HPL-2010-59Templated Recursive Image Composition
Atkins, C. Brian; Lyons, Nicholas P.; Zhang, Xuemei; Tretter, Daniel R.
Abstract: With the rapid expansions of image acquisition and consumption, there is an increasing need for solutions that help ordinary people create image composites. In most solutions today, image layouts are provided as fixed templates. Although fixed templates benefit from a familiar use model and significant design freedom, in practice fixed-template libraries often lack diversity. In a first aspect, a library may not support a particular image count or aspect ratio set; while in a second, it may not provide a particular spatial distribution. To ameliorate both these aspects, we propose a framework called Templated Recursive Image Composition. TRIC is template-based in that every layout is based on a template specification that can itself be presented as an exemplar. However, TRIC is also generative. TRIC can accommodate almost any image set, as long as there is at least one image for every template region in the template specification. Constraints ensure respect for image aspect ratios; for spacing in the layout interior; and for relative proportions and placement of sublayouts that correspond to template regions. We present a description of TRIC along with results that demonstrate its versatility.
External Posting Date: May 6, 2010 [Abstract Only]. Approved for External Publication - External Copyright Consideration
Internal Posting Date: May 6, 2010 [Fulltext]