Volume 8 Number 2 September 2018


A Study and Analysis of Content based medical image retrieval for DICOM Images using Deep Convolutional Neural Network
P. Haripriya, R. Porkodi

Abstract- - Content based medical image retrieval (CBMIR) system is an effective way of supplementing the diagnosis, treatment for various diseases and it is also an efficient management tool for handling large amount of data. The important issue in content-based medical image retrieval is a semantic gap. This research study focused on reducing the semantic gap between low level feature such as visual information captured by the device and high-level semantic concepts are perceived visual information by human vision system. Among several techniques, machine learning techniques are used earlier as actively investigated as a possible direction to bridge the semantic gap for a long term. To provide an effective medical image classification and retrieval service, the intelligent content based medical image retrieval with sematic system is required. So, the recent success of deep learning techniques overcomes the semantic gap problem. In the medical domain DICOM images plays a vital role and it is cursed with huge dimensionality. In this paper, CBMIR employed with deep learning techniques and it is discussed. The experimental data is very complex in nature and it contains the large set of training samples. The DCNN technique has been employed to classify the DICOM images and obtained the high accuracy.


Sports Analysis of FIFA Football World Cup Tournament using Logistic Regression
P. Sudhandradevi, V. Bhuvaneswari

Abstract- - Sports became a prominent part of the human life. Sports analysis provides the expertise on sports-related events. Sports participants have the higher levels of physical activity, psychological health and social welfare. In current strategy sports analytics became a buzz word. The sentiment by sports journalist Grantland Rice said that “not that you won or lost but how you played the game”. Sports science is a widespread academic discipline, applied to areas including athletes performance and Olympic game. The sports data is fine tuned from the fine-tune technique or wearable technology. The objective of the paper is sto find the team who gives their contribution in FIFA Football tournament from the year 1872-2018. The Logistic regression technique is used find the probability of win and loses. This technique has been implemented in R tool based on logistic regression model. The outcome of the prediction gives 76% of accuracy of the model design and also it contributes continent wise football interest among the globe.


Design and Development of Automated Ontology from PubMed abstracts using Rule based approach
G. Suganya, R. Porkodi

Abstract- Ontology is an emerging discipline that has the huge potential to improve information in organization, management and understanding. It has a crucial role to play in the field of information extraction and information retrieval. Gene names extraction is an important problem in the area of biomedical field through which the hidden associations among genes, diseases, mutations and drugs can be extracted and helpful in solving many diseases in human. This paper developed a framework to build an automated ontology from the PubMed abstracts. The design and development of automated ontology consists of two important phases: Identifying the gene names using set of rules and construction of automated ontology from the identified gene names. This work uses 100 PubMed abstracts randomly. The developed automated ontology extracted and visualized the gene names using DLquery and the gene names extracted by this framework using rule based approach compared with existing Genia tagger.


Performance Evaluation of Descriptors Extracted by MSER Detector for Human Action Recognition
R. Rajeswari, P. Ramya

Abstract- Human action recognition helps in automatically analyzing various events in video data. It has been used for recognizing human actions in many applications including surveillance, healthcare and human-computer interface. In order to recognize human actions in videos various feature descriptors and detectors have been proposed in the literature. The feature detectors help in extracting feature descriptors which provide vital information related to the human actions in video frames. One such feature detector is maximally stable extremal regions (MSER) which is widely used for detecting blobs in video frames. In this paper, the performance of various feature descriptors such as Binary Robust Invariant Scalable Keypoints (BRISK), Histogram of Gradients (HOG) and Speeded Up Robust Features (SURF) extracted by MSER for human action recognition is investigated. Experiments are performed on KTH action dataset.


An Heterogeneous Information Processing using Big Data
R. Nivedha, S. Arshiya Sulthana

Abstract- Big Data though it is a hype up-springing many technical challenges that confront both academic research communities and commercial IT deployment, the root sources of Big Data are founded on data streams. It is generally known that data which are sourced from data streams accumulate continuously making traditional batch-based model induction algorithms infeasible for real-time data mining or high-speed data analytics in a broad sense. In this paper, a novel data stream mining methodology, called Stream-based Holistic Analytics and Reasoning in Parallel (SHARP) is proposed. SHARP is based on principles of incremental learning which span across a typical data-mining model construction process, from lightweight feature selection, one-pass incremental decision tree induction, and incremental swarm optimization. Each one of these components in SHARP is designed to function together aiming at improving the classification/prediction performance to its best possible. SHARP is scalable, that depends on the available computing resources during runtime, the components can execute in parallel, collectively enhancing different aspects of the overall SHARP process for mining data streams. It is believed that if Big Data are being mined by incrementally learning a data mining model, one pass at a time on the fly, the large volume of such big data is no longer a technical issue, from the perspective of data analytics. Three computer simulation experimentations are shown in this paper, pertaining to three components of SHARP, for demonstrating its efficacy.