, D. Sharada Mani, P. Seetha Lakshmi, B. Suresh, D. V. Ashok, Mohana Prasad Mendu Proceedings of International Conference on Visual Analytics and Data Visualization Icvadv 2025, 2025 Real-time weed detection and identification in autonomous weeding systems are obtained by combining computer vision and machine learning. In this research study, High-resolution images of the field are taken by cameras, which are then processed with convolutional neural networks (CNNs) to precisely detect various types of vegetation. The proposed system is trained on a large and varied dataset, providing high accuracy under different environmental conditions. Through accurate targeting and eliminating undesirable plants with mechanical claws or targeted sprayers, this method maximizes minimal weed presence while maintaining soil and crop integrity. The process improves sustainable agriculture, lowers operational expenditure, and maximizes overall efficiency.
A Novel Methodology Design to Predict Unexpected Suggestions in E-Commerce through Artificial Intelligence (AI) based Collaborative Filtering G. Karthikeyan, Shaik Rabbani Basha, Nabeel Al-Milli, Mohana Prasad Mendu, T Vijetha, D. Harika Proceedings 2025 International Conference on Recent Innovation in Science Engineering and Technology Icriset 2025, 2025 In the evolving landscape of e-commerce, delivering personalized and unexpected recommendations has emerged as a critical factor in user involvement and sales conversion. This study proposes a new collaborative filtering approach to artificial intelligence (AI), which besides being accurate in its forecast of user preferences is aimed at coming up with serendipitous recommendations that are not part of any predictive models. The suggested system optimizes sophisticated methods such as Neural Collaborative Filtering (NCF), Matrix Factorization, and Graph Neural Networks (GNNs) and couples them with surprise scoring unit together with a context-cognizant personalization. The evaluation of the system was done with the application of a real-world e-commerce dataset comprising more than 1 million users-items interaction. The proposed model attained an average Precision of 0.784, Recallof 0.844, Surpriseof 0.631 and NDCG of 0.681, which surpassed the baseline models, such as Matrix Factorization and User-Based Collaborative Filtering models. The peculiar feature of this methodology is a multi-layer approach which consists in a combination of user behavior, contextual information and novelty-driven optimization. The model becomes dynamically responsive to personal preferences by using temporal patterns, session frequency, and context in real-time like device and location. In addition, the surprise module collects things that are not expected but are valid on the basis of ranking (entropy) and similarity deviation. This study shows that collaborative filtering with AI may highly increase the quality of recommendations facilitated by novelty and context-awareness. The results have heavy implications on e-commerce websites attempting to heighten consumer satisfaction, enhance retention, and induce arbitrary shopping actions.
CNN-Powered Early Detection of Mango Leaf Diseases for Sustainable Fruit Farming Burri Vijaya Kumari, Varagani Tejaswi, Mohana Prasad Mendu, D.V. Ashok, Devalla Manogna, Dulla Srinivas Proceedings of the 6th International Conference on Electronics and Sustainable Communication Systems Icesc 2025, 2025 Fruit cultivation contributes significantly to addressing global food problems and nutritional needs. However, plant diseases are pretty common in tropical South Asia, which ultimately decreases the yield radically, causing huge losses to the farmers. Under such circumstances, early detection of leaf diseases is essential for sustaining healthy crops. Here in, a Convolutional neural network- based method is presented using leaf images for identifying seven common mango diseases by Mango Scan. The model was trained with the help of a special dataset of locally sourced photos on the pattern of mango diseases in India. As a result, it can correctly categorise practically all prevalent mango illnesses. This gave extraordinary performance in $\mathbf{5}$-fold cross-validation evaluations with averages of $\mathbf{9 8. 5 5 \%, ~} \mathbf{9 9. 5 0 8 \%, ~} \mathbf{9 9. 4 5 \%, ~} \mathbf{9 9. 4 7 \%}$, and $\mathbf{9 9. 8 7 8 \%}$ in that order, outperforming the leading industry models such as VGG16 and AlexNet. Eventually, Mango Scan will help in the identification of early indications of illness and thus can be used for improving mango production that would develop and improve the economy of the country.
