Technical Report Number
Digital sculpting is becoming ubiquitous for modeling organic shapes like characters. Artists commonly show their sculpting sessions by producing timelapses or speedup videos. But the long length of these sessions make these visualizations either too long to remain interesting or too fast to be useful. In this paper, we present SculptFlow, an algorithm that summarizes sculpted mesh sequences by repeatedly merging pairs of subsequent edits taking into account the number of summarized strokes, the magnitude of the edits, and whether they overlap. Summaries of any length are generated by stopping the merging process when the desired length is reached. We enhance the summaries by highlighting edited regions and drawing filtered strokes to indicate artists' workflows. We tested SculptFlow by recording professional artists as they modeled a variety of meshes, from detailed heads to full bodies. When compared to speedup videos, we believe that SculptFlow produces more succinct and informative visualizations. We open source SculptFlow for artists to show their work and release all our datasets so that others can improve upon our work.
Dartmouth Digital Commons Citation
Denning, Jonathan D.; Pellacini, Fabio; and Ou, Jiawei, "SculptFlow: Visualizing Sculpting Sequences by Continuous Summarization" (2014). Computer Science Technical Report TR2014-759. https://digitalcommons.dartmouth.edu/cs_tr/362