Information Technology Reference
In-Depth Information
Since the target image is the layout constraint of the input, the output should
follow its feature information, such as color, structure and luminance etc. In Fig. 5,
we compare synthesis results generated by different methods. Clearly, the proposed
method captures much better structural layout for composite textures.
(a)
(b)
(c)
(d)
Fig. 5
Each column from top to bottom shows (a) the input texture, (b) result of image
quilting without finding the minimum boundary cut, (c) result of image quilting and (d) result
of the proposed algorithm, respectively.
