Papers by Rajkumar Kannan
JOURNAL OF CLINICAL AND DIAGNOSTIC RESEARCH, 2018
The proceedings contain 12 papers. The topics discussed include: impact of covid-19 on education ... more The proceedings contain 12 papers. The topics discussed include: impact of covid-19 on education using twitter data;automating road junction identification using crowdsourcing and machine learning on gps transformed data;augmented and personalized digital narratives for cultural heritage under a tangible interface;issues and challenges associated with blockchain in smart cities;identifying dominant nodes in semantic taxonomies;correlation of biological and computer viruses through evolutionary game theory;and a system for collecting, managing, analyzing and sharing diverse, multi-faceted cultural heritage and tourism data.
Transactions on Machine Learning and Artificial Intelligence, 2018
Most of the health organizations provide an array of medical services and request their beneficia... more Most of the health organizations provide an array of medical services and request their beneficiaries to provide their experience’s in the form of opinion/reviews for which they are associated. Doctors of national and international repute have hundreds and even thousands of reviews authored by the health consumers around the globe. For an individual it is difficult and time consuming process to look all the reviews before taking an appropriate decision. Thus it is necessary to summarize the reviews to make an individual to take prompt decision. For a doctor it is also difficult to keep track of patient’s reviews given by the patients in different time intervals, but he may have the summary of his entire patient’s reviews to understand what is the best can be done to the patient’s community. This research paper aims to mine and summarize the medical reviews authored by the health consumers. This article is performed by summarization of text in three steps, the first step is to identi...
International Journal of Computer Science and Artificial Intelligence, 2012
An approach for extracting higher-level visual features for art painting classification based on ... more An approach for extracting higher-level visual features for art painting classification based on MPEG-7 descriptors was implemented in the system "Art Painting Image Colour Aesthetics and Semantics" (APICAS). The approach consists of the following steps: (1) tiling images into non-overlapping rectangles in order to capture more detailed local information; (2) the tiles of the images are clustered for each MPEG-7 descriptor; (3) vector quantization is used to assign a unique value to each tile, which corresponds to the number of the cluster where the tile belongs, in order to reduce the dimensionality of the data. The distribution of significance of the attributes, the importance of the underlying MPEG-7 descriptors as well as analysis of spatial granularity for class prediction in this domain are analyzed.
International Journal of Computer Applications, 2010
In the last three decades, the field of e-assessment has become more important. Several universit... more In the last three decades, the field of e-assessment has become more important. Several universities and higher education institutes have started to provide online assessments. The variance in the e-assessment application domains as well as the market competition has caused several e-assessment systems to be developed. As the universities cover different subjects and courses, they may have more than one e-assessment system based on the department and the course. Therefore, more resources and budgeting planes are required. This paper discusses the possibility of a generic and flexible e-assessment system as a way to solve this problem. Furthermore, it investigates the generality and flexibility requirements in the field of e-assessment and provides appropriate architecture for a modular assessment system with reference to these requirements.
International Journal of Innovation in Education, 2010
Institutions all over the world are continuously exploring ways to use ICT in improving teaching ... more Institutions all over the world are continuously exploring ways to use ICT in improving teaching and learning effectiveness. The use of course web pages, discussion groups, bulletin boards, and e-mails have shown considerable impact on teaching and learning in significant ways, across all disciplines. ELearning has emerged as an alternative to traditional classroom-based education and training and web lectures can be a powerful addition to traditional lectures. They can even serve as a main content source for learning, provided users can quickly navigate and locate relevant pages in a web lecture. A web lecture consists of video and audio of the presenter and slides complemented with screen capturing. In this paper, an automated approach for recording live lectures and for browsing available web lectures for on-demand applications by end users is presented.
Abstract. The aim of this paper is to describe a natural language processing methodology in order... more Abstract. The aim of this paper is to describe a natural language processing methodology in order to identify opinions that are expressed in the text and populate those opinionated words in a domain-driven opinion lexicon. As part of our recent research in opinion mining, our system aggregates students' feedbacks from several faculties and derives meaningful comments that are expressed by them about courses, teaching, evaluation, infrastructure and many others. In this paper, we employ a rule-based approach to extract features, ...
arXiv preprint arXiv:1001.0443, Jan 4, 2010
Abstract: Educational media mining is the process of converting raw media data from educational s... more Abstract: Educational media mining is the process of converting raw media data from educational systems to useful information that can be used to design learning systems, answer research questions and allow personalized learning experiences. Knowledge discovery encompasses a wide range of techniques ranging from database queries to more recent developments in machine learning and language technology. Educational media mining techniques are now being used in IT Services research worldwide. Multi-modal ...
arXiv preprint arXiv:1003.3530, Mar 18, 2010
Abstract: The basic classification techniques for organizing information are thesauri, taxonomy a... more Abstract: The basic classification techniques for organizing information are thesauri, taxonomy and faceted classification. Topic map is relatively a new entrant to this information space. Topic map standard describes how complex relationships between abstract concepts and real world resources can be represented using XML syntax. This paper explores how topic map incorporates the traditional techniques and what are its advantages and disadvantages in several dimensions such as content management, indexing, knowledge ...
