Abstract :
The images are partitioned into numerous mechanisms, thus every mechanism is meaningful, and this is
image segmentation. Given that an image is poised of a number of pixels, the segmentation problem can be treated
as a labeling problem which aims to assign each pixel a label demonstrating an exacting component in the scene. On
the other hand, the segmentation task can be considered as extracting the boundaries between diverse objects, so that
the image is partitioned into meaningful regions according to the boundaries. This paper proposes a new
segmentation technique that combines Dual Tree Complex Wavelet Transform (DTCWT) decomposition with the
watershed transform for an X-Ray medical image. The DTCWT is applied to the intensity image, producing detail
and approximation coefficients. If simply watershed algorithm be used for segmentation of image, then it will have
overlap in segmentation. To resolve this, we used the proposed approach which combines DTCWT and watershed
algorithm. First we used the DTCWT to produce additional medical images, then watershed algorithm is applied for
segmentation of the approximation image, then by using the inverse DTCWT, the segmented image is projected. The
results demonstrate that combining DTCWT and watershed transform can help us to get the very high accuracy
segmentation, even for noisy and satellite images.
Keyword :
X-Ray Imaging, Dual Tree Complex Wavelet Transform, Marker Controlled Watershed Algorithm, Image segmentation