FINDING COMMON LEAF DISEASES USING DEEP CONVOLUTION NEURAL NETWORK – A MACHINE LEARNING APPROACH

Chaitali R Shewale¹, Dr. Sunil Dambhare², Dr. Sandeep U. Kadam³

Assistant Professor, Keystone Scholl of Engineering¹
Professor, Dr. D. Y. Patil Institute of Engineering Management and Research, Pune, India²
Professor, Bhivarabai Swant of Engineering and Research, Pune, India³

Keywords - Maize leaf, CNN, ML, Diseases detection.

ABSTRACT

Plant ailments influence the development of their individual species; along these lines their initial recognizable proof is significant. Many Machine Learning (ML) models have been utilized for the location and arrangement of plant illnesses in any case, after the headways in a subset of ML, that is, Deep Learning (DL), this territory of research seems to have extraordinary potential as far as expanded precision. Many created/changed DL structures are executed alongside a few perception systems to recognize and order the side effects of plant ailments. In addition, a few exhibitionmeasurements are utilized for the assessment of these designs/strategies. For the mostpart, there are eight sorts of normal leaf maladies, including Curvularia leaf spot, overshadow mosaic, dark leaf spot, northern leaf scourge, and darker spot, round spot, rust, and southern leaf curse. Most truly, maize leaf malady is unsafe and will influence maize creation and individuals' lives.

Full PDF File

MAINTAINING AUTHENTICITY OF DIGITAL CERTIFICATE USING BLOCKCHAIN

Sandeep U Kadam¹, Sunil G. Dambhare², Chaitali R Shewale³, Rupesh J Patil⁴

Associate Professor, Bhivarabai Sawant College of Engineering and Research, Pune, India¹
Professor, Dr. D. Y. Patil Institute of Engineering Management and Research, Pune, India²
Assistant Professor, Keystone College of Engineering, Pune, India³
Principal Navsahyadri Grooup of Institutes Faculty of Engineering, Pune, India⁴

Keywords - Blockchain, digital certificate, encryption.

ABSTRACT

According to several researches huge number of graduates are passing out everyyear, the certificate issuing authorities are seems to be compromised for the security credentials of student data. Due to the lack of effective antiforge mechanism, graduation certificates which are copied often get noticed. We can conquer this problem by using digital certificate, though security issues still exist. Blockchain is one of the most recent technologies that can be adopted for the data security. It helps to overcome the problem ofcertificate forgery because of its unmodifiable property. Digital certificate is issued using following procedure. First from student portal and college portal encrypt the entered marks. Then pass this encrypted string to the blockchain. Company portal accepts the marks string from both the portal and passes it to the verification portal. Verification portal decides whether marks are authenticated or not. It will provide the demand unit to verify the genuineness of the paper certificate through mobile phone scanning or website inquiries. Because of the unmodifiable properties of the blockchain, the system not only enhances the authenticity of various paper based certificates, but also electronically reduces the loss risks of various types of certificates.

Full PDF File

STUDY OF MACHINE LEARNING CLASSIFIERS FOR SENTIMENT PREDICTION

Y. Sri Lalitha, Y. Gayathri, Althaf Hussain Basha Sk

^1,2Gokaraju Rangaraju Institute of Engineering and Technology, Hyderabad, Telangana, India
³Department of CSE, Chalapathi Institute of Engineering and Technology, Guntur, Andhra Pradesh, India

Keywords -

ABSTRACT

Product Review Analysis has developed into a crucial application for all businesses. This will give the company the chance to examine customer product reviews and learn what the market thinks of their goods. It necessitates a comprehensive computational analysis of the behaviour of discrete entities with regard to consumer purchasing similarity and the extraction of the customer's perspective on the business entity. Customer satisfaction is the constant yardstick by which corporate performance is judged. In this newly emerging era of e-commerce and social networking, the introduction of a new product requires a thorough examination of consumer opinions on current products and their needs in the product. Since so many reviews are being produced from different sources, it is becoming more and more challenging. The issue of categorizing reviews into positive and negative opinion is addressed in this study. The work presented here used Naive Bayes, Stochastic Gradient Decent, Random Forest, Multinomial, and Logistic Regression techniques to analyze the product reviews.