Neurocomputing, 2020
Abstract Storing and processing millions of parameters in deep neural networks is highly challeng... more Abstract Storing and processing millions of parameters in deep neural networks is highly challenging during the deployment of model in real-time application on resource constrained devices. Popular low-rank approximation approach singular value decomposition (SVD) is generally applied to the weights of fully connected layers where compact storage is achieved by keeping only the most prominent components of the decomposed matrices. Years of research on pruning-based neural network model compression revealed that the relative importance or contribution of each neuron in a layer highly vary among each other. Recently, synapses pruning has also demonstrated that having sparse matrices in network architecture achieve lower space and faster computation during inference time. We extend these arguments by proposing that the low-rank decomposition of weight matrices should also consider significance of both input as well as output neurons of a layer. Combining the ideas of sparsity and existence of unequal contributions of neurons towards achieving the target, we propose sparse low rank (SLR) method which sparsifies SVD matrices to achieve better compression rate by keeping lower rank for unimportant neurons. We demonstrate the effectiveness of our method in compressing famous convolutional neural networks based image recognition frameworks which are trained on popular datasets. Experimental results show that the proposed approach SLR outperforms vanilla truncated SVD and a pruning baseline, achieving better compression rates with minimal or no loss in the accuracy. Code of the proposed approach is avaialble at https://github.com/sridarah/slr .
In the past, several automatic video summarization systems had been proposed to generate video su... more In the past, several automatic video summarization systems had been proposed to generate video summary. However, a generic video summary that is generated based only on audio, visual and textual saliencies will not satisfy every user. This paper proposes a novel system for generating semantically meaningful personalized video summaries, which are tailored to the individual user’s preferences over video semantics. Each video shot is represented using a semantic multinomial which is a vector of posterior semantic concept probabilities. The proposed system stitches video summary based on summary time span and top-ranked shots that are semantically relevant to the user’s preferences. The proposed summarization system is evaluated using both quantitative and subjective evaluation metrics. The experimental results on the performance of the proposed video summarization system are encouraging. Index Terms — Video summarization, video semantics, semantic multinomial, personalization, user pr...
Abstract. An approach for extracting higher-level visual features for art painting classification... more Abstract. An approach for extracting higher-level visual features for art painting classification based on MPEG-7 descriptors is presented in this paper. The MPEG-7 descriptors give a good presentation of different types of visual features, but are complex structures. This prevents their direct use into standard classification algorithms and thus requires specific processing. Our approach consists of the following steps: (1) the images are tiled into non-overlapping rectangles to capture more detailed information; (2) the tiles of the images are clustered for each MPEG-7 descriptor; (3) vector quantization is used to assign a unique value to each tile, which corresponds to the number of the cluster where the tile belongs to, in order to reduce the dimensionality of the data. Finally, the significance of the attributes and the importance of the underlying MPEG-7 descriptors for class prediction in this domain are analyzed.
People, such as students, employees and public, are talking about the institution and its busines... more People, such as students, employees and public, are talking about the institution and its business everyday positively or negatively by means of feedbacks, opinions, comments etc through various social platforms. Their feedbacks and opinions are valuable resources for the institution if listened properly. Since feedbacks are by and large unstructured in nature, understanding and extracting the meaningful information from massive data collections becomes a real challenge. This paper outlines the various tasks that are to be carried out during the knowledge discovery process from the learning environments setting.
The 12th International Conference on Advances in Information Technology, 2021
Social media is more and more utilized by people around the world to express their feelings and o... more Social media is more and more utilized by people around the world to express their feelings and opinions in the kind of short text messages. Twitter has been a rapidly growing microblogging social networking website where people express their opinions in a precise and simple manner of expressions. It has also become a platform for governmental, non-governmental and individual opinions and policy announcements. Detecting sentiments from tweets has a wide range of applications including identifying the anxiety or depression of individuals and measuring the well-being or mood of a community. In addition, the sentiment classification becomes complex when the tweet is written in codemix language which is a mix of two different languages. The main objective of this paper is to classify tweets written in mix of Tamil and English language into positive, negative, or neutral. This is achieved by fine tuning a pretrained multilingual text representation model as well as deep learning transfor...
