Comparative Study of Lung ROI Segmentation with Different Morphological Structuring Element Sizes in Noise Environment

Vol.2  No.1

Year : 2014

Issue : Jan-Mar

Title : Comparative Study of Lung ROI Segmentation with Different Morphological Structuring Element Sizes in Noise Environment

Author Name : V. Kalpana, S.Varadarajan

Synopsis :

Lung cancer is the most challenging problem chasing the scientific and medical fields in several countries around the world. Several medical modalities are specialised to diagnosis the extinct of disease via PET, MRI, CT, US and DICOM. Automated Computer Aided Diagnosing (CAD) system has shown its improvements in the detection of lung cancer using advanced radiology. CAD system merges geometric and intensity models to improve particular area of anatomical structure. Nodules are comparatively smaller than 3-4 cm in diameter and are common abnormalities that are adjacent to vessels or chest wall. Detection of non-spherical shaped nodules is a primary difficulty. Sixty percent of all nodules are not harmful but malignant nodules may be lung cancer tumours. To identify these nodules the methods like intensity thresholding or model based might fail. In this paper the nodules are extracted from the DICOM lung image in the noisy environment such as Gaussian, salt and pepper, Poisson and speckle using morphology and watershed algorithm. The nodules are extracted from the lung portion using different edge detection operators such as Gaussian, Laplacian, prewitt, LOG, Unsharp, average and Sobel in presence of noise and various sizes of the structuring element. These results support to examine and calculate the influence of noises on the DICOM images in extracting the nodules.

For more details:

imanagerpublications.com