VISHWANATH SAMPAT MAHALLE

Vishwanath S. Mahalle, B.E. in Computer Science and Engineering from Shri Sant Gajanan Maharaj College of Engineering(SSGMCE), Shegaon (M.S., India) and M. E. in Computer Engineering from Shri Govindram Sucseria Institute of Technology & Science (SGSITS), Indore(M.P., India) and Ph.D. in Computer Science and Engineering from SGBAU, Amravati.

Vishwanath started his career in 1994 as a Lecturer in Department of Computer Science & Engineering at SSGMCE, Shegaon Professor. With 31 years of experience Vishwanath has now become the Asst. Professor at SSGMCE, Shegaon. He has taught various subjects & guided projects at Under Graduate and Post Graduate level.

He has authored/co-authored numerous research papers in reputed National & International conferences and journals. He has been worked as reviewer of International Conferences also worked as a resource person in AICTE approved STTPs. He was coordinator of various AICTE –ISTE approved workshops. He is a life member of ISTE and ISRD.

EDUCATION

B.E. (Computer Science and Engineering)
M.E. (Computer Engineering)
Ph.D. (Computer Science and Engineering)

RESEARCH, TEACHING, or OTHER INTERESTS

Computer Engineering, Computer Vision and Pattern Recognition

Scopus Publications

CBIR Using ANFIS Classification with DWT Feature Extraction Algorithms
Vishwanath S. Mahalle, Abhijit Banubakode, Maqsood A. Ansari, Vandana C. Bagal
Lecture Notes in Networks and Systems, 2025
A powerful method for interactive content-based image retrieval by variable compressed convolutional info neural networks
Vishwanath S. Mahalle, Narendra M. Kandoi, Santosh B. Patil
Visual Computer, 2024
Enhancing efficiency in content-based image retrieval system using pre-trained convolutional neural network models
Vishwanath S. Mahalle, Narendra M. Kandoi, Santosh B. Patil, Abhijit Banubakode, Vandana C. Bagal
Artificial Intelligence Machine Learning and User Interface Design, 2024
Traditionally, image retrieval is done using a text-based approach. In the text-based approach, the user must query metadata or textual information, such as keywords, tags, or descriptions. The effectiveness and utility of this approach in the digital realm for solving image retrieval problems are limited. We introduce an innovative method that relies on visual content for image retrieval. Various visual aspects of the image, including color, texture, shape, and more, are employed to identify relevant images. The choice of the most suitable feature significantly influences the system's performance. Convolutional Neural Network (CNN) is an important machine learning model. Creating an efficient new CNN model requires considerable time and computational resources. There are many pre-trained CNN models that are already trained on large image datasets, such as ImageNet containing millions of images. We can use these pre-train CNN models by transferring the learned knowledge to solve our specific content-based image retrieval talk. In this chapter, we propose an efficient pre-trained CNN model for content-based image retrieval (CBIR) named as ResNet model. The experiment was conducted by applying a pre-trained ResNet model on the Paris 6K and Oxford 5K datasets. The performance of similar image retrieval has been measured and compared with the stateof-the-art AlexNet model. It is found that the AlexNet architecture takes a longer time to get more accurate results. The ResNet architecture does not need to fire all neurons at every epoch. This significantly reduces training time and improves accuracy. In the ResNet architecture, once the feature is extracted, it will not extract the feature again. It will try to learn a new feature. To measure its performance, we used the average mean precision. We obtained the result for Paris6K 92.12% and Oxford5K 84.81%. The Mean Precision at different ranks, for example, at the first rank in Paris6k, we get 100% result, and for Oxford5k, we get 97.06%.
Knowledge representation in artificial intelligence - A practical approach
Vandana C. Bagal, Archana L. Rane, Debam Bhattacharya, Abhijeet Banubakode, Vishwanath S. Mahalle
Artificial Intelligence Machine Learning and User Interface Design, 2024
In the realm of artificial intelligence, knowledge representation is a vital aspect that enables effective information sharing and processing. Humans excel at sharing trusted information, which is acquired through rigorous testing and validation, resulting in what we commonly refer to as knowledge. The representation of knowledge can take various forms, such as graphs, maps, or textual formats. With the continuous evolution of the IT sector, the introduction of AI has simplified many tasks, often surpassing human capabilities and effortlessly handling even the most basic activities. However, understanding the concept of knowledge representation remains a fundamental question. In this research paper, we delve into the basics of knowledge representation to directly address this question. The understanding of knowledge representation is best achieved by examining the role knowledge plays in specific case studies or systems, which includes scientific reasoning and comprehension of the world. By exploring the intricacies of knowledge representation, we aim to provide a practical approach to its implementation in the field of artificial intelligence.
