Volume 8 Issue 3
Author's Name: B.V. Srikanth, M. Hitha Chowdary, M. Suvishal Sen Kasyap, K. Anjali Reddy.
Abstract— Vision related diseases like glaucoma, diabetic retinopathy, macular degeneration, and cataracts are becoming increasingly common, making early diagnosis essential to prevent blindness and improve quality of life. However, traditional diagnostic methods can be time-consuming, reliant on expert ophthalmologists, and sometimes subject to human error. In many resource-limited areas, access to quality eye care is even more restricted, contributing to the growing challenge of preventable blindness. This project explores an innovative solution leveraging machine learning, particularly deep learning techniques like Convolutional Neural Networks (CNNs), to automate eye disease diagnosis. The goal is to develop a reliable, efficient, and scalable system that can analyze retinal images to detect and classify various eye conditions. By enabling accurate early detection, this approach has the potential to significantly reduce the risk of permanent vision loss and expand access to quality eye care worldwide
Author's Name: Sathana yadav, Shakuntala devi
Abstract— This Rover stands for Remotely Operated Video Enhanced Receiver. Rover is next generation of Bluetooth, infrared & cellular services. Rover technology can be ranging from power full laptop to simple cellular phones. The technology which enables the scalable location aware computing. This involves automation availability of information & services based on current location of user. This user make avail location aware computing through his PDA.PDA is a hand held computer used as palmtop computer. PDA’s commonly have colour screen & audio capabilities to be used as mobile phone, web browsers. Many PDA’s can access the internet, intranet & extranet via Wi-Fi. Many PDA’s employ touch screen technology.
Author's Name: Jayendra Kumar, Devulapally Yathish, Dasari Ragini, Chindam Sreeja
Abstract— A consortium of organizations collaborates and exchanges information to create synergies in their operations. Centralized systems of secure transferring of data cannot provide distributed trust and transparency. Blockchain technology can be used to transfer data securely and transparently. This paper proposes a blockchain based secure transferring of data. It can be used by a consortium of organizations to securely exchange files in a distributed fashion. Hyperledger Fabric, an enterprise blockchain framework, is used for blockchain network setup and the development of smart contracts. The Inter Planetary File System (IPFS) is used for storing files in a distributed way. The paper provides the workflow for identity management and file-sharing processes. The proposed system allows a consortium of organizations to share files with confidentiality, integrity, and availability using blockchain
Author's Name: M. Gokila Priya, R. Gowsalya
Abstract— In Wireless Sensor Networks thousands of nodes are deployed in environment where there is no access to replace batteries. So, designing an energy efficient routing algorithm is a major challenge. Numbers of clustering algorithms such as LEACH, GSA, FCM, and DRINA have been developed to improve the energy balance of the WSN's as energy is the main aspect of WSN's during data transmission. Heuristic algorithms can find best solution in less time. These algorithms are used with mathematical and AI problems to find optimal solution. In this paper combination of fuzzy clustering and PSO algorithm is used to create clusters and find optimal route. Then, the proposed method is implemented in net beans. The proposed method is compare with two tire distributed fuzzy logic based protocol TTDFP which is investigated by energy consumption, packet delivery ratio, events monitored by sink, and event loss ratio.
Author's Name: G. RAJASEKHAR, D. HEMASREE, U. VIVEK RAJ, D. SUMA RAVALI
Abstract— This paper presents the development of a forecasting application for India's energy consumption using machine learning techniques. Energy demand prediction plays a crucial role in ensuring efficient resource allocation and sustainability. This study employs XGBoost, a gradient boosting algorithm, to analyze historical energy consumption data along with meteorological factors. The forecasting system, implemented using Streamlit, provides users with an interactive interface for visualization and prediction. The methodology includes data preprocessing, feature engineering, model training, and evaluation. The model is assessed using Mean Squared Error (MSE) and feature importance analysis. The results demonstrate that the application provides reliable energy consumption forecasts, supporting informed decision-making for policymakers and energy providers.
