Image Processing Reference
In-Depth Information
tributes. According to the indiscernibility relation concept, the mammogram is divided into
two sub-images:
(1)
(2)
where P is the gradient threshold, Q is the noise threshold, f ( i , j ) is the gradient value calcu-
lated from the label-removed image, “ s ” denotes the sub-block. Considering s mn as each pixel
examine its neighbors to decide whether it is noise or not. If it is, then eliminate the noise by
replacing the pixel value with Q . The sub-images that need to be enhanced are defined as fol-
lows:
(3)
(4)
Next, we enhance I 1 and I 2 , respectively, and get the final image by merging the sub-images:
I 2 is enhanced by histogram equalization method and I 1 is transformed below:
(5)
Here, we set ρ = γ = 1.5. After enhancement, the boundary contrast between the mass and
surrounding tissue becomes more obvious.
Search WWH ::




Custom Search