MIT-World Peace University (MIT-WPU) is a top educational institute located in Pune, India. It is recognized within&nbsp;UGC under Government of Maharashtra Act No. XXXV 2017.&nbsp;

It is part of the MIT group of institutions (established in 1983) which covers 10+ campuses across India and has more than 50,000+ students enrolled across various disciplines at a given point of time. The institute houses 15 schools ranging from Engineering &amp; Technology, Design, Pharmacy, Commerce, Governance among others. It offers undergraduate, postgraduate and PhD programs among different disciplines.&nbsp;

&nbsp;

MIT World Peace University

School of Computer Engineering & Technology

Search engine deals with the user provided query to deliver the informative results to users. The volume of data associated with these search engines is very vast and it becomes very difficult to handle just data. These data are dynamic, increase day by day and hence many techniques have been proposed to handle such dynamic data. Existing Tree-based techniques are mostly applicable to queries such as spatial queries, range queries and many more. These techniques are useful only for the queries that have coordinates. But these techniques are not applicable to the queries which do not have coordinates. Keyword-based search has been considered more helpful in many applications and tools. The paper considers objects, i.e., images, tagged with number of keywords and those will be embedded into vector space for evaluation. The main aim here is to achieve higher scalability with the increasing data and speedup of retrieval of results for the users. A query is been implemented in the paper called as nearest keyword set query that deals with multi-dimensional datasets. Hash-based index structure is implemented along with a new algorithm called as ProMiSH, i.e., Projection and Multi-scale Hashing.

World University root node

Engineering and Technology

Lecture Notes in Networks and Systems

Efficient Retrieval from Multi-dimensional Dataset Based on Nearest Keyword

Dynamic texture (DT) is temporal extension of static texture. There are two broad types of dynamic textures: natural and manmade. The examples of dynamic textures are waving tree, sea water, fountain, traffic, moving crowd, etc. Dynamic texture segmentation is a technique used to separate the moving objects from stationary content in the video. Majority of the techniques give good results either for natural or for manmade dynamic textures. The proposed approach for DT Segmentation is combination of local spatiotemporal technique, i.e., local binary pattern-Weber local descriptor and global spatiotemporal technique, i.e., contourlet transform. The local spatiotemporal technique considers the appearance and motion of the object for segmentation. The technique is computationally less complex than optical flow and gives good results for manmade dynamic textures. The global spatiotemporal technique is based on Laplacian pyramid and directional filters. It gives good results for natural dynamic textures. The proposed technique discussed in this paper is unified approach for any type of dynamic texture. The regions of images commonly segmented by both the techniques are considered as the final segmented output. The proposed system works equally well on any kind of dynamic texture. © 2019 Inderscience Enterprises Ltd.

International Journal of Reasoning-based Intelligent Systems

Unified dynamic texture segmentation system based on local and global spatiotemporal techniques

Articular cartilage (AC) of knee plays a vital role in the supporting mechanism of the human body. AC is a smooth, slippery surface which composed of white tissue covering the joint shell. Magnetic Resonance (MR) imaging technique is widely adopted in clinical practice to visualize and detect the defect in Knee anatomical structure. Segmentation of AC using knee MR images is an essential process to biomarker analysis during knee osteoarthritis progression. Hence it has been widely investigated by research communities. Osteoarthritis (OA) is a commonly found knee disease, in which AC is slowly and progressively lost. Early diagnosis of OA plays a significant role to look up treatment possibilities and prevent further degradation of knee AC. Segmentation methods of AC have considered various parameters of image, image quality, and tissue structure, etc. Manual segmentation of cartilage for OA diagnosis, from a large amount of MRI images produced in clinical practice, is a time-consuming and challenging process. Thus semi-automated and fully automated cartilage segmentation methods have been adopted by many researchers to provide reasonable accuracy and reliable detection of AC from knee MR images. This paper reviews knee cartilage segmentation methods from MR images. In this paper, we focus on the recent tread of knee AC segmentation techniques from MR images. First, an introduction to osteoarthritis and its causes are given. Then variety of MR imaging modalities suitable for assessment of knee OA. Finally, performance measures are explained followed by conclusion.

International Journal of Advanced Science and Technology

A Comprehensive Survey of Articular Cartilage Segmentation Methods on Knee MRI

The abundant availability of food is extremely essential for human existence on this planet, thus making agriculture a mandatory practice throughout the world. Within two decades, a severe complication of food will arise. Hence, the advancement and progress of agriculture is essential. With the help of trending technologies like IoT, the environmental factors that affect the crops are easily monitored, and decisions are taken based on them so that crop monitoring is done remotely, which is a magnificent IoT application. Today, agriculture in the world is partially automated, and farmers mostly rely on traditional methods because of which they suffer from scarcity of rainfall and dearth of water. A plethora of techniques exist that provide automation of agricultural activities and hence increase their efficiency so that better crop output is obtained. The automation of agriculture enables reduction in human involvement. The main purpose of this review paper is to present an overview of various techniques adopted in automated agriculture systems which save the farmers' time, money and power. This paper also extracts key features of existing automated agriculture systems based on their advantages and disadvantages. This paper will assist in the development of a system which will overcome the drawbacks of existing automated agriculture systems. © 2019 IEEE.

