Abstract

Measuring inter-document similarity is one of the most essential steps in text document clustering. Traditional methods rely on representing text documents using the simple Bag-of-Words (BOW) model. A document is an organized structure consisting of various text segments or passages. Such single term analysis of the text treats whole document as a single semantic unit and thus, ignores other semantic units like sentences, passages etc. In this paper, we attempt to take advantage of underlying subtopic structure of text documents and investigate whether clustering of text documents can be improved if text segments of two documents are utilized, while calculating similarity between them. We concentrate on examining effects of combining suggested inter-document similarities (based on inter-passage similarities) with traditional inter-document similarities following a simple approach for the same. Experimental results on standard data sets suggest improvement in clustering of text documents

Abstract

User authentication is necessary to secure the data and process on Internet and in digital devices. Static text based authentication are most widely employed authentication systems for being inexpensive and highly scalable. But they are prone to various types of active and passive attacks. The constant need of extending them to increase security is making them less usable. One promising alternative is Graphical authentication systems, which if implemented properly are more secure but have their own drawbacks. In this paper, we discuss in detail the extension of our previous work Passblot [18], a unique graphical authentication system. It generates pseudo random one time passwords using a set of inkblots, unique to each user. Properties of one time passwords ensure the resistance towards various common attacks and the uniqueness of human perception makes it usable. We demonstrate how our system effectively mitigates various attacks and analyse the results from various experiments conducted.

Abstract

In this paper, we explore the use of Bessel features derived from speech utterances, to develop Gaussian mixture speaker models for text independent Speaker Identification. The proposed approach to speaker identification is based on existing methods that employ Gaussian mixtures for the modeling of speakers. However, we have developed the speaker models from the Bessel features derived from the speech utterances, as an alternative to Mel-frequency cepstral coefficients for developing the speaker models. The proposed approach is tested on two databases of ten and twenty speakers respectively and their performance is evaluated. Finally, we have made some suggestions for future work involving the use of Bessel features for text independent speaker identification

Abstract

Edge detection is an important image processing technique used in face recognition. In this paper we propose an edge detection algorithm suitable for extracting edge maps from facial images under noisy conditions. We employ several improvised techniques like composite gradient variation measures, weighted wide convolution kernels and dual maxima detection. This helps in removing spurious edge points along stronger edges, extracting fine edge lines along facial boundaries and curvatures and suppressing noise. This is demonstrated by extracting edges from facial images degraded by Gaussian, speckle and salt and pepper noise. For performance measurement and evaluation Pratt's Figure of Merit and Signal to Noise statistics is used

Abstract

The current research focused on the issues related to human-bird conflicts and subsequent alternatives adopted by the migratory birds at Uppalapadu Lake (India). The study collates the information from field observations and literature surveys and draws conclusions based on the observation data. The study highlights the ways the local community, Forest Department, activist groups and other organizations participate in protecting the birds from harm. It also focuses on some of the management initiatives taken by the Forest Department that are helping the birds during their breeding periods. Overall, the research identifies the ecological importance of the lake with respect to birds and stresses the need for its protection and further expansion.

Abstract

In the present study, three large ecological plots of 3 ha each were surveyed in 3 different forest types of North Andaman Islands. Each 3-ha plot (30 sub-plots of 0.1 ha each) was classified into different classes based on site quality as either excellent, good, moderate or poor using an index developed by utilizing vegetation parameters such as species richness, diversity, density, among others. Analysis revealed most of the area to be under the good category in three forest communities, indicating that forests of North Andaman are potential sites of species richness and diversity. The “general limit of species assemblage” with respect to higher angiosperm taxa in North Andaman was observed based on two independent approaches of sampling: stratified random plots and the large area ecological plots. The general limit of species assemblage was in the range of 14-33 species. The present analysis provides a base …

Abstract

In this work, we study the problem of reporting and counting maximal points in a query rectangle for a set of n integer points that lie on an n×n grid. A point is said to be maximal inside a query rectangle if it is not dominated by any other point inside the query rectangle. Our model of computation is unit-cost RAM model with word size of O(logn) bits. For the reporting version of the problem, we present a data structure of size words and support querying in time where k is the size of the output. For the counting version, we present a data structure of size words which supports querying in . Both the data structures are static in nature. The reporting version of the problem has been studied in [1] and [5]. To the best of our knowledge, this is the first sub-logarithmic result for the reporting version and the first work for the counting version of the problem.

Abstract

The use of manycore architectures and accelerators, such as GPUs, with good programmability has allowed them to be deployed for vital computational work. The ability to use randomness in computation is known to help in several situations. For such computations to be made possible on a general purpose computer, a source of randomness, or in general a pseudo random generator (PRNG), is essential. However, most of the PRNGs currently available on GPUs suffer from some basic drawbacks that we highlight in this paper. It is of high interest therefore to develop a parallel, quality PRNG that also works in an on demand model. In this paper we investigate a CPU+GPU hybrid technique to create an efficient PRNG. The basic technique we apply is that of random walks on expander graphs. Unlike existing generators available in the GPU programming environment, our generator can produce random numbers on demand as opposed to a onetime generation. Our approach produces 0.07 GNumbers per second. The quality of our generator is tested with industry standard tests. We also demonstrate two applications of our PRNG. We apply our PRNG to design a list ranking algorithm which demonstrates the on-demand nature of the algorithm and a Monte Carlo simulation which shows the high quality of our generator.

Abstract

We consider the problem of reporting and counting maximal points in a given orthogonal query range in two-dimensions. Our model of computation is the pointer machine model. Let P be a static set of n points in ℝ2. A point is maximal in P if it is not dominated by any other point in P. We propose a linear space data structure that can support counting the number of maximal points inside a 3-sided orthogonal query rectangle unbounded on its right in O(logn) time. For counting the number of maximal points in a 4-sided orthogonal query rectangle, we propose an O(n logn) space data structure that can be constructed in O(n logn) time and queried upon in O(logn) time. This work proposes the first data structure for counting the number of maximal points in a query range. Das et al. proposed a data structure for the counting version in the word RAM model [WALCOM 2012]. For the corresponding reporting versions, we propose a linear size data structure for reporting maximal points inside a 3-sided query range in time O(logn + k), where k is the size of the output. We propose an O(n logn) size data structure for reporting the maximal points in a 4-sided orthogonal query range in time O(logn + k), where k is the size of the output. The methods we propose for reporting maximal points are simpler than previous methods and meet the best known bounds.

Abstract

This paper initiates formal analysis of a simple, distributed algorithm for community detection on networks. We analyze an algorithm that we calltextsc {Max-LPA}, both in terms of its convergence time and in terms of the" quality" of the communities detected.textsc {Max-LPA} is an instance of a class of community detection algorithms calledtextit {label propagation} algorithms. As far as we know, most analysis of label propagation algorithms thus far has been empirical in nature and in this paper we seek a theoretical understanding of label propagation algorithms. In our main result, we define a clustered version ofer random graphs with clusters where the probability , of an edge connecting nodes within a cluster is higher than , the probability of an edge connecting nodes in distinct clusters. We show that even with fairly general restrictions on and ( for any , , where is the number of nodes),textsc {Max-LPA} detects the clusters in just two rounds. Based on this and on empirical results, we conjecture thattextsc {Max-LPA} can correctly and quickly identify communities on clustereder graphs even when the clusters are much sparser, ie, with for some .