Home
The Evolutionary Poster Composer (evo-poster) is a computational design system for the automatic generation of poster designs based on a Genetic Algorithm.
This way, the purposed system computationally emulates the workflow of traditional letterpress print houses in the 19th century. The typesetting method of these print houses was slightly different from the typical contemporary Graphic Design processes. At the time, letterpress designers created posters while trying to fill a matrix, often in collaboration with the client. The design decisions of these designers were very pragmatic: extensive sentences were composed in condensed typefaces, and short sentences were composed in extended typefaces. Also, the most important parts of the content were emphasized.
Evo-poster is designed to enable the autonomous generation of poster design with exploration, automation, and visualization purposes. From a given content (including text and imagery), the purposed system can generate and evolve poster designs taking into consideration aesthetically and semantics features of the content. The system is also able to identify the most important parts of content and highlight them on the outputs.
The designed users' stories and scenarios clarify the archetypical uses for this system.
evo-poster do this by:
- using an emotional-recognition approach able to identify the emotions related to input text (either to global level or line) and identify the most important parts of the given considering, that the most important parts of the content were that have more emotions related to it;
- generating poster designs__ (composed of imagery and text) and allowing the modification of their visual features in a dedicated interface;
- evolve the visual features of posters__ taking into consideration a set of typographic aesthetical measures and semantics of text.
The development of the proposed system comprises the development of four modules:
- Textual Analysis Module (responsible for the emotional recognition and sentence tokenizer of input text);
- Generator Module (responsible for the generation of poster designs);
- Evaluator Module (responsible for the evaluation of poster designs);
- Evolutionary Module (responsible for the emotional recognition and sentence tokenizer of input text).
The following image presents a tentative workflow for the system (v160122). One may access to editable version here.
- Development and evaluation of the Textual Analysis module;
- Development and evaluation of the Generator module;
- Development and evaluation of the Evaluator module;
- Development and evaluation of the Evolutionary module.