Detection of Aircrafts with Satellite Using R-CNN M Rama, N. Saikiran, Thella Sunitha, CH. Silpa, Mohana Prasad Mendu, Poreddy Jayaraju Conference Proceedings 2025 IEEE 4th International Conference on Data Decision and Systems Icdds 2025, 2025
Post-COVID Cardiovascular Risk Assessment Using an Efficient Machine Learning Model D Srinuvasa Rao, Desidi Narsimha Reddy, Lova Naga Babu Ramisetti, M S Radha Manga Mani, Harikrishna Pathipati, Siva Kumar Pathuri, Mohana Prasad Mendu 2024 3rd International Conference for Advancement in Technology Iconat 2024, 2024 Individuals who were exposed to COVID-19 in different parts of their bodies, including heart attacks, faced major consequences. The frequency of cardiovascular diseases is increasing in the healthcare industry. This illness most commonly affects people aged 20 to 48. In supervised learning, data mining mainly depends on categorization algorithms. Using various combinations of calculations and algorithms, the information obtained from hospital data analysis is used for early-stage prediction of post-Covid cardiovascular disease. Machine learning is one of the strange inventions that have been widely used in a variety of fields, including the application of medical services to disease prediction. We examined the accuracy of machine learning algorithms that can be used in this study to predict heart disease analyzes and predict overall risk. One of the most current data mining methods is classification. The classification method’s primary goal is to keep the data in the correct class. In this work, classification methods such as Logistic Regression Classifier, Naive-Bayes, and Ensemble Cardiovascular Prediction Classification Algorithm were used (ECVDPCA) and the dataset is taken from Kaggle repository. The goal of this study is to find the most exact classification approach and to improve accuracy in predicting Post-Covid Cardio-Vascular illness. Python was used to get the ideal solution in this case to predict the accuracy of the classifier in which the proposed classifier gave the best accuracy when compared with the other two classifiers i.e., 95
RECENT SCHOLAR PUBLICATIONS
An Early Detection of Multi-Class Skin Lesion Using Convolutional Neural Network Architecture and Kernel Soft Plus Extreme Learning MP Mendu, A Shaqadan, SR Basha, VS Kumar, P Khurana, L Deepak 2026 International Conference on Electronic Systems and Intelligent … , 2026 2026
Deep Learning For Dark Web Text Analysis: A Convolutional Approach To Content Categorization SR Gaddam, P HussainBasha, MP Mendu, P Ramalingamma, B Revathi 2025 Seventh International Conference on Research in Computational … , 2025 2025
Detection of Aircrafts with Satellite Using R-CNN M Rama, N Saikiran, T Sunitha, CH Silpa, MP Mendu, P Jayaraju 2025 IEEE 4th International Conference on Data, Decision and Systems (ICDDS … , 2025 2025
Secured Home Automation System Through IEEE 802.15. 4-Based Efficient Internet of Things Communications S Hariharan, GPS Wesley, MP Mendu, B BR, T Vijetha 2025 2nd International Conference on Artificial Intelligence and Knowledge … , 2025 2025
CNN-Powered Early Detection of Mango Leaf Diseases for Sustainable Fruit Farming DS Burri Vijaya Kumari, Varagani Tejaswi, Mohana Prasad Mendu, D.V.Ashok ... 6th International Conference on Electronics and Sustainable Communication … , 2025 2025
A Novel Methodology Design to Predict Unexpected Suggestions in E-Commerce through Artificial Intelligence (AI) based Collaborative Filtering G Karthikeyan, SR Basha, N Al-Milli, MP Mendu, T Vijetha, D Harika 2025 International Conference on Recent Innovation in Science Engineering … , 2025 2025
Automatic Weed Detection and Removal Using AI KK Babu, DS Mani, PS Lakshmi, B Suresh, DV Ashok, MP Mendu 2025 International Conference on Visual Analytics and Data Visualization … , 2025 2025
Post-COVID Cardiovascular Risk Assessment Using an Efficient Machine Learning Model DS Rao, DN Reddy, LNB Ramisetti, MSRM Mani, H Pathipati, SK Pathuri, ... 2024 3rd International Conference for Advancement in Technology (ICONAT), 1-6 , 2024 2024
Factors influencing the Behavior of the mobile phone users to switch their service Providers in Andhra Pradesh. MM Prasad, DP Kumar International Journal in Management & Social Science 4 (10), 253-267 , 2016 2016 Citations: 8
Web Information collection using Personalized Ontology Model MP Mendu, K Kiran 2013
MOST CITED SCHOLAR PUBLICATIONS
Factors influencing the Behavior of the mobile phone users to switch their service Providers in Andhra Pradesh. MM Prasad, DP Kumar International Journal in Management & Social Science 4 (10), 253-267 , 2016 2016 Citations: 8
An Early Detection of Multi-Class Skin Lesion Using Convolutional Neural Network Architecture and Kernel Soft Plus Extreme Learning MP Mendu, A Shaqadan, SR Basha, VS Kumar, P Khurana, L Deepak 2026 International Conference on Electronic Systems and Intelligent … , 2026 2026
Deep Learning For Dark Web Text Analysis: A Convolutional Approach To Content Categorization SR Gaddam, P HussainBasha, MP Mendu, P Ramalingamma, B Revathi 2025 Seventh International Conference on Research in Computational … , 2025 2025
Detection of Aircrafts with Satellite Using R-CNN M Rama, N Saikiran, T Sunitha, CH Silpa, MP Mendu, P Jayaraju 2025 IEEE 4th International Conference on Data, Decision and Systems (ICDDS … , 2025 2025
Secured Home Automation System Through IEEE 802.15. 4-Based Efficient Internet of Things Communications S Hariharan, GPS Wesley, MP Mendu, B BR, T Vijetha 2025 2nd International Conference on Artificial Intelligence and Knowledge … , 2025 2025
CNN-Powered Early Detection of Mango Leaf Diseases for Sustainable Fruit Farming DS Burri Vijaya Kumari, Varagani Tejaswi, Mohana Prasad Mendu, D.V.Ashok ... 6th International Conference on Electronics and Sustainable Communication … , 2025 2025
A Novel Methodology Design to Predict Unexpected Suggestions in E-Commerce through Artificial Intelligence (AI) based Collaborative Filtering G Karthikeyan, SR Basha, N Al-Milli, MP Mendu, T Vijetha, D Harika 2025 International Conference on Recent Innovation in Science Engineering … , 2025 2025
Automatic Weed Detection and Removal Using AI KK Babu, DS Mani, PS Lakshmi, B Suresh, DV Ashok, MP Mendu 2025 International Conference on Visual Analytics and Data Visualization … , 2025 2025
Post-COVID Cardiovascular Risk Assessment Using an Efficient Machine Learning Model DS Rao, DN Reddy, LNB Ramisetti, MSRM Mani, H Pathipati, SK Pathuri, ... 2024 3rd International Conference for Advancement in Technology (ICONAT), 1-6 , 2024 2024
Web Information collection using Personalized Ontology Model MP Mendu, K Kiran 2013
