Cut-out image mosaics

To be presented at the 6th Symposium on Non-Photorealistic Animation and Rendering (NPAR 2008)


Jeff Orchard, Craig Kaplan

David R. Cheriton School of Computer Science

University of Waterloo

An image mosaic is a rendering of a large target image by arranging a collection of small source images, often in an array, each chosen specifically to fit a particular block of the target image.  Most mosaicking methods are simplistic in the sense that they break the target image into regular tiles (e.g., squares or hexagons) and take extreme shortcuts when evaluating the similarity between target tiles and source images.  In this paper, we propose an efficient method to obtain higher quality mosaics that incorporate a number of process improvements.  The Fast Fourier Transform (FFT) is used to compute a more fine-grained image similarity metric, allowing for optimal colour correction and arbitrarily shaped target tiles.  In addition, the framework can find the optimal sub-image within a

source image, further improving the quality of the matching.  The similarity scores generated by these high-order cost computations are fed into a matching algorithm to find the globally-optimal assignment of source images to target tiles.  Experiments show that each improvement, by itself, yields a more accurate mosaic.  Combined, the innovations produce very high quality image mosaics, even with only a few hundred source images.


(above) Cut-out image mosaic involving various target tile shapes, colour correction, and the ability to select sub-images from images in the source database.  The target and source images were taken from the Library of Congress collection on flickr.

Some of our higher-resolution image mosaics on flickr

Download the paper... do it!