PROJECT TITLE :

Salient Region Detection via High-Dimensional Color Transform and Local Spatial Support

ABSTRACT:

In this paper, we tend to introduce a completely unique approach to automatically detect salient regions in an image. Our approach consists of worldwide and local options, which complement every different to compute a saliency map. The first key idea of our work is to create a saliency map of a picture by employing a linear combination of colors during a high-dimensional color area. This can be based mostly on an observation that salient regions often have distinctive colors compared with backgrounds in human perception, but, human perception is sophisticated and highly nonlinear. By mapping the low-dimensional red, inexperienced, and blue color to a feature vector in an exceedingly high-dimensional color space, we tend to show that we have a tendency to can composite an correct saliency map by finding the optimal linear combination of color coefficients in the high-dimensional color space. To additional improve the performance of our saliency estimation, our second key idea is to utilize relative location and color distinction between superpixels as features and to resolve the saliency estimation from a trimap via a learning-primarily based algorithm. The further local features and learning-primarily based algorithm complement the worldwide estimation from the high-dimensional color remodel-primarily based algorithm. The experimental results on three benchmark datasets show that our approach is effective compared with the previous state-of-the-art saliency estimation strategies.