This work describes a new algorithm for automatic layout of variable content documents. The aim is to demonstrate that high-quality page designs can be efficiently obtained from dynamic and variable content (such as content from the internet, sales brochures, bills, and others), which is a current trend in the production of personalized documents. The approach described here, contrary to other works published in this field, does not make use of approximative methods (such as genetic algorithms) to assign regions of a page to the contents of a document. The choice of a deterministic algorithm was made due to the performance and reliability requirements in the context of Variable Data Printing, which is one of the applications of this work. The method works by making successive divisions on a page, allocating regions according to the areas required by each element (content), instead of constraining their geometries as in previous works in this field. The choice of how to divide a page is based on a simple heuristic, capable of preserving some important document design criteria such as alignment, homogeneity, etc. as well as the reading order provided by the input and the possibility of grouping related items. Besides that, the algorithm is also capable of distributing content in a document that is composed by several pages. As a case study, the algorithm was developed to produce document layouts that are similar to newspapers covers, both in design and style. Finally, this work presents concepts regarding adaptive documents, a quick survey of the most relevant works in this field, a detailed description of the algorithm that was developed, as well as an empirical evaluation of the results, regarding qualitative aspects from the generated documents and, to a lesser extent, the performance of the proposed algorithm.
展开▼