Fundam. Informaticae, 2008
Dance video is one of the important types of narrative videos with an interesting and semantics-r... more Dance video is one of the important types of narrative videos with an interesting and semantics-rich content. Modeling the expressive dance semantics is a real challenging task, compared to the other types of videos such as news, sports and video streams. This paper proposes a formal conceptual model based on the theory of regular tree automata called, DanceGrammar. The DanceGrammar and its tree automaton represent the dance pieces that are associated with the accompanying song. A dance piece incorporates many dance video semantics such as dancer objects, actors, dance movements, emotions and spatio-temporal relationships among the dancers and their actors. This paper then outlines the translation rules to convert the tree automaton of the Dance- Grammar into W3C XML Schema and presents the underlying XML Schema representations.
Dance video is one of the important types of narrative videos with semantic rich content. This pa... more Dance video is one of the important types of narrative videos with semantic rich content. This paper proposes a new meta model, Dance Video Content Model (DVCM) to represent the expressive semantics of the dance videos at multiple granularity levels. The DVCM is designed based on the concepts such as video, shot, segment, event and object, which are the components of MPEG-7 MDS. This paper introduces a new relationship type called Temporal Semantic Relationship to infer the semantic relationships between the dance video objects. Inverted file based index is created to reduce the search time of the dance queries. The effectiveness of containment queries using precision and recall is depicted. Keywords: Dance Video Annotations, Effectiveness Metrics, Metamodeling, Temporal Semantic Relationships.
Alia Abdelmoty, Cardiff University, UK Alex Abramovich, Gordon College, Israel Khalil A. Abuosba,... more Alia Abdelmoty, Cardiff University, UK Alex Abramovich, Gordon College, Israel Khalil A. Abuosba, Arab Academy of Banking and Financial Sciences, Jordan Grigoris Antoniou, University of Crete, Greece Mehmet Emin Aydin, University of Bedfordshire, UK Costin Badica, University of Craiova, Romania Susmit Bagchi, Samsung Electronics (SISO), India Zeki Bayram, Eastern Mediterranean Univ., TRNC Tibor Bosse, Vrije Universiteit Amsterdam, The Netherlands Stijn Christiaens, Vrije Universiteit Brussel, Belgium Frank F. P. Dignum, Universiteit Utrecht, The Netherlands Tharam S. Dillon, Curtin University of Technology, Australia Weichang Du, University of New Brunswick, Canada Peter Eklund, University of Wollongong, Australia Riza Cenk Erdur, Agean University, Turkey Vadim Ermolayev, Zaporozhye State University, Ukraine Martin Gaedke, Chemnitz University of Technology,Germany Manolis Gergatsoulis, Ionian University, Greece M.Adeeb Ghonaimy, Ain Shams University, Egypt Aditya K. Ghose, Universit...
Recent advances in hardware sophistication related to graphics display, audio and video devices m... more Recent advances in hardware sophistication related to graphics display, audio and video devices made available a large number of multimedia and hypermedia applications. These multimedia applications need to store and retrieve the different forms of media like text, hypertext, graphics, still images, animations, audio and video. Dance is one of the important cultural forms of a nation and dance video is one such multimedia types. Archiving and retrieving the required semantics from these dance media collections is a crucial and demanding multimedia application. This paper summarizes the difference dance video archival techniques and systems. Keywords: Multimedia, Culture Media, Metadata archival and retrieval systems, MPEG-7, XML.
International Journal of Networking and Virtual Organisations
In India, nowadays one per 88 children is born with autism spectrum disorder (ASD) contrary to a ... more In India, nowadays one per 88 children is born with autism spectrum disorder (ASD) contrary to a ratio of one per 110 children few years back. In this paper, we propose a social sensor network for individuals with ASD to offer medical promotions as intervention and first aid. They should be in care and medical surveillances. We are monitoring enjoyment, frustration, boredom, and stress by using body-worn sensors and video streaming. This sensor collects the information from the human body and transmits it to their smart phones then to the external server. We introduce an algorithm named as spectrum clustering algorithm (SCA) which clusters people based on activities of the child. Communication is provided between the clusters of people by using context recommendation algorithm. We use prerogative response algorithm (PRA) for load balancing on server while processing the requests. Our processes are simulated using Omnet++ simulator and provide effective results.
Uploads
Papers by Rajkumar Kannan