Accepted Papers
- Randomization in Image Steganography Using RGB
Saikat Banerjee and Lovey Rana, SOIT, Centre for Development of Advanced Computing (CDAC),India.
ABSTRACT
The paper states the approach of embedding text secret message into an JPEG image file by dividing the entire image into blocks of 4 pixels and finding the key position using the values of RGB from the first block and then arranging the entire blocks into array and selecting particular blocks randomly for embedding the secret message. The idea is to finding a random pattern that is not easy to recognize by an attacker and along with that spreading the secret message throughout the image to cause a smooth distribution that will not affect the image quality much with a high capacity for data embedding.
- A comprehensive scalable framework for Bioinformatics using Map-reduce for Heterogeneous Multicore Systems
Prashanth Thinakaran, Srinivas Avireddy, Prasanna Ranganathan, Sumalatha Ramachandran, Anna University,India.ABSTRACT
In this paper we propose Map reduce based Application Programming Interface framework for Heterogeneous Multi-core systems in a cloud environment. Exploiting the heterogeneity by harnessing the capabilities of both CPU'S and GPGPU's simultaneously warrants a significant improvement in performance at the same time conserving the energy per compute instruction. Existing approaches exploit parallelism either from CPU or from GPGPU cores. The proposed API exploits the parallelism across both the CPU as well as GPU nodes simultaneously by adopting novel heuristics in scheduling and work load balancing thus paving a path for efficient resource utilization of heterogeneous compute nodes in cloud environment. Data processing in Bioinformatics involves a lot of embarrassingly parallel algorithms which are mostly executed in sequential manner consuming a lot of computational time/power leading to poor resource utilization. Using the proposed Map Reduce API, an adaptive Map Reduce based framework for Heterogeneous Multi-core systems is intended bring in parallelism based enhancements for varied type of algorithms. The mapping of computation to the underlying architectural resources (core/node compute components) is completely automated in MR-API, ensuring the extraction of maximum potential of the compute nodes. The existing approaches demand the programmers to provide this mapping schema which in turn demands a lot of change in the algorithm devised to adapt to the underlying architecture. The proposed API overcomes this issue and hence becomes very much suitable for exploiting parallelism in the field of Bioinformatics by effectively making use of CPU and GPGPU's. MR-API adopts novel heuristics in automating the process of mapping computations to processing elements including map-reduce in order to tap the parallelism of the application in a cloud environment. The application of MR-API framework to map computations to computational elements on a CPU+GPU nodes, it is also established that the proposed API framework is scalable, fault tolerant and adaptable to various benchmark sets and system configurations. The working of the proposed framework is as follows: the given source code in split into CPU and GPU specific instructions, followed by spawning threads for CPU units and GPU accelerators and finally scheduling instructions which is driven by load balancing heuristics. The optimization of the generated final code is by building Directed acyclic graphs which gives a vivid picture concerning the outright dependencies across the instructions. The proposed API was tested using standardized benchmarks such as Binomial, Blackscholes, matrix multiplication and a speed up of 50-60% has been obtained when compared against sequential CPU processing. The API was adapted to the Bioinformatics algorithms which involved varied combinations of workload sets such as light computation for large dataset and heavy computation for small dataset. The prominent Bioinformatics based algorithms such as BLAST, Self Organizing Maps, Keyword frequency and stem slicing were considered and a speed ups of about 60-70% were obtained as against the existing sequential processing algorithms. Similarly the fault tolerance, scalability, energy efficiency, resource utilization are also established for the proposed framework.
- Statistical Analysis based Hypothesis Testing Method in Biological Knowledge Discovery
Md. Naseef-Ur-Rahman Chowdhury, Suvankar Paul, and Kazi Zakia Sultana,Chittagong University of Engineering & Technology (CUET),Bangladesh.l, and Kazi Zakia Sultana
ABSTRACT
The correlation and interactions among di erent biologicalentities comprise the biological system. Although already revealed interactions contribute to the understanding of di erent existing systems,researchers face many questions everyday regarding inter-relationships among entities. Their queries have potential role in exploring new relations which may open up a new area of investigation. In this paper,we introduce a text mining based method for answering the biological queries in terms of statistical computation such that researchers can come up with new knowledge discovery. It facilitates user to submit their query in natural linguistic form which can be treated as hypothesis. Our proposed approach analyzes the hypothesis and measures the p-value of the hypothesis with respect to the existing literature. Based on the measured value, the system either accepts or rejects the hypothesis from statistical point of view. Moreover, even it does not nd any direct relationship among the entities of the hypothesis, it presents a network to give an integral overview of all the entities through which the entities might be related. This is also congenial for the researchers to widen their view and thus think of new hypothesis for further investigation. It assists researcher to get a quantitative evaluation of their assumptions such that they can reach a logical conclusion and thus aids in relevant researches of biological knowledge discovery. The system also provides the researchers a graphical interactive interface to submit their hypothesis for assessment in a more convenient way.
