Please use this identifier to cite or link to this item:
|Title:||MRI image segmentation and volume extraction for clinical study||Authors:||Kallam Hanimi Reddy||Keywords:||DRNTU::Engineering::Electrical and electronic engineering||Issue Date:||2009||Abstract:||Image segmentation aims to separate objects of interests from the background in an image. It has an important role in image processing, with a variety of applications such as image visualization, object detection, and volumetric measurement. In clinical applications, tumor volume measurement plays an important role in which physicians need to quantify tumor growth over a period of time. This volume measurement is done by segmenting the tumor from its surroundings in a sequence of 2D images. Due to the fact that growing number of patients or increase in medical data, manual segmentation becomes a difficult task as it needs more time and the accuracy of segmentation depends on the expertise. For this reason, we need automatic or semi-automatic techniques for segmentation and analysis of medical images. This work has been focusing on automatic segmentation of the tumor in Magnetic Resonance (MR) images and its quantitative volume measurement, which depends strongly on the accuracy of the segmentation result. Accurate segmentation can be achieved by using an improved level set method which utilizes the image information in local regions for doing segmentation slice-by-slice. The resultant volume is compared with the ground truth volume obtained by manual segmentation. The performance evaluation is done by using statistical measures such as positive predictive value, Jaccard similarity index, and tumor volume accuracy measurement and the results are promising.||URI:||http://hdl.handle.net/10356/42767||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Theses|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.