Informatics Publishing Limited

S. Murugavalli ¹

Affiliations
1 Panimalar Engineering College, Chennai, IN

Abstract

Image segmentation plays a vital role in medical image analysis and detection. In this paper, automatic brain tumor detection from magnetic resonance image using wavelet based genetic algorithmic approach was implemented. First, the MR images are preprocessed using discrete wavelet transform, and then the genetic algorithm is applied to detect the tumor pixels. All volumes were processed for abnormal classification. Tumor segmentation was performed on the abnormal slices and the results were compared with a ground truth tumor volume. A total of 100 real time data were acquired from magnetic resonance imaging system.

Keywords

Discrete Wavelet Transform Genetic Algorithm, MRI Brain Tumor, and Segmentation.

Full Text

     

M. Rajeev Kumar ¹, J. Senthil ¹, S. Murugavalli ²

Affiliations
1 Department of Information Technology, Veltech Dr.RR & Dr.SR Technical University, Avadi, Chennai, IN
2 Department of Computer Science and Engineering, Panimalar Engineering College, Chennai, IN

Abstract

Iris recognition, the ability to recognize and distinguish individuals by their pattern, is the most reliable biometric in terms of recognition and identification performance. However,performance of these systems is affected by the heterogeneous images (regarding focus, contrast, or brightness) and with several noise factors (iris obstruction and reflection) when the cooperation is not expectable from the subject. Current Iris recognition system does not deal with the noise data and substantially increase their error rates in these conditions. An Iris classification method is proposed on the segmented and normalized iris image that divides the image into six regions, followed by independent feature extraction in each region. This will provide the iris signature in terms of binary values, then that are compared with each region for the identification. Iris liveness detection also called fake identification aims to ensure that an input image sequence is from a live subject instead of an iris photograph, a video playback, a glass eye or other artifacts. However, efforts on iris liveness detection are still limited, though iris liveness detection is highly desirable. In addition to this Fake identification is also done in this paper. Fake, the original image is forged by fixing lenses over the iris portion. This can be identified by using fast Fourier transform.

Keywords

Noncooperative Iris Recognition, Iris Classification, Feature Extraction, Biometrics, Fake Identification.

Full Text

     

B. Indhuja ¹, S. Mala Devi ¹, S. Murugavalli ²

Affiliations
1 Department of Computer Science and Engineering, Panimalar Engineering College, Chennai, IN
2 Computer Science and Engineering Department, Panimalar Engineering College, IN

Abstract

Speech Recognition (SR) is a process that transcribes speech into text using a computer. Speech recognition system is a speech-to-text conversion wherein the output of the system displays text corresponding to the recognized speech. A step towards a more natural, “human-like” communication between machines and users in need of information is represented by the introduction of speech language technologies into Information Retrieval system. The integration of the Information Retrieval system (IR) and the Automatic speech recognition (ASR) system degrades the performance. The failure to recognize a keyword in continuous speech may drastically affect the performance of the IR system. The proposal is to build an ASR system that is trained to identify the word we pronounce in restricted domain. Then for each of the recognized word, the system is expected to find the match which is then extracted by the IR system.

Keywords

Automatic Speech Recognition (ASR), Information Retrieval (IR), Word Error Rate (WER).

Full Text

     

I. Thamarai ¹, S. Murugavalli ¹

Affiliations
1 Sathyabama University, Chennai 600 119, Tamil Nadu, IN

Abstract

Objectives: Software effort estimation is a critical task in the software development process due to the intangible nature of software. A new model for software effort estimation using Differential Evolution Algorithm called DEAPS is proposed in this paper. Method: In this methodology, the complete set of historical project base is reduced to a set of similar projects using Euclidean distance metric. Then Differential Evolution Algorithm which is an Evolutionary Computation method is used for optimization and the most relevant project is retrieved. The proposed method is validated on Desharnais dataset. Findings: DE has a very effective mutation process which improves the ability of exploration. So we got promising results which indicate that the use of this model could significantly improve the efficiency of Analogy based Software Effort Estimation. The metrics used are MMRE, MdMRE and pred (25%). The results are compared with previous findings and the results clearly show that the proposed method is better than the existing methods. Application: This methodology can be used to minimize the errors in the software estimation so that financial loss and delay in the completion of project may be avoided.

Keywords

Algorithmic and Non-Algorithmic Models, Differential Evolution Algorithm, Evolutionary Computation, Software Effort Estimation.

Full Text