Full PDF File

A SEGMENT BASED SCHEME FOR SECURE DATA SHARING IN CLOUD

Abu Salim and Rajesh Kumar Tiwari

Department of Computer Science and Engineering, Glocal School of Technology and Computer Science. Glocal University, Saharanpur, U.P.

Keywords - Segmentation, cloud storage, symmetric key cryptography, asymmetric key cryptography, and proxy re-encryption. 

ABSTRACT

Cloud computing makes it possible to share data, which brings a plethora of advantages to the consumers of the platform. These advantages include extensive access to networks, resource polling, rapid elasticity, and measurable services. The security risk is increased when the data are relocated to an outside place on a CSP (cloud service provider) server. The key security risks associated can be classified in terms of confidentiality, integrity, and availability. Depends on the type of consumers that make use of cloud storage, data is divided into the various groups. In our plan, we separated the data into three categories: the first kind, in which the data is used by only one person, which is the owner; the second type, in which the data is accessed by a small number of consumers; and the third type, in which the data is utilized by a significant number of users. In order to accommodate these three categories of data, the storage capacity available in the cloud will be divided into three groups: S1, S2, and S3. S1 will be used only by the owner, whereas S2 and S3 will be used by a restricted and a big number of users respectively. The usage of a symmetric and asymmetric cryptographic technique in conjunction with proxy re-encryption has been implemented in order to provide secure access and protection for these three categories of data.

Full PDF File

EVALUATION OF CLASSIFICATION OF BRAIN IMAGES INTO  ALZHEIMER DISEASE AND NORMAL

SUMANTH S and A SURESH

Research Scholar, Periyar University, Salem, Tamilnadu, India

Principal, Maisurii Women’s College of Arts and Science, Kakapalayam, Salem, Tamilnadu, India

Keywords - Alzheimer’s disease (AD), Correlation based Feature Selection (CFS), Multi Layered Perceptron Neural Network (MLPNN) and fuzzy classifiers.

ABSTRACT

Alzheimer’s disease (AD) is the most common form of dementia in elderly people worldwide. Most existing pattern classification methods just use one individual modality of biomarkers for diagnosis of AD or MCI, which may affect the overall classification performance. Correlation based Feature Selection (CFS) is a simple filter algorithm that ranks feature subsets and discovers the merit of feature or subset of features according to a correlation based function. This work classifies the brain image as Alzheimer or normal. To classify whether the image is Alzheimer or normal Multi-Layered Perceptron Neural Network (MLPNN) and fuzzy classifiers are used. MLP have been evolved over the years as a very powerful technique for solving a wide variety of problems. Much progress has been made in improving performance of MLP and in understanding how these neural networks gets
operate.

Full PDF File

IMPLEMENTATION OF DATA MINING TECHNIQUES FOR EXTRACTION OF KNOWLEDGE MANAGEMENT

Yethiraj N G and Sumanth S

Assistant Professor, Department of Computer Science
Maharani’s Science College for Women, Palace Road, Bangalore -560001

Assistant Professor, Department of Computer Science
Smt. VHD Central Institute of Home Science, Seshadri Road, Bangalore-560001

Keywords -Data Warehouse, Data Mining (DM), Artificial intelligence (AI), Ontology, Knowledge Management (KM), and Information Technology (IT)

ABSTRACT

Knowledge Management systems benefit corporations that take advantage of the Artificial Intelligence technology. As enterprises are being driven toward KM systems to meet competitive pressures and create value, they are increasingly finding that these systems can facilitate reuse of existing knowledge and create new knowledge in an effort to allow better decision making process. In this paper we argue that Data mining and Data warehousing with allied AI concepts can make a significant contribution to knowledge management initiatives.

Full PDF File