UNIGE document Chapitre d'actes
previous document  unige:123330  next document
add to browser collection
Title

Transferring Neural Representations for Low-Dimensional Indexing of Maya Hieroglyphic Art

Authors
Can, Gulcan
Hu, Rui
Gayol, Carlos Pallán
Krempel, Guido
Spotak, Jakub
Odobez, Jean-Marc
show hidden authors show all authors [1 - 9]
Published in Computer Vision – ECCV 2016 Workshops. Amsterdam (The Netherlands) - 8-10 and 15-16 October 2016 - Cham: Springer International Publishing. 2016, p. 842-855
Abstract We analyze the performance of deep neural architectures for extracting shape representations of binary images, and for generating low-dimensional representations of them. In particular, we focus on indexing binary images exhibiting compounds of Maya hieroglyphic signs, referred to as glyph-blocks, which constitute a very challenging dataset of arts given their visual complexity and large stylistic variety. More precisely, we demonstrate empirically that intermediate outputs of convolutional neural networks can be used as representations for complex shapes, even when their parameters are trained on gray-scale images, and that these representations can be more robust than traditional handcrafted features. We also show that it is possible to compress such representations up to only three dimensions without harming much of their discriminative structure, such that effective visualization of Maya hieroglyphs can be rendered for subsequent epigraphic analysis.
Keywords Shape retrievalNeural networksDimensionality reduction
Identifiers
ISBN: 978-3-319-46603-3
Full text
Structures
Research groups Computer Vision and Multimedia Laboratory
Viper group
Citation
(ISO format)
ROMAN RANGEL, Edgar Francisco et al. Transferring Neural Representations for Low-Dimensional Indexing of Maya Hieroglyphic Art. In: Computer Vision – ECCV 2016 Workshops. Amsterdam (The Netherlands). Cham : Springer International Publishing, 2016. p. 842-855. doi: 10.1007/978-3-319-46604-0_58 https://archive-ouverte.unige.ch/unige:123330

224 hits

308 downloads

Update

Deposited on : 2019-09-18

Export document
Format :
Citation style :