Mammography Image Segmentation: Chan-Vese Active Contour and Localised Active Contour Approach

<p>Breast cancer is one of the most common diseases diagnosed among female cancer patients. Early detection of breast cancer is needed to reduce the risk of fatality of this disease as no cure has been found yet for this illness. This research is conducted to improve the Gradient Vector Flow (GVF) Snake Active Contour segmentation technique in mammography segmentation. Segmentation of the mammogram image is done to segment lesions existence using Chan-Vese Active Contour and Localized Active Contour. Besides that, the effectiveness of these both methods are then compared and chosen to be the best method. Digital Database of Screening Mammograms (DDSM) is used for the purpose of screening. First, the images undergo pre-processing process using the Gaussian Low Pass Filter to remove unwanted noise. After that, contrast enhancement applied to the images. Segmentation of mammograms is then conducted by using Chan-Vese Active Contour and Localized Active Contour method. The result shows that Chan-Vese technique outperforms Localized Active Contour with 90% accuracy.</p>

Download Full-text

SEGMENTASI CITRA PARU-PARU MENGGUNAKAN METODE KONTUR AKTIF DENGAN VALIDASI ROC

Komunikasi Fisika Indonesia ◽

10.31258/jkfi.16.2.91-95 ◽

2019 ◽

Vol 16 (2) ◽

pp. 91

Author(s):

Sintha Syaputri ◽

Zulkarnain Zulkarnain

Keyword(s):

Active Contour ◽

Operating Characteristic ◽

Outer Edge ◽

Contour Method ◽

Body Parts ◽

X Ray ◽

Average Percentage ◽

Contour Segmentation ◽

Active Contour Method ◽

Chest X Ray

Segmentation is the process of separating parts of objects from the background by dividing images that have different object intensities with each other such as in imaging of body parts. Active contour segmentation was used for medical imaging that resistant to noise around objects. This study used 5 chest X-Ray images, specifically to the lungs with a grayscale format measuring 256 x 256 pixels, through the preprocessing process and filtering a Gaussian filter, each image was inputted to the R2015a version of the matlab GUI program. Then the segmentation had done by using the active contour method. In this method a curve in the form of a small circle was placed on the edge of object to be segmented. The curve will move according to the shape of the outer edge of the lung based on the values of active contour parameters such as Alpha, Beta, Gamma, Kappa, WEline, WEdge, WEterm and Iteration. Validation was done by using the ROC (Receiver Operating Characteristic) method and were obtained an average percentage with an accuracy value of 96.26%, a specificity of 96.47% and a sensitivity of 76.54%.

Download Full-text

Feature Based Active Contour Method for Automatic Detection of Breast Lesions Using Ultrasound Images

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.573.471 ◽

2014 ◽

Vol 573 ◽

pp. 471-476 ◽

Cited By ~ 3

Author(s):

Telagarapu Prabhakar ◽

S. Poonguzhali

Keyword(s):

Breast Cancer ◽

Early Detection ◽

Active Contour ◽

Automatic Segmentation ◽

Ultrasound Image ◽

Contour Method ◽

Speckle Reduction ◽

Initial Contour ◽

Ultrasound Images ◽

Active Contour Method

Breast cancer has been increasing over the past three decades. Early detection of breast cancer is crucial for an effective treatment. Mammography is used for early detection and screening. Especially for young women, mammography procedures may not be very comfortable. Ultrasound has been one of the most powerful techniques for imaging organs and soft tissue structure in the human body. It has been used for breast cancer detection because of its non-invasive, sensitive to dense breast, low positive rate and cheap cost. But due to the nature of ultrasound image, the image suffers from poor quality caused by speckle noise. These make the automatic segmentation and classification of interested lesion difficult. One of the frequently used segmentation method is active contour. If this initial contour of active contour method is selected outside the region of interest, final segmentation and classification would be definitely incorrect. So, mostly the initial contour is manually selected in order to avoid incorrect segmentation and classification. Here implementing a method which was able to locate the initial contour automatically within the multiple lesion regions by using the wavelet soft threshold speckle reduction method, statistical features of the lesion regions and neural network and also we are able to automatically segment the lesion regions properly. This will help the radiologist to identify the lesion boundary automatically.

Download Full-text