2019 IEEE Pune Section International Conference, PuneCon 2019

Automated Systems for Smart Agriculture

In an era where customer relationship management plays an important role in business success, firms endeavour to develop profitable associations with their existing and future consumer base. With customer-driven industries becoming highly competitive, there is a huge demand for efficient business intelligence (BI) approaches to facilitate high performance results while optimising on the costs invested, mainly in marketing activities. Therefore, researchers around the world are exploring BI techniques to maximise customer satisfaction and customer loyalty. Recently, machine learning models are being adopted to guide businesses for investing on new marketing strategies through intelligent mining of their existing customer data. This paper adopts a neural network approach to analyse and develop a model of business rules establishing relationships that can exist between customer demographics and the final purchase made by the customers. Then this model can be employed for the prediction of the purchasing patterns that might occur in the future. The dataset obtained from Kaggle's 'Allstate Purchase Prediction challenge' consisting of a car insurance customer base is employed for developing our prediction model. The experimental results confirm that the neural network approach adopted is able to extract the purchasing rules exhibited by the customers of a firm by discovering the customer behaviour patterns and their relationships with different demographic variables. An efficient prediction of the customer purchasing patterns would help organisations to make appropriate business proposals to their prospective customers in order to achieve higher customer satisfaction. © 2019 IEEE.

Prediction Model Development using Neural Network Approach

Effective human machine interaction systems areneed of the time so the work carried out deals with one of suchsignificant HMI tasks- automatic emotion recognition. Theexperimentation carried out for this study is focused to facialexpressions based emotion recognition. Two techniques ofemotion recognition based on hybrid features are designed andexperimented using JAFFE database. The first techniquereferred as "Hybrid Method1" is designed around featuredescriptor obtained through local directional number &amp;principal component analysis and feed forward neural networkused as classifier. The second technique referred as "HybridMethod 2" is designed around feature descriptor obtainedthrough histogram of oriented gradients, local binary patternand Gabor filters. PCA- principal component analysis is used fordimensionality reduction of feature descriptor and k-nearestneighbors as classifier. The average emotion recognitionaccuracy achieved through method 1 and method 2 is 85.24%and 93.86% respectively. Effectiveness of both the techniques iscompared on the basis of performance parameters such asaccuracy, false positive rate, false negative rate and emotionrecognition time. Emotion recognition has wide applicationareas so the work carried out can be applied for suitableapplication development.

Fulltext

International Journal of Recent Technology and Engineering (IJRTE)

Facial Features Based Hybrid Methods for Emotion Recognition

eLearning is the use of technology in providing education via various online resources. It has the potential to change the whole education system. Recommendation Systems are schemes that aim to provide the user with the most appropriate and precise products by filtering helpful data from a wide information base. A recommender system in the eLearning environment is a software agent that tries to intelligently recommend actions to a learner by various methods. This paper is a compilation of various techniques that are used by recommender systems in eLearning platforms. Also it focuses on deep learning techniques to enhance eLearning Recommendation system.

Recommendation System for Enhancing eLearning using Deep Learning

Blind people have been using a simple white foldable stick as an aid for their daily activities since many years. The idea in this project is to develop an intelligent solution - A Smart Blind Stick, an IoT-based project that has Ultrasonic Sensors using which the distance of any obstacle can be measured and provides output in the form of buzzers. RF transmitter and receiver will be used to locate the stick. Additionally, GPS navigation will be provided for the complete navigation and guidance of the blind person. Live location tracking of the blind will be facilitated to the nearest help centres or their relatives in case of any emergency. Face detection

Recent Trends in Information Technology and its Application

Smart Stick for Blind: A Survey

Personality is an important parameter as it differentiates various individuals from one another. Personality prediction is an evergreen area of research. Predicting personality with the help of data through social media is a promising approach as this method does not require any questionnaires to be filled by users thus reducing time and increasing credibility. Thus having knowledge of personality is an interesting domain for researchers to work on. Predicting personality has many applications in real world. Use of social media is increasing day by day. Huge amount of textual data as well as images continue to explode to the web daily. Current work focuses on Linear Discriminate Analysis, Multinomial Naive Bayes and AdaBoost over Twitter standard dataset. © 2019 IEEE.

2019 4th International Conference on Information Systems and Computer Networks, ISCON 2019

Using textual data for Personality Prediction:A Machine Learning Approach

Use of social media is increasing day by day. A huge amount of textual data, as well as images, continue to explode to the web daily. Personality prediction is an evergreen area of research. Predicting personality with the help of data through social media is a promising approach as this method does not require any questionnaires to be filled by users thus reducing time and increasing credibility. In the current work we have fetched real-time Twitter data through a standard API and thereby created a sample experimental dataset. We have employed several popular machine learning algorithms such as kNN, Naive Bayes and SVM. It was observed that Naive Bayes yielded the highest accuracy of 71.67% as compared to the other classifiers. © 2019 IEEE.