-
Ocmr - Online Calligraphic Manuscript Recognition in Cloud
Jeeva Rathanam. G, Ashwin Srinivas. S,Anna University,India..
ABSTRACT
Now-a-days the usage of the portable computing devices such as iPhones, tablets etc., has increased and non-keyboard based methods which are used for data entry are receiving more attention in the research communities and commercial sectors. In cloud the documents are written in different languages and scripts. Most of the text recognition algorithms are made to work only in a particular script and treat any input text has being written only in the script under consideration.
Therefore, in cloud environment the online document analyzer must first identify the script before employing a particular algorithm for text recognition. Automatic transcription of multilingual documents and search for documents on the web containing a particular script are some of the most important applications of handwritten script's recognition. In handheld devices the usage of handheld scripts has increased which accept handwritten input has created a growing demand for algorithms that can efficiently analyze and retrieve handwritten data. Online documents may be written in different languages and scripts. The final documents are stored in cloud environment for future access and retrieval of the document in any language. A single document page in itself may contain text written in multiple scripts.
We proposed a method to classify words and lines into one of the six major scripts: Devnagari, Han, Cyrillic, Hebrew, Roman or Arabic. The Classification is based on different spatial and temporal features that are extracted from the strokes of the different words. The proposed system attains an overall accuracy of 87 percent at word level. In cloud, the scripts are analyzed and translated for end user. The results are displayed in the handheld devices as per the language or script required by the end-user. By using this proposed module the online input document has segmented into individual lines and words to facilitate the feature extraction and further recognition process.
-
Effective High Branch Coverage Testing Process by
Genetic Algorithm and Symbolic Execution
M.Parthiban,Anna University,India.
.ABSTRACT
Testing is the vital method to maintain excellence guarantee by gather information about the activities of the software being industrial or customized. In organize to effectively test software; an amount of different testing technique must be performed. White-box testing or structural testing is a technique of testing software that tests workings of an application, as divergent to its functionality. Usually, software quality has been assured through manual testing which is monotonous, difficult, and frequently gives poor coverage of the source code mainly when availing of random testing approaches. This led to more work in the formal validation in software testing. In order to sufficiently test software, various testing techniques must be performed. One class of testing techniques used extensively is white box testing in which properties of the software code are used to ensure certain branch coverage.
Typically, a testing tool targets one type of structural test and the software unit is the program, file or particular methods. Generally, in Object Oriented Programs, branch coverage is an important issue. To solve this issue, Evolutionary Testing Techniques and Symbolic Execution Techniques have been developed in the existing works. Evolutionary Testing suffer from path problem, in which search progression is led away from its goal. The tools used for Evolutionary Testing Techniques cannot be able to guide test generation directly, since they don't use line up structure or semantic information. The tools used for Symbolic Execution Techniques alone will not be the effective tools, because they do not helpful for generating method sequences, which produce required receiver-object states or non-primitive-argument states. Both techniques do not have the ability to use execution feedback to avoid generating exception-throwing tests, and to produce both fault-revealing and regression tests, so they cannot achieve the highest branch coverage.
In this work, branch coverage is given importance, because it is one of the most challenging criteria in the literature. The proposed combination of two different tools allows generating, automatically branch coverage software test cases for programs written in Java. This novel integration of two techniques Genetic Algorithm and Symbolic execution addresses the structural testing problems with high branch coverage.
- Reliability of PPDDalgorithmincloud computing
Zeinab Shams, Akbar Farhoodi Nejad , and Razieh ghiasi,University Of Qom, Iran.ABSTRACT
In this paper we study effect of reliability in a scheduling algorithm, namely Processor-set Partitioning and Data Distribution Algorithm (PPDD), in order to scheduling divisible loads originating from multiple sites in single-level tree in which multiple loads among all processors are distributed while the overall processing time of all jobs is minimized. In this paper, we first survey the various existing scheduling algorithms in cloud computing environment and then tabulate their various parameters. Besides, addressing reliability in scheduling the multiple loads from multiple sites in the DLT (divisible load scheduling theory) paradigm a first time could be an interesting contribution of this paper. Furthermore, we assume arrival rate of failures follows a Poisson distribution and use the Check Pointing Technique for obtaining fault-tolerance in order to update power processing per each processor (site). In fact, we consider probability of failure at each processor and recovery time of any failure in order to make processing power on each processor more equitable. Since reliability is in contrast with processing time, therefore, taking the failure probability and recovery time into account increases the overall processing time. Also, some sensitivity analysesare applied to evaluate the proposed algorithm. By this mechanism, the proposed algorithm gets closer to reality and its performance becomes better in comparison to PPDD algorithm.