International Journal of Computer Science and Technology
IJCST 9.2 ver-1 (April-June 2018)

S.No. Research Topic Paper ID Download

Improvement of K-means using Ruleset

Mary Ambika Babu, Neeba E A


The development of clustering analysis technology, there have been many application-based clustering algorithms, such as text clustering. K-means is a method of clustering observations into a specific number of disjoint clusters. The “K” refers to the number of clusters specified. Various distance measures exist to determine which observation is to be appended to which cluster. The algorithm aims at minimizing the measure between the centroid of the cluster and the given observation by iteratively appending an observation to any cluster and terminate when the lowest distance measure is achieved. The two big limitations that the K-Means algorithm has, number of cluster, K, must be determined beforehand and random selection of initial cluster centre, proposed K-Means improved algorithm based on the minimum rule set. This method proposed the concept of the minimum rule covering set. In order to solve the two big limitations of K-Means algorithm effectively. Performance analysis between traditional k-means and improved k-means is evaluated using Chi-Square method, Entropy and F-Measure method.
Full Paper


Bone Age Assessment using Deep Learning

Mariyam Sunny Kattukaran, Abey Abraham


The assessment of skeletal bone age is very subjective and tiring process. As a result, computer assisted techniques are developed to replace hand operated examination techniques in medical industry. The research aims to find out a new computer aided technique which is based on convolutional neural network. It is developed to check the skeletal maturity as well as gender and race. The most important advantage of this proposal is that it minimizes the segmentation problem bared by the existing systems. Skeletal bone age assessment is the most used clinical practice to scrutinize endocrinology, genetic and growth disorders in the youngsters. It is usually executed using radiological analysis of left hand wrist using Greulich and Pyle technique or Tanner Whitehouse technique. Nevertheless, both the techniques have many cons including human observation variability and consumption of time.
The study proposes deep learning techniques to check skeletal bone age and also to generalize the gender and race. This bone age assessment technique works on public data set for all age ranges, races and genders.
Full Paper


Study on Differences between Microsoft Skype for Business and Microsoft Skype Broadcast

Siddarth Kaul, Dr Anuj Jain


The paper proposes to clearly elaborate the differences and understanding of Microsoft Skype for business and Microsoft Skype Broadcast as known Skype for Business is a Unified communication Platform (Formerly Known as Lync) for audio & Video Conferencing, IM and Image and File Transfer capabilities which is a successful platform in the Market whereas Skype Broadcast is a portal for Online schedule of Meeting Broadcast and audio & Video Conferencing on Microsoft Portal.
Full Paper


Relative Study for Mammographic Images using Edge Detection Algorithm

Apoorva Jain


Image processing refers to manoeuvring of the grey point information contained within the pixel of a digital image. Image edge detection is one of the significant inside of image processing. In this paper, I will show various edge detection operators, and their comparative result images for mammographic images. A mammogram is an X- Ray representation of breast. These images are used as a screening means to detect early breast cancer in women experiencing no symptoms or in women experiencing symptoms such as ache and nipple absolve. There is resemblance between normal and cancerous breast tissues; early detection of breast cancer is tricky. Cancer analysis with the support of image processing includes the subsequent things-Image acquirement, enhancement, post-processing and diagnosis. I used MATLAB software to implement edge detection algorithms for early detection of breast cancer in women.
Full Paper


The Energy Efficient Routing and High Security Transmission in Mobile Ad-hoc Networks

G V Eswara Rao, K V Chandra Sekhar


In this paper we exhibit an overview of secure ad hoc routing conventions for mobile wireless networks. A mobile ad hoc network is a gathering of nodes that is associated through a wireless medium shaping quickly evolving topologies. The generally acknowledged existing routing conventions intended to suit the requirements of such self-sorted out networks don’t address conceivable dangers going for the interruption of the convention itself. The suspicion of a trusted situation isn’t one that can be sensibly expected; consequently a few endeavors have been made towards the plan of a safe and powerful routing convention for ad hoc networks. We quickly introduce the most prevalent conventions that take after the table-driven and the source-started on-request approaches. In view of this discourse we at that point define the risk demonstrate for ad hoc routing and display a few particular assaults that can focus on the activity of a convention. With a specific end goal to dissect the proposed secure ad hoc routing conventions structurally we have arranged them into classifications; arrangements in view of cryptography, arrangements in light of symmetric cryptography, notoriety based arrangements and a class of addon components that fulfill particular security prerequisites. An examination between these arrangements can give the premise to future research in this rapidly developing region.
Full Paper


