S. U. Aswathy ¹, G. Glan Deva Dhas ², S. S. Kumar ²

Affiliations
1 Department of Computer Science, Noorul Islam University, Kanyakumari - 629180, Tamil Nadu, IN
2 Electronics and Instrumentation, Noorul Islam University, Kanyakumari - 629180, Tamil Nadu, IN

Abstract

In medical image processing, Brain tumor segmentation from MRI scan slice without human intervention has become one of the most challenging area in research.MR image slices usually contain a significant amount of noise caused by operator interactions, environmental or external factors, machines used etc, which in turn may cause serious segmentation inaccuracy. Our main aim is to recognize a tumour and its quantification from a specific MRI scan of a brain image to obtain the best segmentation in minimum time. In this paper, we are trying to develop a robust segmentation technique, by combining two segmentation algorithms, expectation maximization and level set. This algorithm framework composed of three stages. 1. Pre-processing is used for background separation of brain; it can be also called as pre-segmentation method 2. Abnormal regions in brain is detected by using Expectation Maximization (EM) algorithm and in 3. A level set method to sharpen the segmented EM output to get sharp and accurate boundaries. At the end of the process, the tumour region is extracted from the MR images and its exact position and shape is determined with minimum time. The Experimental results clearly define the effectiveness of our approach in its accuracy and computation time when compared with other EM based method.

Keywords

Expectation-Maximization, Feature Extraction, Image Segmentation, Level Set, Pre-processing

Full Text

   

Devi Devapal ¹, S. S. Kumar ²

Affiliations
1 College of Engineering, Pathanapuram – 689696, Kerala, IN
2 Noorul Islam Centre for Higher Education, Kumaracoil – 629180, Tamil Nadu, IN

Abstract

Objectives: Synthetic Aperture RADAR (SAR) is a satellite imaging technology which is affected by speckle noise having granular pattern. Speckle is multiplicative noise and occurs due to the interference of the signal with the backscattered echoes. Methods/Analysis: Speckle degrades the image quality and makes further segmentation and classification of images difficult. Despeckling can be done in spatial and transform domain. In this paper the various transform domain despeckling techniques like wavelet, shearlet, contourlet and curvelet are compared. The results are analyzed using performance parameter like ECF, SSIM and ENL. Findings: Comparison of the various methods is done by using synthetic images and real images. ECF and SSIM are used to evaluate synthetic images and ENL is used for real images. In the case of real images the ENL value is highest for curvelet compared to wavelet, shearlet and contourlet. In the case of synthetic images SSIM and ECF value is high for curvelet compared to other methods. ENL value is lowest for shearlet transform whereas SSIM and ECF values are lowest for contourlet transform. From the results, it can be concluded that curvelet outperforms all the other methods. Novelty/ Improvements: Transform Domain techniques has got wide spread applications in the field of denoising, segmentation and classification. Using curvelet for despeckling can improve the further segmentation and classification process.

Keywords

ECF, ENL, SAR, SSIM, Speckle

Full Text

   

P. Y. Muhammed Anshad ¹, S. S. Kumar ²

Affiliations
1 Department of Electronics and Communication, Noorul Islam University, Kanyakumari – 629180, Tamil Nadu, IN
2 Department of Electronics and Instrumentation, Noorul Islam University, Kanyakumari – 629180, Tamil Nadu, IN

Abstract

Chondroblastoma is a type of benign bone tumor originates from chondroblasts. The rate of misdiagnose is much higher among benign bone tumors. The treatment method is different in both cases. Computer Aided Diagnosis (CAD) system helps more for reducing the rate of misdiagnosis. Segmentation is used for the selection of area of interest from medical images. The possible ways of diagnosis is X-Ray, CT, MRI and Biopsy. Active contour modeling is tested using region based approach. Various levels set and Gaussian distribution function is used for segmentation process. The proposed system separates similar intensity values of mean and different intensity values of variance. This system segments similar intensity regions as suspected area. At the end of this work, the system can extract the exact position and shape of the chondroblastoma affected region in bone.

Keywords

Chondroblastoma, Computer Aided Diagnosis, Gaussian Distribution Function, Level Set, MR Images Segmentation.

Full Text

   

Devi Devapal ¹, S. S. Kumar ², Christy Jojy ³

Affiliations
1 College of Engineering, Pathanapuram - 689696, Kerala, IN
2 Noorul Islam University, Tuckalay - 629180, Tamil Nadu, IN
3 Lourdes Matha College of Science and Technology, Thiruvananthapuram - 695574, Kerala, IN

Abstract

Synthetic Aperture Radar (SAR) is a satellite imaging technology which produce high resolution images and can be used in varied atmospheric conditions. SAR obtains higher resolution remote sensing images from broad areas of terrain. It can be used day and night under all weather conditions. Unwanted data which affects the actual details of the image is called as noise. Inherent property of SAR is Speckle Noise of multiplicative nature which degrades SAR image quality. The backscattered echoes from objects undergo interference constructively and destructively which results in speckled image. The process of removing speckle noise is called as despeckling. The essential feature of despeckling technique is to effectively filter homogeneous areas while preserving textures and edges. The literature on despeckling methods for SAR images in both spatial and transform domain has been reviewed in this paper in chronological order.

Keywords

Despeckling, Speckled Image, Speckle Noise, Synthetic Aperture Radar

Full Text