Newspaper Magnification with Preserved Entry Points

Sebastian Gallardo (1) , María Cristina Riff (2) , Dorian Mazauric (3) , Pierre Kornprobst (1)

1 BIOVISION - Biologically plausible Integrative mOdels of the Visual system : towards synergIstic Solutions for visually-Impaired people and artificial visiON
2 UTFSM - Universidad Tecnica Federico Santa Maria [Valparaiso]
3 ABS - Algorithms, Biology, Structure

We present a new task called "Newspaper magnification with preserved entry points", which consists in making the print edition more readable and usable in small displays by preserving the quality of entry points (e.g., headlines, images, illustrations), i.e., all elements that allow to scan a page and focus in the article they want. Pinch-to-zoom is the classical feature available to magnify the content. However, the pinch-to-zoom feature has a well-known drawback: one quickly loses positioning inside the page and thus can not scan it anymore, significantly impacting reading comfort and navigation. To solve this problem, we propose a formulation that preserves entry points during magnification while keeping all stories on a page always visually present. Our approach proposes a new packing problem to generate alternative layouts considering computational aesthetics criteria, which we solve by a genetic algorithm. As a proof-of-concept, we focus here on text-only content and headlines readability and usability in this paper. Results show that our approach improves entry point quality in the generated layouts compared

Keywords: Newspapers; document transformation; layout generation; readability; aesthetics; evolutionary optimization

Domains: Neural and Evolutionary Computing [cs.NE]; Document and Text Processing

Sebastian Gallardo, María Cristina Riff, Dorian Mazauric, Pierre Kornprobst. Newspaper Magnification with Preserved Entry Points. 2023. ⟨hal-04210840⟩