Statistical Analysis of Part of Speech (Pos) Tagging Algorithms for English Corpus
Part of speech (POS) Tagging is the procedure of allocating the portion of speech tag or supplementary philological class signal to every single and every single word in a sentence. In countless Usual Speech Processing presentations such as word intellect disambiguation, data recovery, data grasping, analyzing, interrogating, and contraption clarification, POS tagging is imitated as the one of the frank obligatory tool. Categorizing the uncertainties in speech philological items is the mystifying goal in the procedure of growing an effectual and correct POS Tagger. n this paper we difference the presentation of a insufficient POS tagging methods for Bangla speech, e.g. statistical way (n-gram, HMM) and perception established approach. A supervised POS tagging way needs a colossal number of annotated training corpuses to tag properly. At this early period of POS-tagging for English. In this work we craft an earth truth set that encompasses tagged words from sampled corpus. We additionally investigated the presentation of POS taggers for disparate kinds of words.
Published by: Swati Tyagi, Gouri Shankar Mishra
Author: Swati Tyagi
Paper ID: V2I3-1157
Paper Status: published
Published: June 3, 2016
Hybrid Call Security System using Encryption & Steganography
— Most of the users use internet for various voice or video calling applications. Also many companies utilize these applications for their corporate calls (inbound/outbound business calls) with the users outside of their network. To achieve the goal of voice communication security, a number of audio security and audio processing algorithms are in use individually or in a combination to provide the effective voice security. Hacking attacks on these applications can cause great losses to the user security which can lower the number of active users and so the business popularity. In the proposed voice call security model, we have proposed a hybrid approach using compression, encryption and steganography to enable to highest level of security in the voice calling while adding the minimum possible delay in the voice packets delivery. Our proposed framework focuses on the security of voice communications, which can take place over a wired phone, cellular connection or internet. The proposed framework consists of three major components to secure the voice communications over internet or intranet i.e. band pass filter, cryptography or steganography. The voice signal would be decomposed using band pass filters, followed by application of cryptography on all of the bands. All of the signal components after decomposition and encryption would be combined and placed in another voice signal using steganography method. This framework have designed to provide multi-layer security to the sensitive voice calling channels between the VVIP, VIP and other important personal of the nation. The results have proved the effectiveness of the system. The system had been tested for its security level, possibilities of breaching attacks, accuracy, noise (WGN) reduction, compression levels, encryption levels, elapsed time and many other aspects. The proposed framework has proved to be effective in all situations related to the voice communication security.
Published by: Manmeet Kaur, Samreen Sekhon Brar, Namrata Chopra
Author: Manmeet Kaur
Paper ID: V2I3-1156
Paper Status: published
Published: May 31, 2016
Smart Routing for WSN for the Energy Balanced Routing over Hierarchical Deployments
The data aggregation is the method of combining the data streams coming from multiple nodes. The data aggregation methods reduce the routing decision cost implied by the routers or nodes in the path. The data aggregation can be proven to be efficient in the terms of energy efficiency or transmission efficiency. The proposed model is entirely based upon the data aggregation and data forwarding technique. In this research project, we have tried to develop the new era aggregation model for the wireless sensor networks. The proposed model is intended to enhance the energy efficiency by improving the aggregation model for the sensor networks to enable the efficient delivery mechanism for the data. The WSNs are the directed graph networks, where the data is being sent into the similar direction of base transceiver station (BTS). Because the data in the directed graph networks flow in a particular direction, it always made the aggregation method efficient. The proposed model aggregation model is based upon the smart amalgamation of heuristic and greedy algorithm based aggregation. The greedy allows the algorithm to cover any number of data streams for the aggregation, whereas heuristic is responsible for the group formation before going for the aggregation. The multiple aggregation groups are produced by using the heuristic approach in the proposed model. So in this thesis, we have proposed and implemented the smart heuristic and greedy based hybrid aggregation and data transmission algorithm. The proposed model has shown its effectiveness in the case of aggregation and data transmission. The proposed model has proved its efficiency in the form of network load, route persistence, energy consumption and latency. The proposed model has proved it better than the previous approaches on the basis of the above listed performance parameters.
Published by: Namrata Chopra, Manmeet Kaur
Author: Namrata Chopra
Paper ID: V2I3-1155
Paper Status: published
Published: May 31, 2016
Extracting News from the Web Pages by using Concept of Clustering with Neural Genetic Approach
Web news extraction is an investigation area which has been widely discovered. It has resulted in some systems which has good extraction abilities with little or no human involvement. The present system looks into the perception of web newscast from a lone web site which takes a parallel format and the idea commonly is not as efficient when multiple web news pages are considered which belong to different sites. My work proposes a web extraction layout which is rather same for most of the web news The purpose of web news extraction is to enhance information retrieval which stores news articles related to a particular event for competitive business analysis Researches in this area have shown many approaches different from the other based on the need, the extractor should be chosen.
Published by: Nishan Singh Saklani, Saurabh Sharma
Author: Nishan Singh Saklani
Paper ID: V2I3-1154
Paper Status: published
Published: May 30, 2016
Characterization of Stainless Steel 316l Coated by Thermal Spray Coating
An experimental investigation was carried out to find the characteristic of coated stainless steel 316L, coating was done by D-gun method. Surface properties corrosion is check out of both coated base metals and uncoated base metals. Two specimens of coating were prepared with coating thickness of 150 microns (single layer) and 300 microns (multilayer). Pitting corrosion test was performed to analysis the corrosion properties of base metal and coated metals, corrosion test was conducted for 96 hours and solution used was stagnant seawater composition and temperature of solution was at a temperature of 420C. Characterization of Inconel 718 and stainless steel 316L was done by using XRD. Result obtained from corrosion test showed that weight loss of base metal SS 316L was 0.0002 gram which 50% more than the weight loss in multilayer coated specimens that 0.0001 gram.
Published by: Prajapati Amit Kumar, Sudhanshu Pandey, Ram Mishra, Ramdoot Yadav
Author: Prajapati Amit Kumar
Paper ID: V2I3-1153
Paper Status: published
Published: May 26, 2016
Blur Detection using Hybrid Classifier
Popular entertainment and communication services of internet or mobile applications is multimedia content such as image, audio and video that may suffer from low quality problem. Blur is the one of the factors that degrades the quality of image or frames in video. Enhancement or restoration of blurred image requires detection of blurred region or kernel. Therefore, blur detection is the initial and main step of blur phenomena followed by blur classification and restoration process. In this paper, we presented overview on a few defocus and motion blur detection methods with their applications. Some of this methods based on features of blurred kernel while others not. These methods can be either direct or indirect. Direct methods only identify the blurred region and segment it from un-blurred one. While indirect methods first detect and then restore the blurred region. We discussed both type of blur detection methods.
Published by: Indu Kharb, Abhishek Sharma
Author: Indu Kharb
Paper ID: V2I3-1152
Paper Status: published
Published: May 20, 2016
