Ant Colony Alpha Matte: A New Approach for Natural Image Matting

In this paper, we present an interactive algorithmto separate foreground and background regions of naturalimages (natural image matting) using ant colony optimization.Today, image matting is one of the most challenging andinteresting research fields in image processing. In our approachinstead of preparing a trimap, the user specifies foreground andbackground regions by some red and blue scribbles. Then byminimizing local energy function of all pixels alpha matte isestimated. Our approach not only needs a little interaction withthe user but also by applying ant colony algorithm on colorimages, finds homogenous regions of the image and yields goodresults compared with other methods. In other words, the localenergy of a pixel is obtained using traveled path by the pixel antand since the ant tends to move to pixels similar to beginningpixel, homogenous regions of the image are detected. Moreover,we use some techniques like vectorization in the implementationof our algorithm in order to decrease time complexity.Experimental results show our algorithm advantages

کلیدواژه ها:

Image matting ، ant colony ، energy function vectorization ، scribbles