Author's Name: C.V.Ramasamy, M. Mohana Priya
Abstract— This Nowadays increasingly enterprises and organizations are hosting their information into the cloud, in order to decrease the IT maintenance cost and develop the data reliability. However, facing the numerous cloud vendors as well as their heterogeneous pricing policies. customers may well be at a loss with which cloud(s) are suitable for storing their information and what hosting policy is cheaper. The all-purpose status quo is that customers usually put their information into a single cloud (which is subject to the vendor lock-in risk) and then simply trust to chance. This paper proposes a novel information hosting scheme (named CHARM) which integrates two answer functions desired. The first is selecting several suitable clouds and an appropriate unemployment strategy to store information with minimized monetary costand definite availability. The second is triggering a transition process to re-distribute information according to the variations of information access pattern and pricing of clouds. We evaluate the performance of CHARM using both trace driven simulations and prototype experiments. The results show that compared with the major existing schemes; CHARM not only saves around 20% of monetary cost but also exhibits sound adaptability to information and price adjustments
Author's Name: Meenatchi Sundaram, M.Sabari nathan, N.Bala, M. Rafiq
Abstract— The improvement in the properties of SHTS subsequent to applying AFRP is talked about right now and its polymerization impact on reinforcing. To build up a correlation on the ongoing exploration pattern right now, extraordinary method for retrofication plot was engaged with this examination, by following an act of winding or helical wrapping of AFRP to accomplish proceeds with firmness with a uniform solidarity over the stature of the section. To examine the proposed reinforcing plan, a similar report has been finished as for the customary methodology. A progression of test examination was done to think of the outcome and later a concise conversation has been finished with respect to the use of AFRP in various fields of Designing. Absolutely 21 examples were casted both in even and winding jacketing and tried tentatively under hub compressive burden by supporting a few parameters to watch the variety in the difference in the properties of SHTS to check the pivotal burden conveying limit alongside the firmness and Young's modulus. The trial examination demonstrated that there is a surprising improvement in the properties of AFRP reinforced examples as for various parameters after the application AFRP and the impact of its polymerization with the holding specialist. Along these lines after the fortifying of segment examples with AFRP, the general augmentation in the heap ringing limit of the SHTS was 23.27% and furthermore the proposed plan of winding wrapping gave a better outcome as looked at than the conventional technique for level stripping.
Author's Name: R.Vivekanandan, S.M.Manoj kumar, S.Rajesh
Abstract— This A Model based Controller for the Continuous Stirred Tank Reactor (CSTR) is designed and developed using Proportional, Integral and Derivative (PID) Controller for the process of the biological system in the bioreactor. The model of the controller is based on the First principle method guided by experimental results. The mathematical model is a linearized representation of the state space at the point of operation and also the representation of the input output transfer function. The PID controller evaluation and simulation is simulated using process system MATLAB. The PID controller is the most versatile controller in the industry that provides the desired control action needed by adjusting those parameters to control a mechanism. To predict the plant restriction, the PID controller utilizes the control action and also the process model is explicated. The parameters of the reactor such as flow, tank level and temperature has to be controlled in the process of Continuous Stirred Tank Reactor (CSTR). In CSTR microorganism inoculated under different conditions. The processed control algorithm enhances the conditions for the growth of micro-organisms in the bioreactor by effectively regulating the tank level and temperature of the process.
