Adversarial Synthesis of Retinal Images from Vessel Trees

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2017

Autores

Costa, P (Autor) Outra; Galdran, A (Autor) Outra; Meyer, MI (Autor) Outra; Ana Maria Mendonça (Autor) FEUP; Aurélio Campilho (Autor) FEUP

Ata de Conferência Internacional

Título: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Páginas: 516-523

14th International Conference on Image Analysis and Recognition, ICIAR 2017

5 July 2017 through 7 July 2017

Indexação

ISI Web of Knowledge - 10 Citações

Scopus - 11 Citações

Outras Informações

ID Authenticus: P-00M-X2M

DOI: 10.1007/978-3-319-59876-5_57

Abstract (EN): Synthesizing images of the eye fundus is a challenging task that has been previously approached by formulating complex models of the anatomy of the eye. New images can then be generated by sampling a suitable parameter space. Here we propose a method that learns to synthesize eye fundus images directly from data. For that, we pair true eye fundus images with their respective vessel trees, by means of a vessel segmentation technique. These pairs are then used to learn a mapping from a binary vessel tree to a new retinal image. For this purpose, we use a recent image-to-image translation technique, based on the idea of adversarial learning. Experimental results show that the original and the generated images are visually different in terms of their global appearance, in spite of sharing the same vessel tree. Additionally, a quantitative quality analysis of the synthetic retinal images confirms that the produced images retain a high proportion of the true image set quality. © Springer International Publishing AG 2017.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 8