Coarse features, such as scene composition and subject together with fine details, such as strokes and line styles, are useful clues for painter and style categorization. In this work, to automatically predict paintingâs artist and style, we propose a novel deep multibranch neural network, where the different branches process the input image at different scales to jointly model the fine and coarse features of the painting. Experiments for both artist and style classification tasks are performed on the challenging Painting-91 dataset, that includes 91 different painters and 13 diverse painting styles. Our method outperforms the best method in the state of the art by 14.0% and 9.6% on artist and style classification respectively.
Bianco, S., Mazzini, D., Schettini, R. (2017). Deep multibranch neural network for painting categorization. In 19th International Conference on Image Analysis and Processing, ICIAP 2017; Catania; Italy; 11-15 September 2017 (pp.414-423). Springer Verlag [10.1007/978-3-319-68560-1_37].
Deep multibranch neural network for painting categorization
Bianco, S
;Mazzini, D;Schettini, R.
2017
Abstract
Coarse features, such as scene composition and subject together with fine details, such as strokes and line styles, are useful clues for painter and style categorization. In this work, to automatically predict paintingâs artist and style, we propose a novel deep multibranch neural network, where the different branches process the input image at different scales to jointly model the fine and coarse features of the painting. Experiments for both artist and style classification tasks are performed on the challenging Painting-91 dataset, that includes 91 different painters and 13 diverse painting styles. Our method outperforms the best method in the state of the art by 14.0% and 9.6% on artist and style classification respectively.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