Author's Name: Nitin yadav, Shanmugadevi
Abstract— We propose an Energy-efficient location-aware clone detection protocol in densely deployed WSNs, which can guarantee successful clone attack detection and maintain satisfactory network lifetime. Specifically, we exploit the location information of sensors and randomly select witnesses located in a ring area to verify the legitimacy of sensors and to report detected clone attacks. The ring structure facilitates energy-efficient data forwarding along the path towards the witnesses and the sink. We theoretically prove that the proposed protocol can achieve 100 percent clone detection probability with trustful witnesses. We further extend the work by studying the clone detection performance with untrustful witnesses and show that the clone detection probability still approaches 98 percent when 10 percent of witnesses are compromised. Moreover, in most existing clone detection protocols with random witness selection scheme, the required buffer storage of sensors is usually dependent on the node density, while in our proposed protocol, the required buffer storage of sensors is independent of number of nodes but a function of the hop length of the network radius h. Extensive simulations demonstrate that our proposed protocol can achieve long network lifetime by effectively distributing the traffic load across the network
Author's Name: P. Swarajya Lakshmi, B. Akash Rao, N. Sidhartha, K. Kalyan Chary
Abstract— Customer segmentation plays a vital role in modern business strategies, especially in the competitive e-commerce landscape. Understanding customer needs and identifying potential buyers at the right time is crucial for businesses aiming to stay ahead. By categorizing customers into distinct segments, businesses can create targeted marketing strategies that enhance customer satisfaction and increase sales. This paper explores the application of clustering techniques, with particular emphasis on the K-Means algorithm. Known for its efficiency, simplicity, and proven effectiveness, K-Means offers businesses a powerful tool for achieving accurate and actionable customer segmentation results. The paper further investigates how K-Means can be applied to large-scale datasets, ensuring scalability and adaptability to various business needs. By examining case studies and real-world applications, it highlights how this algorithm contributes to informed decision-making and the development of personalized marketing campaigns. Ultimately, this approach not only improves customer engagement but also drives business growth through optimized resource allocation and targeted communication strategies.ers.
Author's Name: Mangal Kalyani, Gupta Raju
Abstract— In this paper we are merely want to explore digital image processing which is a motivating field because it provides the higher image information for human understanding and process the image for storage, transmission, and illustration for machine perception. Image process could be a technique to reinforce raw pictures received from cameras/sensors placed on satellites, aircrafts or photos taken in traditional everyday life for varied applications. Pre-processing is the initial process applied on images which are at lowest level of abstraction. Also Shape analysis methods play an important role in systems for object recognition, matching, registration, analysis, accurate detection of image, Research in shape analysis has been motivated in part, for fast recognition of image from large data base. Accessing the desired and relevant image from large data base in an efficient manner is another motive for research. Various techniques of shape detection are based on shape boundary or interior.
Author's Name: Raju Patel, Rishil Mayank
Abstract— The use of different types of pavement block has been increased in India a decade ago. The specific requirements of paver block has been increased in the areas of footpaths and parking areas. Port land Cement Concrete and Asphalt Concrete are the most common roadway and highway construction material used. It is very necessary to use the materials which is durable and have the high compressive strength and can absorb the water to greater extent to have the greater water absorption value. In this project we have given greater stress on about the use of materials and machinery involved for the manufacture of different types of paver block in construction Purposes. For design of paver block the tests is to be performed is used to find the compressive strength and water absorption. The blocks having greater compressive strength and of less cost and can be used for much more time have been studied out in this project.
Author's Name: P. Ravinder Rao, Geethika Y, Y Shivaiah, A. Ranjith Reddy
Abstract— This paper introduces EcoMed, an advanced system utilizing Convolutional Neural Networks (CNN) to accurately identify medicinal plants and diagnose diseases affecting them. The system addresses the critical need for efficient, automated tools in plant identification and disease management. By analyzing leaf images, EcoMed empowers agricultural professionals, herbalists, and researchers to enhance plant health monitoring and expand knowledge of medicinal plants. The paper details the system architecture, data processing techniques, CNN model design, and evaluation metrics, demonstrating superior performance in identification and diagnosis accuracy.
