Close
Metadata

%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2017/08.21.20.59
%2 sid.inpe.br/sibgrapi/2017/08.21.20.59.40
%T Handwriting Synthesis from Public Fonts
%D 2017
%8 Oct. 17-20, 2017
%A Balreira, Dennis Giovani,
%A Walter, Marcelo,
%@affiliation Institute of Informatics - Universidade Federal do Rio Grande do Sul
%@affiliation Institute of Informatics - Universidade Federal do Rio Grande do Sul
%E Torchelsen, Rafael Piccin,
%E Nascimento, Erickson Rangel do,
%E Panozzo, Daniele,
%E Liu, Zicheng,
%E Farias, Mylène,
%E Viera, Thales,
%E Sacht, Leonardo,
%E Ferreira, Nivan,
%E Comba, João Luiz Dihl,
%E Hirata, Nina,
%E Schiavon Porto, Marcelo,
%E Vital, Creto,
%E Pagot, Christian Azambuja,
%E Petronetto, Fabiano,
%E Clua, Esteban,
%E Cardeal, Flávio,
%B Conference on Graphics, Patterns and Images, 30 (SIBGRAPI)
%C Niterói, RJ
%S Proceedings
%I IEEE Computer Society
%J Los Alamitos
%K handwriting synthesis, public fonts.
%X Handwriting synthesis generates renderings of text which look like they were written by a human but are in fact synthesized by a model. From an input sample of the desired handwriting, we introduce an algorithm that finds the best match between characters using as source for the output text the large collection of publicly available fonts designed to look like handwriting. For each character in the desired output text, we find the best match among the public fonts using a metric that matches both the shape and appearance of the input real character. Once we have the set of best characters we build the output sentence or paragraph by concatenation of individual characters. Our results show that even though human calligraphy is highly individual and specialized, visually similar renderings are possible for many applications that do not demand full similarity. On a user study with 12 subjects, our synthesis results were considered, on average, 71% similar to the input samples.
%@language en
%3 PID4960255.pdf


Close