Evaluation Metrics for Automated Typographic Poster Generation

Sérgio M. Rebelo,J. J. Merelo,João Bicker,Penousal Machado

2024-02-10

Abstract:Computational Design approaches facilitate the generation of typographic design, but evaluating these designs remains a challenging task. In this paper, we propose a set of heuristic metrics for typographic design evaluation, focusing on their legibility, which assesses the text visibility, aesthetics, which evaluates the visual quality of the design, and semantic features, which estimate how effectively the design conveys the content semantics. We experiment with a constrained evolutionary approach for generating typographic posters, incorporating the proposed evaluation metrics with varied setups, and treating the legibility metrics as constraints. We also integrate emotion recognition to identify text semantics automatically and analyse the performance of the approach and the visual characteristics outputs.

Artificial Intelligence,Human-Computer Interaction,Multimedia

What problem does this paper attempt to address?

The paper attempts to address the challenging task of evaluating designs in the process of automatically generating typographic posters. Specifically, the authors propose a set of heuristic metrics to assess typographic design, focusing on its readability, aesthetics, and semantic features. These metrics aim to evaluate the visibility of the text, the visual quality of the design, and the effectiveness of the design in conveying the content's semantics.

To validate the effectiveness of the proposed metrics, the authors employed a constrained evolutionary approach to generate typographic posters and integrated these metrics into different settings, using readability metrics as constraints. Additionally, the authors incorporated sentiment recognition technology to automatically identify the text's semantics and analyzed the performance of this method and the visual characteristics of its output. Experimental results indicate that the proposed metrics can effectively guide the evolutionary process, producing complete and readable designs from various text inputs while considering aesthetic and semantic factors.