Author's Name: Santhosh, Sai Laxmi
Abstract— Identifying of the plant diseases is essential in prevention of yield and volume losses in agriculture Product. Studies of plant diseases mean studies of visually observable patterns on the plant. Health surveillance and detecting diseases in plants is essential for sustainable development agriculture. It is very difficult to monitor plant diseases manually. It requires a lot of experiences in work, expertise in these field plant diseases and also requires excessive processing time. Therefore; image processing is used to detect plant diseases. Disease detection includes steps such as acquisition, image Pre-processing, image segmentation, feature extraction and Classification. We describe these methods for the detection of plant diseases on the basis of their leaf images; automatic detection of plant disease is done by the image processing and machine learning. The different leaf images of plant disease are collected and feature extracted of the various machine learning methods
Author's Name: Shubham Kumar, Kanchan Bhatt
Abstract— Weather Forecasting is the attempt to predict the weather conditions based on parameters such as temperature, wind, humidity and rainfall. These parameters will be considered for experimental analysis to give the desired results. Data used in this project has been collected from various government institution sites. The algorithm used to predict weather includes Neural Networks(NN), Random Forest, Classification and Regression tree (C &RT), Support Vector Machine, KNN Nearest neighbor. The correlation analysis of the parameters will help in predicting the future values. This web based application we will have its own chat bot where user can directly communicate about their query related to Weather Forecast and can have experience of two-way communication.
Author's Name: Rarishma Kaushik, Rahesh Parmar
Abstract— Air pollution and its harm to human health has become a serious problem in many cities around the world. In recent years, research interests in measuring and predicting the quality of air around people has spiked. Since the Internet of things has been widely used in different domains to improve the quality for people by connecting multiple sensors. In this work an IOT based air pollution monitoring with prediction system is proposed. The internet of Things is a action interrelated computing devices that are given unique identifiers and the capability of exchange information over a system without anticipating that human to human or human to machine communication. The deep learning algorithm approach is to evaluate the accuracy for the prediction of air pollution. The main objective of the project is used to predict the air Quality. The large dataset works with LSTM for better air quality prediction. The prediction accuracy of air quality with LSTM, the evaluation indicator Root means square error is chosen to measure performance.
Author's Name: P. Kavitha, S. Ramadevi
Abstract— Sentimental analysis is also called "opinion mining" analyses attitudes and classifies text views. It relates to the use of natural language processing, text, and linguistic processing. A huge amount of data is created with the rapid growth of web technologies. Social networking sites are now popular and normal places where feelings can be shared by short messages. These sentiments involve happiness, sadness, anxiety, fear, etc. The analysis of short texts tends to recognize the crowd's sentiment. Sentiment Analysis on IMDb moviereviews describes a reviewer's general feeling or impression of a movie. Since the perceptions of humans improve the effectiveness of products & since a movie'ssuccess or failure depending on its review, costs are rising, and a good sentiment analysis model needs to be developed, that classifies moviereviews. Machine learning methods use ML algorithms to carry out sentiment analysis as a standard classification problem using syntactic and language characteristics. There are some methods of machine learning used for sentiment analysis in this paper. Most of the sentiment analysis is performed using SVM, RF, ANN, and NB, Algorithms of DT, BN, & KNN.
Author's Name: Aaditya Sharma, Aman Sharma, Punit Sharma, Punit Ekka
Abstract— Physically challenged people or the people with disabilities are those section of the society in which the society embarks no decision or gives any judgement about their physical abilities of the working body parts and the other requirements of them according to their physical fitness as well, the present generation has the awareness regarding keeping themselves fit and sportingly active but on the other hand the people who are physically challenged are unable to work on the same. The project hereby tries to implement the solution for the same by giving the various physical exercises that could be done by the physically challenges people according to their working body parts and according to their ability accordingly.
