WireRoom: Model-guided Explorative Design of Abstract Wire Art

ACM Transactions on Graphics (Proceedings of SIGGRAPH 2021)


Zhijin Yang1    Pengfei Xu1*    Hongbo Fu2    Hui Huang1

1Shenzhen University    2City University of Hong Kong



Fig. 1. The wire shapes interactively designed with our WireRoom framework. All these wire shapes are with a single-wire structure. Bottom: the input models used for guiding the creation of these wire shapes.


Abstract

We present WireRoom, a computational framework for the intelligent design of abstract 3D wire art to depict a given 3D model. Our algorithm generates a set of 3D wire shapes from the 3D model with informative, visually pleasing, and concise structures. It is achieved by solving a dynamic travelling salesman problem on the surface of the 3D model with a multi-path expansion approach. We introduce a novel explorative computational design procedure by taking the generated wire shapes as candidates, avoiding manual design of the wire shape structure. We compare our algorithm with a baseline method and conduct a user study to investigate the usability of the framework and the quality of the produced wire shapes. The results of the comparison and user study confirm that our framework is effective for producing informative, visually pleasing, and concise wire shapes.



Fig. 3. The workflow of our framework. Given an input model, we first extract the landmark sites and the feature lines (left). Then our automatic wire generation algorithm returns a set of wire shapes (middle). The user may select the most desired one and perform subsequent edits to obtain the final wire shape (right, displayed in two views).





Fig. 6. The wire shapes generated with different combinations of a and b. The red rectangle highlights the wire shape generated with the default parameters. When a = 0, i.e., the first wire shape in the top row, the generated path approximates to the geodesic path on the unwarped surface.
Fig. 8. The wire shapes generated with different combinations of a* and b*. The red rectangle highlights the wire shape generated with the default parameters.




Fig. 7. Two wire shapes with a tree-like structure. The red rectangles highlight the middle part expansions.
Fig. 9. The top-ranked wire shapes with different values of λ. The red rectangle highlights the wire shapes generated with the default parameter.


Fig. 10. Left: the main interface of our framework. Right: the interactive tools provided by our framework. Please refer to the accompanying video for interactive sessions.




Fig. 13. Top: 10 models with different complexities were tested in Study I. These models were differentiated in 5 pairs according to the structural similarities or categories. Red spheres represent the automatically detected landmark sites. Bottom: average completion time, average general operation number, and average curve editing number for each model in the two configurations; the user evaluation of the automatic wire generation algorithm and the interactive tools. The unpaired t-test confirmed that there exist significant differences between WireRec and NonWireRec in the completion time, the general operation number, and the curve editing number for almost all models. The error bar represents the standard error of the mean.




Fig. 15. The selected wire shapes from the creations in Study I. These wire shapes received the highest scores in two configurations for each model in Study II. Top: the wire shapes created in WireRec. Bottom: the wire shapes created in NonWireRec. The participants might perform wire editing after selecting the most desired wire shape from the candidates in WireRec. More wire shapes created by the participants can be found in the supplemental materials.



Fig. 18. Gallery of 3D wire shapes produced with our framework. Each wire shape is rendered in two views. Bottom: two wire shapes are produced for each model. All these wire shapes are with a single-wire structure.



Acknowledgements

We thank the reviewers for their constructive comments, the user study participants for their time. We thank Ruth Jensen, Maria Szabo, and Bud Bullivant for allowing us to use the images of their wire artworks. This project was partially supported by NSFC (62072316), the Research Grants Council of HKSAR (CityU 11212119), the Centre for Applied Computing and Interactive Media (ACIM) of School of Creative Media, GD Talent Program (2019JC05X328), GD Science and Technology Program (2020A0505100064, 2015A030312015), DEGP Key Project (2018KZDXM058), National Engineering Laboratory for Big Data System Computing Technology, and Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ).


Bibtex

@article{WireRoom21,

title={WireRoom: Model-guided Explorative Design of Abstract Wire Art},

author={Zhijin Yang and Pengfei Xu and Hongbo Fu and Hui Huang},

journal={ACM Transactions on Graphics (Proceedings of SIGGRAPH 2021)},

volume={40},

number={4},

pages={128:1--128:13},

year={2021},

}



Downloads (faster for people in China)

Downloads (faster for people in other places)