The label pixels in the image are converted to the binary image with some unwanted object. Label filtering is applied to remove isolated pixels as noise by using the concept of the connected pixels. As per connected regions correspond to individual objects, we need to identify separate connected regions. Label filtering tries to isolate the individual objects by using eight-connected neighborhood algorithm and label propagation [22]. Once the algorithm is completed, only the resulting classes exceeding a certain number of pixels are labeled as gingival. Fig. 6 shows the image after label filtering.