Based on Fig. 6, the segmentation in 30 × 40 image gives
a computation time of 3.6 seconds, while 60 × 80 image
gives a computation time of 118.1 seconds which is almost 2
minutes. Although the difference of these two image sizes is
not large, their computation time can differ greatly, which is
about 114.5 seconds difference. It can be said that using
normalised cuts algorithm for image segmentation on image with large size may severely hamper the segmentation
efficiency. This shows that large image segmentation using
normalised cuts algorithm may not meet the requirement of
certain computer vision application in reality.
To retain the use of normalised cuts algorithm but done
in an optimal way, two-stage segmentation approach is
applied here for larger image. The same image with a larger
size of 75 × 100 is processed by using the two-stage
approach segmentation which is also shown in Fig. 6. Even
though the result of the image segmentation done in twostage
approach may not be as good as the previous two
images done with the ordinary approach, it is still able to
segment out the important objects although it is lacking of
some extend of accuracy in the segmentation.