@inproceedings {pub4701,
	title = {Voxel-based Three-dimensional Neural Style Transfer},
	author = {Timo Friedrich AND Barbara Hammer AND Stefan Menzel},
	year = {2021},
	month = {September},
	abstract = { In recent years, Neural Style Transfer has been successfully applied in the crea-tive process for generating novel artistic 2D images by transferring the style of a painting to an existing content image. These techniques which rely on deep neural networks have been extended to further computational creativity tasks like video, motion and animation stylization. However, only few research has been conduct-ed to utilize Neural Style Transfer in the spatially three-dimensional space. Exist-ing 2D/3D hybrid approaches avoid the extra dimension during the stylization process and add postprocessing or differentiable rendering to transform the re-sults to 3D. In this paper, we propose for the first time a complete three-dimensional Neural Style Transfer pipeline based on a high-resolution voxel rep-resentation. Following our previous research, our architecture includes the stand-ardized gram matrix style loss for noise reduction and visual improvement, the bipolar exponential activation function for symmetric feature distributions and best practices for the underlying classification network configuration. In addition, we propose regularization terms for voxel-based 3D Neural Style Transfer opti-mization and demonstrate their capability to significantly reduce noise and unde-sired artefacts. We apply our 3D Neural Style Transfer pipeline on a set of style targets in combination with an organic and industrial content shape. Both style transfer results are evaluated using 3D shape descriptors which confirm the sub-jective visual improvements.},
	publisher = {Springer, LNCS, LNAI, LNBI},
	booktitle = {International Work-Conference on Artificial Neural Networks}
}