Abstract:
Whenever we see an image the visual system focuses on the prominent features
present in an image. These prominent features are known as the salient features
which we are obtaining by generating the saliency map. As done before we are also
using the natural statistics that is measuring the saliency from data collection of
natural images.
We have used the ICA filters to generate the saliency map. Our proposed work is to
improve the blur result given by the ICA filter by using different techniques. The
results of our work are very clear. The prominent feature is improved by applying the
techniques.
The bottom up saliency plays a vital role for the basic saliency map but the difference
majorly occurs, when we improve the saliency map by applying different filters and
edge detection techniques thus resulting in object detection and improving the
performance of our saliency map.
