Classification of fungal and bacterial leaf diseases using machine learning techniques
In Farming and Gardening, leaf diseases have grown to be a challenge as it can cause considerable reduction in both quality and measure of agricultural yields. Thus, automated recognition of diseases on leaves plays a vital role in Farming and Gardening sector. This Thesis imparts a simple and computationally dexterous method used for leaf disease identification and grading using digital image processing and machine learning method. The proposed system is divided into two phases, in the first phase the plant is recognized on the basis of the features of leaf, it includes pre-processing of leaf images, and feature extraction followed by Support vector machine and decision tree based training. In the second phase the disease present in the leaf is classified, this process includes K-Means based segmentation of defected area, feature extraction of the defected portion and the Support vector machine and decision tree based classification of disease. Then the disease grading is done on the basis of the amount of disease present in the leaf.
Published by: Naveen Mandal, Yogesh Rathore
Author: Naveen Mandal
Paper ID: V4I3-1992
Paper Status: published
Published: June 27, 2018
Sensitive label privacy protection on social network data
Privacy is one of the fundamental issues when publishing or sharing social community data for social technology studies and business evaluation. Lately, researchers have developed privacy models much like okay-anonymity to save you node re-identification via shape data. However, even if those privacy fashions are enforced, an attacker may still have the ability to deduce one’s private statistics if a group of nodes in large part share the same touchy labels (i.e., attributes). In other words, the label-node courting isn't always nicely covered by pure structure anonymization methods. Moreover, current strategies, which depend upon area enhancing or node clustering, may also significantly alter key graph properties. Items shared through Social Media may affect more than one user’s privacy e.g., photos that depict multiple users comments that mention multiple users, events in which multiple users are invited, etc. The shortage of multi-celebration privateness management guide in modern mainstream Social Media infrastructures makes users unable to as it should be manipulated to whom those objects are absolutely shared or no longer. Computational mechanisms which can be capable of merge the privacy preferences of more than one customer’s right into an unmarried policy for an item can assist resolve this problem. But, merging more than one customers’ privateness preferences isn't always a smooth venture, because privateness options may war, so techniques to clear up conflicts are needed. To tackle this problem, in this, we propose the first computational mechanism to resolve conflicts for multi-party privacy management in Social Media with privacy policy inference of user-uploaded images that is able to adapt to different situations by modeling the concessions that users make to reach a solution to the conflicts.
Published by: Pradnya Sangade, Nitin Shivale
Author: Pradnya Sangade
Paper ID: V4I3-1984
Paper Status: published
Published: June 26, 2018
An experiment of improvement in solar panel efficiency using solar concentration by number of mirrors
Energy demand in India increasing as population growth increasing but it is difficult to fulfill energy demand as the resources are limited. India has a significant capacity for solar energy generation from renewable energy sources - solar photovoltaic, solar thermal, small water, wind, biomass, bio fuels and others. While proven technologies have emerged over the last few decades, the economical practicality of renewable energy is yet to be achieved and requires creation and adoption of innovative business and financial models. India’s energy problems include: Over 40 per cent of the households in India still don’t have electricity. One third of our total energy availability is from non-commercial sources (organic manure and fertilizer). Per capita energy consumption is one third of the world's average and per capita electricity consumption. India faces a supply gap of about 11% shortage and 14% of peak power shortage supply demand. Solar energy is at forefront of clean, renewable energy, and this pace is increasing due to the progress in solar panels production and efficiency as well as rapid unstable fuel costs. Photovoltaic (PV) solar cells are the most easily achievable solar technology, and they work best in bright days, in which the event does not have little or an obstacle in the sunlight. Photovoltaic (PV) technologies convert energy into electricity directly from the sunlight. Sunlight semiconductor strikes the substance and produces electrons, releasing electrons from their nuclear bonds. PV panels have no moving parts and usually live with less maintenance for twenty years or more. Landlords can set up PV panels to reduce or eliminate their monthly electricity bill, and utilities can create large "farms" of PV panels to provide pollution-free electricity to their customers. When the sun is at its highest point in the sky, then the power generation can offset the costly electricity generated when daily demand is greatest. Concentrated photovoltaic technology (CPV) uses optics such as mirrors and lenses to focus on sunlight on solar cells to generate electricity. CPV benefits on photovoltaic concentration because it is the number of solar cells required for the same power generation. With the duration and intensity of the sunlight, the temperature also puts a big impact on the performance of the PV module because high temperature reduces production efficiency. This experiment tells a practical approach to increase the efficiency of the solar panel by the number of uses of the mirror mechanism. These reflectors are cheap, easy to use, quite simple to use and no additional equipment or device is required to use. But experimental results show a remarkable increase in the overall production of solar panel. Experimental readings received from (a) without reflectors, (b) with one reflector, (c) with two reflectors and (d) with three reflectors.) with numbers of reflectors are compared.
Published by: Alok Ranjan, Om Prakash Tiwari, Anil Yadav
Author: Alok Ranjan
Paper ID: V4I3-1982
Paper Status: published
Published: June 26, 2018
A survey paper on privacy of medical data storage on the cloud
The proliferation of smartphones and use of reliable cloud technologies have given rise to various cloud-based applications. Remote disease prediction based on real-time medical data is one such application that has become very popular among research communities. One of the challenges involved here is the privacy of medical data storage on the cloud. Encrypted data requires decryption before it can be used by machine learning algorithms for disease prediction. This paper presents an analysis of research methods proposed to provide privacy of medical data on the cloud and their limitations.
Published by: Roomana Hasan, Dr. Kailash Shaw
Author: Roomana Hasan
Paper ID: V4I3-1977
Paper Status: published
Published: June 26, 2018
Long term load forecasting using K-Means clustering and ANN approach
This study implements the Kmean+ANN based Load forecasting technique to predict the load of Amritsar and Pathankot station. Artificial Neural Network (ANN) is one of the emerging methods used for forecasting the load. This method shown good results in different power station problems included planning, protection, designing, control & security analysis and fault diagnosing. In this paper, it is illustrated that from last few years the load forecasting has been widely adopted and this is due to increase in the demand for electric power and this had also resulted in an increase of generating sources expenditure. In most of the cases, it had been implemented in utilizes to determine the number of resources required to fulfill the demands of the project. The major focus area of the research was load prediction of two different stations for 9 upcoming years. In this study work past 18 years load information was collected and different years were select as base years and after that, the author had implemented the k-mean clustering and ANN method to predict load for upcoming nine years.
Published by: Shelly Bhanot, Shivani Mehta
Author: Shelly Bhanot
Paper ID: V4I3-1980
Paper Status: published
Published: June 25, 2018
Mobile web based android application for department management
Mobile applications are too handy and easy to use. They are used as an option to the traditional process. Applications developed have users which through proper authentication can access the system. Teacher and students both have their own apps and have username and password to log in and so anyone cannot log in. Students can view all the details like attendance percentage, defaulter list, internal marks, practical submitted using this app interface. Teachers can take attendance, enter marks, keep track of each student who has submitted the practicals, generate automatic defaulter and critical defaulter list. The project also adds mobility and automation of the information.
Published by: Santosh Patkar, Sunil Sahu, Shekar Singh
Author: Santosh Patkar
Paper ID: V4I3-1963
Paper Status: published
Published: June 25, 2018
