In this paper, we present an adaptive approach in order to match and retrieve near duplicate images at different scales. Matching only local Features does not necessarily identify visually similar images. Global features are fast at matching but less accurate. Many existing methods either use local features or CNN features for image or video retrieval task. In this paper, we combined the use of SURF local points and CNN features extracted around SURF points in order to match near duplicate image pairs. Image pairs are segmented into blocks and CNN features of the image block containing matched SURF features are extracted and matched. Regions around matched image blocks are grown adaptively and matching is carried out until CNN mismatch is observed. To verify our proposed approach, experiments are carried out on benchmarking California-ND and Holiday dataset. Compared to traditional approaches for image retrieval, our approach not only retrieves relevant images but also provides detail of localized matched patch. For California-ND dataset and Holiday dataset, we achieve remarkable mAP (mean average precision) score up to 0.86 and 0.74 respectively.

No abstract

like commodity, crude just to name a few [12]. Unlike other Abstract-Stock market analysis and prediction has been one technical indicators, which try to find statistical relation of of the widely studied and most interesting time series analysis current price with future price, candlestick indicator tries to problems till date. Many researchers have employed many find the investor's sentiment on a given stock. The idea is different models, some of them are linear statistic based while simple "what will be the price of a stock tomorrow depends some non linear regression, rule, ANN, GA and Fuzzy logic on what people think about that stock at the end of the day". based. In this paper we have proposed a novel model that tries to I o .. v predict short term price fluctuation, using Candlestick Analysis. Iode tO indenif as a pattern s, we have This is a proven technique used for short term prediction of stock used SOM and CBR as a pattern recognizer and pattern price fluctuation and market timing since many years. Our matching mechanism. Decision derived by the CBR (SOM) is approach has been hybrid that combines Self Organizing Map in turn combined with other technical indicator like with Case Based Reasoning to indentify profitable patterns Stochastics and Volume, in order to improve the accuracy of (candlestick) and predicting stock price fluctuation based on the predication. pattern consequences. All available models for predicting stock price behavior can be divided into three category 1) Machine Learning based Index Terms-Candlestick Analysis, Short Term Stock (Black box), 2) Expert knowledge based (White box) and 3) Prediction, Self Organizing Map, Case-Base Reasoning Combination of both (Hybrid). In Black box models [3] [5] the system learns itself from previous experiences. Such a model does not require an expert, but the representation of the knowledge is complex and not verifiable (like ANN) where as S tock data analysis has remain one of the challenging time in White box models [2] [4] [6] the representation of the series analysis problem over the years. Because of the knowledge is simple and verifiable but requires expert to complexity and intrications, the problem has received impart knowledge in the system (like Fuzzy rule, CBR). attention from many researchers. Owing to its high expected System is only as good as the expert's knowledge. Hybrid returns, there are many starting from local investor to a big models [7] [8] try to combine the benefits of both. In this paper finance houses who are interested in finding out how the stock we have proposed a hybrid model which combines SOM with market works? Many techniques have been derived over a CBR. period of time by many researchers and investors to predict

The performance of various classification algorithms greatly depends on the characteristics of the data to be classified. There is no single classifier that works best on all given problems. The purpose of this study is to develop the computer vision based cashew grading system in conjunction with most accurate classification technique. The performance of different classification techniques including Multi-Layer Perceptron, Naive Bayes, K-Nearest Neighbor, Decision tree, Support Vector Machine are evaluated using WEKA toolbox to have most suitable classification technique for the cashew grading system. Subsequently, the classification technique that has the potential to significantly improve the performance of the system is suggested to be utilized in cashew grading system.

Social network is one of the most important complex networks, which aims to describe the interactive relationship among a group of active actors that represent different kind of structure. Many systems in the real world such as human societies and different types of components can be modeled as social networks. We can represent such a network in terms of graphical community. Social Network Analysis provides inherent research due to success of social media sites and social content sharing facility. Social Network Analysis provides key terms to provide platform for industry to generate survey of product and facilitate to introduce new innovation ideas to public entity. Now a day, as increase the use of social media sites provide the entrepreneurs and user to define new concept of community creation that represents the relationship of users that might be interested in same kind of activity. To create such communities introduce new research area for researcher. This community detection is different from traditional clustering. In This paper, we propose new algorithm for community detection in social network to get some meaningful and important information.