Hexagonal image representation for 3-D photorealistic reconstruction

Sato, Hidenori; Matsuoka, Hiroto; Onozawa, Akira; Kitazawa, Hideaki

doi:10.1109/icpr.2002.1048391

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hexagonal lattice is also used for medical sensors, such as DEPFET (Neeser et al, 2000) or retina implant system (Schwarz et al, 1999). Moreover, hexagonal lattice is a well-known and studied grid (Sato et al, 2002;Shima et al, 2010;Asharindavida et al, 2012;Hoogeboom et al, 2018) and offers advantages compared to square lattice (Middleton et al, 2001) such as higher sampling density and a better representation of curves. In addition, some more benefits have been shown by (Sousa, 2014;He and Jia, 2005;Asharindavida et al, 2012) such as equidistant neighborhood, clearly defined connectivity, smaller quantization error.…”

Section: Application Example: the Hexagonal Casementioning

confidence: 99%

Indexed Operations for Non-rectangular Lattices Applied to Convolutional Neural Networks

Jacquemont

Antiga

Vuillaume

et al. 2019

Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Application

View full text Add to dashboard Cite

The present paper introduces convolution and pooling operators for indexed images. These operators can be used on images that do not provide Cartesian grids of pixels, as long as a list of neighbor's indices can be provided for each pixel. They are foreseen being useful for convolutional neural networks (CNN) applied to special sensors, especially in science, without requiring image pre-processing. The present work explains the method and its implementation in the Pytorch framework and shows an application of the indexed kernels to the classification task of images with hexagonal lattices using CNN. The obtained results show that the method gives the same performances as the standard convolution kernels. Indexed convolution thus makes deep neural network frameworks more general and capable of addressing unconventional image lattices. The current implementation, as well as code to reproduce the experiments described in this paper are made available as open-source resources on the repository www.github.com/IndexedConv.

show abstract