Cube Art

SIGGRAPH Asia 2016 Posters

teaser image


In this study, we propose Cube Art, a representation of a 3D model assembled from same cubes. Here the term same is defined as two cubes that are regarded as the same, in that they can be rotated in space such that the textures (or the Dice numbers) on the corresponding sides become equal. Since such a cubic model lacks most of original geometric properties, the model has poorer expressive- ness than the original one. To enhance the resulting shading to be more expressive, each cube needs to be properly positioned such that the textures on each face of the cube approximate the shading of the original models. The goal of this study is to determine the positioning of each cube so that the textures on each face of the cubes best approximate the original shading model under face visibility constraints of the model’s orthogonal directions. We measure the similarities between each face’s texture of the cubes and the textures resulting from the orthogonal projections onto a hexahedron of the original shading model for each cube position, and choose the best cube position. Our results show that properly rotated cubes enhance the perception of the shading and its aesthetics.


Dice Art (Dragon)
Transformation Sequence



This work was supported by JSPS KAKENHI Grant number 16K12433.


  title = {Cube Art},
  author = {Kita, Naoki and Miyata, Kazunori},
  booktitle = {SIGGRAPH ASIA 2016 Posters},
  series = {SA '16},
  year = {2016},
  isbn = {978-1-4503-4540-8},
  location = {Macau},
  year = {2016},
  pages = {30:1--30:1},
  articleno = {30},
  numpages = {1}
  url = {},
  doi = {10.1145/3005274.3005305},