Document Clustering: How to Measure Quality of Clusters in Absence of Ground Truth

Iti Sharma, Harish Sharma


Demand of simple and scalable clustering algorithms for text documents is increasing as the volume of data generated by through internet is exploding. There are no known classes for such data and extrinsic measures of quality are not sufficient to guide about which algorithm is better for an application. This paper suggests four different intrinsic measures that can be used to evaluate cluster output and hence the clustering method to suit a particular application. The proposed metrics measure homogeneity and coherence of documents in a cluster as well as the overlap among different clusters in an interpretable form.
Full Paper


A Review on Face Detection and Recognition Techniques

Bijoy Chhetri, Ishwar Sharma, Sarita Thapa, Chitrakala Pradhan, Santa Kumari


The main purpose of this paper is to review the methods of face detection and face recognition with machine learning as a core recognition algorithm. The different methods that are involved has been categorically mentioned where face detection and recognition is taken as the first step for information drawing for many image processing application.
The live capturing and distinguishing of an object and detection and recognition is generally a challenging task. In this review, an attempt has been made to do research on existing algorithms and schemes is defined the limitation of different method and system in place has been discussed.
Full Paper


Digital Preservation and Data Compression

Vipul Sharma, Roohie Naaz Mir


The spread of computing has led to an explosion in the volume of data to be stored on hard disks and sent over the Internet. This growth has led to a need for “digital preservation” & “data compression”, that is, the ability to preserve the data digitally and to reduce the amount of storage or Internet bandwidth required to handle this data respectively. This paper provides information on different preservation standards, strategies, problems and drawbacks related to preserving data digitally and also provide a survey of different compression algorithms. The focus is on the most prominent data compression algorithms like Huffman and LZW.
Full Paper


Cloud Service Recommendation Approach Using Group Consensus Experiences and Contextual Data

Dr. Hlaing Phyu Phyu Mon


Owing to increase in the number of cloud services with various kinds of service attributes in cloud computing environment, it is usually hard for the users to request suitable service attributes for a service they want. This paper recommends some suitable service attributes by learning user context attributes and group experiences. K-means clustering is adopted to classify historical user groups context attributes and the context itemsets obtained from k-clusters are further mined using associative mining between context and service properties oriented to cloud services to recommend what service attributes should be requested according to context properties of current user. In addition, user groups subjective experiences are also considered upon recommended service list to reduce the exception differences. Finally, the proposed approach is validated and compared with proficient service prediction approach called CAMF.
Full Paper


The Efficient Web based Low-Bandwidth System for Telemedicine Applications

Gowri Sree Lakshmi Neeli


Telemedicine to take a shot at low bandwidth is a website and this application intended to work with portable web association (or) in Offline mode (if there should arise an occurrence of no system availability in remote zones) furnishes an easy to use interface with which the patient and specialist can have a vivacious cooperation. The patient can communicates his manifestations to the specialist and can take proposals from him. With this application the patient can spare his/her time traveling from far separations to the doctor’s facilities. Telemedicine has increased enormous prominence in creating nations where rustic populace is denied access to even fundamental medicinal services. Directly telemedicine is demonstrating amazingly reasonable and possible arrangement connecting with rustic populace and crossing over uniqueness in quality and access to social insurance amongst urban and provincial areas. The telemedicine market has seen dynamite development of late for the most part in light of merging Information innovation Communication and Healthcare. This paper looks at the present condition of telemedicine in creating nations. It additionally examines telemedicine execution cases, lessons gained from the cases, and finishes up with potential researchable basic achievement factors that record for the development and unassuming triumphs of telemedicine. The paper likewise quickly examines about the headways in execution of telemedicine in created countries.
Full Paper


Logical Knowledge-based Advanced Cost Estimation Methodology (LKACEM) Applied to Metal Matrix Composite Aero-Engine Blisk