Teager Energy Operator: A Signal Processing Approach for Detection and Classification of Power Quality Events
V. S. Mahalle, G. N. Bonde, S. S. Jadhao, S. R. Paraskar
Proceedings of the 2nd International Conference on Trends in Electronics and Informatics Icoei 2018, 2018
In electrical power system power quality (PQ) disturbances are frequent; to detect these disturbances and classify them is must, so that mitigation action is carried out accordingly. This paper presents a signal processing based approach for detection and classification of PQ events, where Teager Energy Operator (TEO) is used to detect the event and statistical parameter are used to classify them. The TEO is applied on most of PQ problems such as voltage sag, swell, transients and interruptions. By applying TEO to transient PQ events, the detection flag is obtained at start and end of the event. The length of the detection flag depends on the selection of average value for setting the threshold. The standard deviation is determined for the starting duration of detection flag for every event. After setting the threshold, the events can be classified. This method is simple, faster and easy to implement.
Secure ranked query processing in location based services accessing outsourced spatial databases
Priyanka W. Falke, Vshwanath S. Mahalle
Proceedings of the International Conference on Inventive Systems and Control Icisc 2017, 2017
This paper introduce a geographic information is present on the internet which is in the scattered format. When user search for any place then various locations based service providers find out the results. The commonly used location based service providers are Google, Bing, Yahoo etc. we considers a novel distributed system for collaborative location-based information generation and sharing which become increasingly popular due to the explosive growth of Internet-capable and location-aware mobile devices. The results that are given by location based service providers not fully processed. The system consists of a data collector, data contributors, location-based service providers (LBSPs), and system users. The data collector gathers reviews about points-of-interest (POIs) from data contributors, while LBSPs purchase POI data sets from the data collector and allow users to perform spatial top-k queries which ask for the POIs in a certain region and with the highest k ratings for an interested POI attribute. The data contributor is the peoples or users who provide the information about the place. The user can gives comment and reviews to the product or the place. It's not necessary the LBSP provides proper information. In practice, LBSPs are untrusted and may return fake query results for various bad motives, e.g., in favour of POIs willing to pay. In our approach when the data contributor searches for any query, it will process by the LBSP. But before result showing to the user, it will process by the data collector. There are various results for the same query. So each query processed and special top result according to the location will provide to the users. This paper presents three novel schemes for users to detect fake spatial snapshot and moving top-k query results as an effort to foster the practical deployment and use of the proposed system.
Learning to recommend descriptive tags for health seekers using deep learning
Vidhi L. Chawda, Vishwanath S. Mahalle
Proceedings of the International Conference on Inventive Systems and Control Icisc 2017, 2017
Health plays an important role for human happiness and well being. Automatic disease prediction is important to overcome the issues of health seekers. Generally people use Google to search their queries and that search engine respond them with the answer but that answer is in scattered format. User not gets exact answer for his / her queries. So we are going to implement this paper. Here we propose a novel deep learning scheme to infer the disease according to questions of health seekers. In this work we first analyze and categorize needs of health seekers and ask for manifested symptoms for disease prediction. Then user will search for their query. Then the query get processed to give prediction of disease to the user or health seekers. Here concept of hidden layers is get used. First medical signatures mines from raw features. These features and signatures deems as input node in one layer and hidden nodes in subsequent layer. This paper presents idea of deep learning architecture which is used in health care domain for the diagnosis of diseases.
Overview of secure distributed de-duplication system with improved reliability
Shweta A. Junghare, V. S. Mahalle
Proceedings of the International Conference on Inventive Systems and Control Icisc 2017, 2017
