Region-Based Tumor Segmentation in Mammograms

It is a simple approach compared to other methods, as it involves dividing an image into different sections based on predetermined. Compared to others, it is a straightforward method because it entails separating an image into different sections based on predetermined criteria [58 (link)]. There are two primary kinds of region-based segmentation: (1) region splitting and merging and (2) region growing. Region growing allows the removal of a region from an image using defined criteria, such as intensity. It involves selecting a starting seed point. It is important to note that unlike region growing, region splitting and merging work on the entire image [59 ].
In the present study extracting the region of interest (ROI) involves using both thresholding and region-based techniques. The tumor in the INbreast dataset samples cites moreira2012inbreast is labeled by a white bounding box, as shown in Figure 4. For extracting ROI, the tumor region is first determined by setting a threshold value based on the white color pixels in the image. The threshold for all images is determined to be 80 after several attempts, independent of tumor size. After identifying the greatest area inside this threshold within the image, the tumor is automatically cropped. Figure 4 shows ROI extracted using threshold and region-based methods.
The method for extracting ROI can be summarized in four steps:

Thresholding the grayscale mammogram image to create a binary image.

Labelling and counting the binary image objects, then retaining only the largest one, which is the tumor, as defined by the white bounding box.

Assign the largest area within the threshold value to “1” and the rest a value of “0.”

Multiply binary image with original mammogram image for obtaining final ROI without including other parts of breast or artifacts.