Author's Name: Khaily Mohammed Shah, Sathyam Pandey
Abstract— A human brain responds at a much faster rate to images and the information it contains. An image is considered as proof of past events that have occurred, but in today's world where editing tools are made available so easily tampering of images and hiding the original content has become too mainstream. The identification of these tampered images is very important as images are considered as vital sources of information in crime investigation and in various other fields. The image forgery detection techniques check the credibility of the image. Various research has been carried out in dealing with image forgery and tampering detection techniques, this paper highlights various the type of forgery and how they can be detected using various techniques. The fusion of various algorithms so that a complete reliable type of algorithm can be developed to deal mainly with copy-move and image splicing forgery. The copy-move and image splicing method are main focus of this paper.
Author's Name: Dasari Savith, Adusumalli Sai Lochan, Kandula Shivani
Abstract— The proper execution of identification and classification of fruit diseases are crucial to maintaining optimal crop yields and reducing economic loss in the agricultural sector. Traditional methods of manual inspection are error-prone, susceptible to bias, and require extensive labor effort. We, in the present research, introduce a methodology based on deep learning for automatic fruit disease classification and identification via Convolutional Neural Networks (CNNs) in combination with GrabCut segmentation.[3] The suggested system scans fruit images using GrabCut to segment images efficiently, isolating affected regions from the background. Subsequently, the extracted features are evaluated with a CNN model to detect the severity and type of the disease.[6] Our method achieves maximum accuracy by leveraging the deep learning capability of detecting complex image patterns, thus leading to enhanced precision in disease classification. The model is trained and tested with a dataset comprising varied diseased and healthy images of fruits, demonstrating marked improvement in performance compared to traditional machine learning techniques. The system offers a robust and efficient solution for real-time disease detection, advantageous for farmers through the application of timely interventions in preventing further disease propagation and potential loss in yield.
Author's Name: Aditya Swami, shish Kumar
Abstract— By this research, we want to display the key factors of how effective blogging as technology is. We collected information curated by several researchers on the same for understanding their point of view as well. By doing this we found out what makes blogging the best way of sharing information online. We also analyzed the technological aspect of modern-day blogging which includes modern technologies being used, privacy rules and control as well as data analytics. By these approaches, we showed how blogging is not only the best medium for sharing information but can also help in other fields as well like marketing, education, data analysis, community development. With this research, we also wanted to highlight the downsides of using blogging or microblogging.
Author's Name: G. Vishnu Murthy, P. Tarun Kumar, M Ramu , D Lakshmi Charith
Abstract— With the rapid advancement of the Internet of Things (IoT), automated defect detection has become increasingly significant across various industries. This paper presents an IoT-based defect detection system for projectors, designed to improve efficiency and minimize errors associated with manual inspections. Leveraging real-time sensor data and machine learning algorithms, the proposed system detects potential defects such as overheating, misalignment, and lens malfunctions. The study examines different IoT-enabled approaches, evaluates key performance metrics, and provides experimental results that demonstrate the system's effectiveness.
Author's Name: Shubham Patil, Saurav Daware, Ameya Bhagat
Abstract— This paper is aimed at developing an Online College Management System that is of importance to the educational institute or college. This system is named College ERP using MERN stack. This system may be used to monitor college students and their various activities. This application is being developed for an engineering college to maintain and facilitate ease of access to information. For this the users must be registered with the system. College ERP is an Internet based application that aims at providing information to all levels of management within an organization. This system is used as an information management system for the college. For a given student and staff (technical and nontechnical) can access the system to either upload and access some information from the database.
Author's Name: Tulshiram Bansod , Monali Shivram
Abstract— With the quick rise in living standards, people's shopping passion grew, and their desire for clothing grew as well. A growing number of people are interested in fashion these days. However, when confronted with a large number of garments, consumers are forced to try them on multiple times, which takes time and energy. As a result of the suggested Fashion Recommendation System, a variety of online fashion businesses and web applications allow buyers to view collages of stylish items that look nice together. Clients and sellers benefit from such recommendations. On the one hand, customers can make smarter shopping decisions and discover new articles of clothes that complement one other. Complex outfit recommendations, on the other hand, assist vendors in selling more products, which has an impact on their business. FashionNet is made up of two parts: a feature network for extracting features and a matching network for calculating compatibility. A deep convolutional network is used to achieve the former. For the latter, a multi-layer completely connected network topology is used. For FashionNet, you must create and compare three different architectures. To achieve individualised recommendations, a two-stage training technique was created.
