SVM BASED INTELLIGENT SYSTEM FOR CLASSIFICATION OF MRI BRAIN IMAGES

S.Sakkaravarthi 1, Dr.K.G.Srinivasagan 2, S.MuthuKumar 3
1 Assistant Professor, Department of CSE, Sree Sowdambika College of Engineering, Aruppukottai, Tamilnadu, India.
2 Professor & Head - Department of CSE-PG, National Engineering College, Aruppukottai, Tamilnadu, India.
3 Professor & Head - Department of CSE, Sree Sowdambika College of Engineering, Aruppukottai, Tamilnadu, India.

     

Medical imaging plays a vital role in diagnosing the diseases as well as the study of human anatomy and physiology. MRI is an advanced medical imaging technique especially used for capturing the human brain. The manual interpretation of brain tumor slices based on visual examination by physician may lead to missing diagnosis and time consuming when a large number of MRI brain images are analyzed. To avoid human based diagnostic error, automated brain tumor classification is preferred. For automated brain tumor classification, various techniques are available. Those techniques suffer due to misclassification which is unfavoured by physicians. Still the problem is open and research is going to promote better result in very fast manner. Still the problem is open and research is going to promote better result in very fast manner. Automated MRI Brain image classification using Support Vector Machine classifier is proposed to classify brain image into normal or abnormal. Brain abnormality can be further classified into benign or malignant. In this paper, MRI brain image is pre-processed using median filter. Then, statistical based texture features are extracted from Gray Level Co-occurrence Matrix (GLCM) of an input brain image. After feature extraction, relevant features are obtained using forward feature selection technique to reduce the feature space. The selected features are given as input to Support Vector Machine classifier. Finally, Support vector machine classifier is utilized to perform two functions. The first is to differentiate between normal and abnormal. The second function is to classify the type of abnormality in benign or malignant tumor. 

View full abstract...

A MODIFIED ACO USING ROULETTE WHEEL SELECTION FOR SOLVING COMBINATORIAL OPTIMIZATION PROBLEM

Sathya V
PG Scholar, Dept of CSE, Sri Ramakrishna Engineering College, Coimbatore, Tamilnadu, India .

  

   In Cloud computing, private cloud is a branch where resource sharing is becoming more popular now days. The importance of resource sharing has lead to the development of many algorithms. The existing works in the field of resource sharing used many optimization techniques. The two techniques namely Service Composition Optimal Selection (SCOS) and Optimal Allocation of Computing Resources (OACR) are combined and named as Dual Scheduling of Cloud Services and Computing Resources (DS-CSCR). This dual scheduling is not suitable for large scale problems and hence, a new Ranking Chaos Optimization (RCO) was introduced. RCO executes with low time consumption. But the design of heuristic function is complex and hence, a Modified Adaptive Chaos Optimization (MACO) technique with Roulette Wheel Selection (RWS), is used for allocating tasks to the resources. Finally the comparison of the proposed work with the existing RCO proves that, the resource allocation can be made with low time consumption and the stability of the virtual machine increases. 

View full abstract...

LOW POWER AND HIGH PERFORMANCE ADDRESS GENERATOR FOR WIMAX DEINTERLEAVER

D.Poornima Devi
PG Scholar, Department of Electronics and Communication(VLSI), SriVidya college of Engineering and technology, Virudhunagar, Tamilnadu, India.

     The aim is to generate the address generation circuitry of Deinterleaver used in the WiMAX transreceiver using the Xilinx Field Programmable Gate Array (FPGA). The floor function associated with the implementation of FPGA is very difficult in IEEE 802.16e standard. So we eliminate the requirement of floor function by using a simple mathematical algorithm. Some modulations like QPSK, 16-QAM and 64-QAM along with its code rates make our approach to be novel and high efficient. By using the majority logic circuit and biorthogonal decoding it can be used to reduce power consumption and latency compared to the previous paper. Power consumption can be reduced up to 34mw and latency can be reduced up to 1.206ns.

View full abstract...

FREE ENERGY ELECTRIC BI-CYCLE

Jothilingam S1, Pavunraj P2
1 UG scholar, Department of Electrical and Electronics, Vickram College of Engineering, Madurai, Tamil Nadu, India.
2 Faculty of Electrical and Electronics, Vickram College of Engineering, Madurai, Tamil Nadu, India.

     The main work deals with free energy concept. The objective is to produce energy and also utilize it without any external energy source or external charging. As we are using two different modes to power as a hybrid model. The next source is a dc generator (24 volt) which is arranged in such a way that it produces energy when the front wheel of the cycle starts rotating. Finally we have used a dc motor (12 volt) to drive the cycle. There is lead acid battery (12 volt) used for storing the power produced by the generator and the stored energy is utilized by the motor to drive the system simultaneously. This paper describes the process of planning, designing, and testing a hybrid electric bicycle. It provides a lot of detail into the challenges of modifying an existing mechanical system to one that is based on both human propulsion as well as a set of electro-mechanical interfaces that provide assists. Through designing an electro-mechanical system, with various non-human inputs and feedback channels, a major challenge was centralizing the control of the system. After establishing criteria for speed, control, efficiency, and weight, we began a process of selecting parts and developing models for how the overall system including the rider could be integrated in a way that is both safe, and easy to use. 

View full abstract...

A MULTI-AGENT BASED APPROACH TO RECONFIGURABLE MULTIPROCESSOR ARCHITECTURE MANAGEMENT

P Arunraj 1, K R Sarath chandran2
1 Master of Engineering, Department of CSE, SSN College of Engineering, Chennai, Tamilnadu, India.
2 Assistant Professor, Department of CSE, SSN College of Engineering, Chennai, Tamilnadu, India.

    

 One of the most important research area in computer hardware is reconfigurable computing and it is becoming gradually more attractive for many useful applications. Reconfiguration is the process of changing the structure of a reconfigurable device at start-up time. In Reconfiguration there are many issues, among them reconfiguration time is the main factor. Media applications such as voice and video processing as in HDTV and some real time applications are becoming increasingly complex and consume more time and more computing power. The main objective of this paper is to reduce the reconfiguration time by applying Multi-Agents to the system. By doing so, they provide better price per performance and consume less power computation. A comparison is made between the traditional architecture without agent and proposed architecture with agents.

View full abstract...

STUDY OF DESIGN IN SOLAR VAPOUR ABSORPTION REFRIGERATION SYSTEM

¹ X.S.Ashok Jain, ² M.Edwin
¹ Assistant Professor, Department of Mechanical Engineering, Einstein College of Engineering, Tirunelveli, India.
² Assistant Professor, Department of Mechanical Engineering, University college of Engineering Nagercoil, Anna University Tirunelveli Region, India.

     This paper presents the modeling of a solar-powered, absorption cooling system using an evacuated tube collector and water-Lithium bromide (LiBr-H2O) solution. As solar processes are inevitable transient in their operation, the energy ratios and temperatures will vary with time and Coefficient of Performance (COP). It is noted that a well modified collector could improve the performance of the system. Therefore a configured solar based LiBr-H2O absorption cycle is suggested after a review made in the selection of collectors. An extensive study on the design procedure of temperature, energy balance and mass conservation is made improving the coefficient of performance analytically.

View full abstract...