Day by day the use of memory is increases rapidly. A lot of data created every day so the data management is become a critical task. Most of time much amount of memory wastage because of same copies present on various locations. The process of eliminating the repeated or duplicates copies of data is called as Data deduplication. This data deduplication process is widely used in cloud storage to decrease storage space and upload bandwidth. However there is only one copy of each file stored even if such a file is owned by huge number of users. By using, deduplication system progress of storage utilization and reliability is increases. In addition, the dare of privacy for sensitive data also take place when they are outsourced by users to cloud. In this we use new distributed deduplication systems with upper dependability in which the data chunks are distributed from corner to cornering multiple cloud servers. The safety needs of data privacy and tag stability are also accomplish by introducing a deterministic secret sharing scheme in distributed storage systems, instead of using convergent encryption as in previous deduplication systems. A deduplication technique, on the other hand, can reduce the storage cost at the server side and save the upload bandwidth at the user side.
Resolve the classification problem on secure encrypted relational data
Anjali J. Rathod, V. S. Mahalle
Proceedings of the International Conference on Inventive Systems and Control Icisc 2017, 2017
Data mining is a powerful new technique to discover knowledge within the large amount of the data. A number of theoretical and practical solutions to query processing have been proposed under various scenarios. With the recent popularity of cloud computing, data owners now have the opportunity to outsource not only their data but also data processing functionalities to the cloud. Because of data security and personal privacy concerns, sensitive data (e.g., medical records) should be encrypted before being outsourced to a cloud, and the cloud should perform query processing tasks on the encrypted data only. These tasks are termed as Privacy Preserving Query Processing (PPQP) over encrypted data. These protocols protect the confidentiality of the stored data, user queries, and data access patterns from cloud service providers and other unauthorized users. Several queries were considered in an attempt to create a well-defined scope. These queries included the k-Nearest Neighbor (kNN) query, advanced analytical query, and correlated range query. This paper presents protocols utilize an additive cryptography base privacy preserving data mining technique at different stages of query processing to achieve the best performance all computations can be done on the encrypted data.
Secure spatial Top-k query processing via untrusted location based services accessing outsourced databases
Priyanka W. Falke, V. S. Mahalle
Proceedings of 2017 International Conference on Innovations in Information Embedded and Communication Systems Iciiecs 2017, 2017
This paper introduce a geographic information is present on the internet which is in the scattered format. When user search for any place then various locations based service providers find out the results. The commonly used location based service providers are Google, Bing, Yahoo etc. we considers a novel distributed system for collaborative location-based information generation and sharing which become increasingly popular due to the explosive growth of Internet-capable and location-aware mobile devices. The results that are given by location based service providers not fully processed. The system consists of a data collector, data contributors, location-based service providers (LBSPs), and system users. The data collector gathers reviews about points-of-interest (POIs) from data contributors, while LBSPs purchase POI data sets from the data collector and allow users to perform spatial zonal top-k queries which ask for the POIs in a certain region and with the highest k ratings for an interested POI attribute. The data contributor is the peoples or users who provide the information about the place. The user can gives comment and reviews to the product or the place. It's not necessary the LBSP provides proper information. In practice, LBSPs are untrusted and may return fake query results for various bad motives, e.g., in favour of POIs willing to pay. In our approach when the data contributor searches for any query, it will process by the LBSP. But before result showing to the user, it will process by the data collector. There are various results for the same query. So each query processed and special top result according to the location will provide to the users. This paper presents three novel schemes for users to detect fake spatial snapshot and arranging top-k query results as an effort to foster the practical deployment and use of the proposed system.
An overview of secure friend matching in mobile social networks
Ruchika Ganvir, Vishwanath Mahalle
Iciiecs 2015 2015 IEEE International Conference on Innovations in Information Embedded and Communication Systems, 2015
Enhancing the data security in Cloud by implementing hybrid (Rsa & Aes) encryption algorithm
Vishwanath S Mahalle, Aniket K Shahade
2014 International Conference on Power Automation and Communication Inpac 2014, 2014
Identification of the real source of DDOS attack by FDPM in IP traceback system
Gaurav D. Barokar, V.S. Mahalle
Proceedings Uksim Amss 8th European Modelling Symposium on Computer Modelling and Simulation EMS 2014, 2014