Author's Name: G.SrikanthReddy, P.Preetham, K.CharankumarReddy
Abstract— This project focuses on developing a machine learning-based system to detect fake social media profiles using three algorithms: K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Random Forest. Fake profiles pose a significant threat to online platforms by spreading misinformation, engaging in fraud, and manipulating public opinion. The study aimed to classify profiles as real or fake based on various features such as account age, activity level, and friends count. After training and testing the models, Random Forest achieved the highest accuracy at 93%, followed by SVM at 90% and KNN at 86%. The results indicate that Random Forest is the most effective algorithm for this classification task due to its ability to handle complex data and avoid overfitting. This system has potential applications in improving online security and could be further enhanced with larger datasets and real-time detection capabilities in future research.
Author's Name: Madar Bandu, R.Poojitha
Abstract— Business intelligence is a computer technology used to identify, obtain and analyze business data for example revenue, products, costs, sales revenue etc. However, this type of trajectory creates closed data that requires significant time to get an idea of everything the user needs. Despite its shortcomings, Business Intelligence (BI) has been able to generate fast data that could take a long time to research, making it an ideal tool for tracking emerging markets and markets in the community. Business Intelligence (BI) is an umbrella term that combines software with the development of Information systems and makes available the required information for any company. Most agree that BI works by capturing, storing, understanding and analyzing unprocessed data and making information accessible to it to improve business performance.
Author's Name: Aditi Shedge , Shaily Shah
Abstract— Air pollution has become a major issue in large cities because increasing traffic, industrialization and it becomes more difficult to manage due to its hazardous effects on the human health and many air pollution-triggering factors. This paper puts forth a machine learning approach to evaluate the accuracy and potential of such mobile generated information for prediction of air pollution. Temperature, wind, humidity play a vital role in influencing the pollution dispersion and accumulation, majorly influencing the prediction of pollution levels. Thus, this paper includes the atmospheric condition information registered throughout the study period in order to understand the influence of these factors on air pollution monitoring. Data driven modelling is an efficient way of extracting valuable information from generated data sets, however it is less efficient when the data is incomplete or contains inaccuracies. This modelling approach has true potential for real time operations because it can detect non-linear spatial relationships between sensing units and could aggregate results for regional investigation. Neural networks comparatively showed good capability in air quality prediction than support vector regression.
Author's Name: K.Jayendra Kumar, B.Krishna kanth, R.Akshay Kumar, B.Sharanya
Abstract—In the Urban mobility is increasingly hindered by traffic congestion, which results in inefficiencies, environmental damage, and economic losses. This study proposes an intelligent traffic control system based on cloud computing, big data analytics, and machine learning that maximizes traffic flow and urban mobility. By gathering and processing real-time traffic information from heterogeneous sources, including sensors, GPS units, and social media, the system utilizes Random Forest algorithms for predictive modeling. The system provides realtime traffic control measures, such as adaptive signal timing and route guidance, which react to conditions in real time. Scalability and effective management of data are provided through the use of a cloud-based system. Real-time inputs and observation are made easier through a friendly interface, and predictive analysis and testing are enabled through vehicle simulation. This integrated solution has the potential to minimize congestion, shorten travel times, and enhance safety overall, and offer a scalable and flexible urban traffic management solution.
