A Time Efficient Clustering Algorithm for Gray Scale Image Segmentation

Performance Measures

Determining the fineness of segmentation results, minimizing the noise factor present in the segmented image and performing the segmentation in a least amount of time are some of the factors that affect the outcome of segmentation. In this section, we discuss briefly the following quality measures we have used in our work that affect the performance of gray scale image segmentation using clustering approach.

Structural Content (SC)

The value of SC (Jaskirat et al., 2012) to a great extent influences the quality of a segmented image. SC measure is given by:

where in(i, j) is the input image and seg(i, j) is the target segmented image and m & n are image matrix rows and columns respectively.

A smaller value of SC means that the image is of better quality.

Root Mean Square Error (RMSE)

Generally speaking, RMSE (Jaskirat et al., 2012) gives the difference between the values predicted by a model and the values actually present in that model. In term of image processing, it corresponds to the amount of deviation present in the output segmented image as compared to that in the original input image. RMSE measure is given by:

A smaller value calculated for RMSE means that the image is of good quality.