Nikhil Thakur, Pathmeswaran Raju, Michal Krzyzanowski, Phani Chinchapatnam


Cost estimation is an important activity for advanced understanding of product/process knowledge that is used to plan activities accordingly. For composite material parts, choices of materials and their methods of manufacturing are broad and thus complexity is high. Design is complex too involving tight tolerances leading to need of a new and advanced costing system. Proper knowledge management followed by improving the current cost estimation methods is a viable solution. This paper proposes a logical knowledge-based advanced cost estimation methodology that uses a mathematical set theory-based knowledge management system designed by utilising a generic product life-cycle for Knowledge Information & Data collection. This acquired KID is represented as parent sets, subsets and elements. The knowledge structure so created is coupled to a mixed method of cost estimation for developing logical advanced cost estimation system that can be used for both composite and conventional costing. Standard rules governing the parameters of cost and their relationships with the knowledge base is included in this methodology as a part of logical layer which interacts with other layers to form a reliable cost estimate. This methodology is then applied to develop a factory cost model for an aero-engine blisk design made up of metal matrix composite. This is done by using both simple and advanced software tools. Finally the outcome from these softwares are analysed by comparison study. The comparison is shown graphically for machining, material and overall cost parameters as a difference in their output cost values. As an outcome it is proved that the methodology is flexible for use with different softwares, is capable of reliable estimates, is less complex and thus can be used for cost estimation.
Full Paper


Discovering the Consigned Technique for Detection and Prevention of SQLIA’s

Immanuel Wonderful C.J, J. Hari Babu


SQL injection is a procedure that endeavors a security liability happening in the database layer of an application. The defenselessness is available when client input is either erroneously separated for string exacting break characters inserted in SQL proclamations or client input isn’t specifically and along these lines startlingly executed. SQL injection is a trap to SQL query or order as an info potentially through the pages. They happen when information gave by client isn’t appropriately approves and is incorporated straightforwardly in a SQL query. By utilizing these vulnerabilities, an attacker can submit SQL charges straightforwardly access to the database. In this paper we display all SQL injection attack writes and furthermore unique procedure and apparatuses which can identify or keep these attacks. To address this issue, we display a broad audit of the distinctive sorts of SQL injection attacks. We additionally present and consider recognition and prevention procedures against SQL injection attacks.
Full Paper


A Novel Pulse-Coupled Neural Network using Gabor Filters for Medical Image Fusion

Fayadh Alenezi, Ezzatollah Salari


Medical image fusion is important in biomedical applications for non-invasive diagnosis. Image fusion aims to reduce defects associated with single images created from different modalities. These defects make the single-image result less informative and therefore less useful for medical diagnosis. An ideal fused image, created from two or more multimodality images, increases the accuracy, information content and improves visual properties. Current state-of-the-art image fusion techniques have not successfully resolved the poor visual properties, leading to lessthan-ideal information quality. This paper describes a novel technique to improve information quality of fused images, with many practical applications in the biomedical, military, and remote sensing. The proposed algorithm combines the action of Gabor filtering, maximum pixel intensity selection and Pulse Coupled Neural Network (PCNN) implementation. The results are then used to create a fused image. As a proof of concept, several images are evaluated with standard criteria and compared with results from existing image fusion methods.
Full Paper


Aspect Oriented Software Development: Concepts, Issues and Real-World Applications

Priyanka. R. Sarode, Dr. R. N. Jugele


Aspect-oriented software development supports better separation of concerns by presenting a new modular unit, called an aspect, for modularization of crosscutting concerns. As a new type of modular unit, aspect should have clear edges that define the way they communicate with the rest of the system and how they affect other elements. Several programming languages and mechanisms proposed for implementing aspect-oriented systems, and these systems are beginning to use for real-world applications. Based on the collected data this paper gives the conceptual discussion on the key concepts, pitfalls of Aspect Oriented Software Development and discuss the major industrial projects using Aspect Oriented Software Development.
Full Paper


Cloud Computing: Fundamentals

Pedro Ramos Brandao


Deepening the concept of Cloud Computing, describing and analyzing the underlying technologies in this concept’s technological implementation. The relationship of distributed computing and virtualization with Cloud Computing. The need for automation in Cloud Computing.
Full Paper