Author's Name: Santhosh, Sai Laxmi
Abstract— Identifying of the plant diseases is essential in prevention of yield and volume losses in agriculture Product. Studies of plant diseases mean studies of visually observable patterns on the plant. Health surveillance and detecting diseases in plants is essential for sustainable development agriculture. It is very difficult to monitor plant diseases manually. It requires a lot of experiences in work, expertise in these field plant diseases and also requires excessive processing time. Therefore; image processing is used to detect plant diseases. Disease detection includes steps such as acquisition, image Pre-processing, image segmentation, feature extraction and Classification. We describe these methods for the detection of plant diseases on the basis of their leaf images; automatic detection of plant disease is done by the image processing and machine learning. The different leaf images of plant disease are collected and feature extracted of the various machine learning methods
Author's Name: Ayana Gopal and P. Rameswara Anand
Abstract— During the COVID-19 pandemic, numerous research studies focused on sentiment analysis of fake news on Twitter (now X). These studies revealed various patterns and themes in the tweets. The studies reviewed utilized diverse methods such as social network analysis, NLP, and mixed-methods approaches to uncover patterns of COVID-19 misinformation on Twitter. The studies reviewed utilized diverse methods such as social network analysis, NLP, and mixed-methods approaches to uncover patterns of COVID-19 misinformation on Twitter. By analyzing these headlines, we strive to understand the trends and characteristics of negative fake news and how similar the various fake news is. The experiment result shows better performance over the state-of-art algorithm in terms of accuracy, precision, recall and F1 score.
Author's Name: Abhay raj Singh, Suryakala Praveen
Abstract— Abstract- In this paper analyze the performance of Wireless Local Area Networks (WLANs), it is important to identify what types of network settings can cause bad performance. Low throughput, high packet loss rate, delayed round trip time (RTT) for packets, increased retransmissions, and increased collisions are the main attributes to look for when analyzing poor network performance. We use the OPNET Modeler to simulate the RTS/CTS mechanism to evaluate the performance of IEEE 802.11 MAC protocol. We have simulated two scenarios with and without RTS/CTS mechanism enabled on network nodes. We have concluded our findings by comparing the total WLAN retransmissions, data traffic sent/received, WLAN Delay of two scenarios. RTS/CTS mechanism is helpful to reduce the number of retransmissions if hidden node problem persists in network scenarios.
Author's Name: P.Swarajya Lakshmi, Shaistha Muskan, Mathangi Anusha, Akireddy Abhinav
Abstract—The primary purpose of image captioning is to generate a caption for an image. Image captioning needs to identify objects in image, actions, their relationship and some silent feature that may be missing in the image. After identification the next step is to generate a most relevant and brief description for the image that must be syntactically and semantically correct. It uses both computer vision concepts for identification of objects and natural language processing methods for description. It’s difficult for a machine to imitate human brain ability however research in this field have shown a great achievement. Deep learning techniques are enough capable to handle such problems using CNN and LSTM. It can be used in many intelligent control systems and IOT based devices. In this survey paper, we are presenting different approaches of image captioning such as retrieval based, template based, and deep learning based as well as different evaluation techniques.
Author's Name: Kala Selvi, Ramanathan, Premkumar
Abstract— This work gives a mechanism for doing authentication and authorization between managed element and server from a single database using a Centralized controller which can control a multiple switches. This work allows having one or more authentication servers for the switches to authenticate against which centralizes the authentication databases, making it easier to manage switch. Moreover, switch continues to support the pre-existing local authentication which works as a fallback in case of loss of connectivity to authentication server. Command authorization on per user basis is added which makes possible to have authorization of user to execute specific commands. Old access level authorization is continued to support as well. Protocol client is added and integrated into the existing system. As a part of this this work Remote authentication is supported meaning that authentication has not to be done by each switch by its own. Authentication database is shared with each other by switches now. Therefore each switch need not to be configured individually for a specific user and password in the network which will make the process of adding/modifying users very fast as opposed to time consuming in a large switch network and it is no more a security concern also. Chances of misconfiguration and mismatch are minimized. Keywords: AAA, API, Authentication, Authorization, C, Database, Ftp, NAS, Session, Switch, Telnet, SSH.
