4D Reconstruction of Blooming Flowers

Computer Graphics Forum 2016 

Qian Zheng1,3+        Xiaochen Fan2+        Minglun Gong4       Andrei Sharf5        Oliver Deussen2,3       Hui Huang1,2*   

1Shenzhen University          2SIAT         3University of Konstanz        4Memorial University of Newfoundland         5Ben Gurion University 

+Joint First Authors      *Corresponding Author


Figure 1:Reconstruction of a blooming Orchid from a noisy and incomplete point cloud sequence. Note that smaller models representing early stages of the blooming (left ones on the bottom row) are scaled up for a better visualization.


Abstract 

Flower blooming is a beautiful phenomenon in nature as flowers open in an intricate and complex manner whereas petals bend, stretch and twist under various deformations. Flower petals are typically thin structures arranged in tight configurations with heavy self-occlusions. Thus, capturing and reconstructing spatially and temporally coherent sequences of blooming flowers is highly challenging. Early in the process only exterior petals are visible and thus interior parts will be completely missing in the captured data. Utilizing commercially available 3D scanners, we capture the visible parts of blooming flowers into a sequence of 3D point clouds. We reconstruct the flower geometry and deformation over time using a template-based dynamic tracking algorithm. To track and model interior petals hidden in early stages of the blooming process, we employ an adaptively constrained optimization. Flower characteristics are exploited to track petals both forward and backward in time. Our methods allow us to faithfully reconstruct the flower blooming process of different species. In addition, we provide comparisons with state-of-the-art physical simulation-based approaches and evaluate our approach by using photos of captured real flowers.


Figure 2: The blooming sequence of a Golden Lily. At early stages, the petals have complex shapes and packed in a small space. Hence, they are poorly represented in the captured point clouds (top). Nevertheless, our approach recovers the full model sequence (bottom).


Figure 3: An opening Water Lily consists of narrow and thin petals. As a result, it is hard to separate different petals from the scan data (top). Our approach is able to reconstruct the complete sequence (bottom) through tracking a template model.


Figure 4: Comparison between our reconstructed models (top) and captured image sequence (bottom) of a blooming Lily flower. Note that different viewpoints are used at different blooming stages and the filaments are not modeled in our approach.Figure 4: Comparison between our reconstructed models (top) and captured image sequence (bottom) of a blooming Lily flower. Note that different viewpoints are used at different blooming stages and the filaments are not modeled in our approach.


Figure 5: Side-by-side comparison between our modeling results (top) and the results from the state-of-the-art flower blooming simulation (bottom). While the simulated results look more regular, our data-driven approach captures botanic reality.


Reference

The point cloud sequences and reconstructed mesh sequences are provided here. The point cloud is .pcd format, and the mesh is .obj format.
If you find the data useful in your research, please cite this paper. 

Download from below..


Acknowledgments

We would like to thank the anonymous reviewers for their constructive comments. This work was supported in part by NSFC (61522213, 61379090, 61331018), 973 Program (2015CB352501), Guangdong Science and Technology Program (2015A030312015, 2014B050502009, 2014TX01X033, 2016A050503036), Shenzhen Innovation Program (JCYJ20151015151249564), National Foreign 1000 Plan (WQ201344000169), Guangdong Leading Talents Plan (00201509) and NSERC (293127). 


BibTex

@ARTICLE{Flower2016,
    title = {4D Reconstruction of Blooming Flowers},
    author = {Qian Zheng, Xiaochen Fan, Minglun Gong, Andrei Sharf, Oliver Deussen, Hui Huang},
    journal = {Computer Graphics Forum},
    volume = {},
    issue = {},
    pages = {},
    year = {2016}
}
Copyright © 2016-2018 Visual Computing Research Center