Joint Map and Symmetry Synchronization

Sun, Yifan; Liang, Zhenxiao; Huang, Xiangru; Huang, Qixing

doi:10.1007/978-3-030-01228-1_16

Cited by 13 publications

(11 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our approach outperforms baseline approaches across all categories. On categories in the first group, our approach yields slightly better results than that of [Sun et al 2018] and [Huang et al 2014]. This is due to the fact that the fraction of correct initial maps is significant, and matrix based map synchronization techniques are already delivering good results.…”

Section: Experimental Evaluationmentioning

confidence: 88%

“…In particular, for rotational symmetric objects, our approach only yields modest results. To address this issue, it would be interesting to combine the lifting approach described in [Sun et al 2018] and the tensor formulation described in this paper. Another limitation of our approach is that the input graph 6 is favored to be a random graph or a geometric graph.…”

Section: Discussion Limitations and Future Workmentioning

confidence: 99%

“…We use 128 sample points across all experiments. The last baseline is [Sun et al 2018], which leverages the Kronecker product operator to synchronize maps among symmetric objects. As a by-product, it also promotes consistent correspondence pairs across the input shapes, which ultimately boost the quality of synchronized shape maps.…”

Section: Experimental Evaluationmentioning

confidence: 99%

“…Quantitative evaluation of our tensor map approach and prior baseline approaches on 11 categories of the SHREC07 dataset [Giorgi et al 2007]. We compare against four baseline approaches: Cosmo17 [Cosmo et al 2017], Nguyen11 [Nguyen et al 2011], Huang14 [Huang et al 2014], and Sun18 [Sun et al 2018]. We show both (Left) results without factoring out the underlying symmetry and (Right) results after factoring out the underlying symmetry.…”

Section: Figmentioning

confidence: 99%

See 3 more Smart Citations

Tensor maps for synchronizing heterogeneous shape collections

et al. 2019

Self Cite

View full text Add to dashboard Cite

Establishing high-quality correspondence maps between geometric shapes has been shown to be the fundamental problem in managing geometric shape collections. Prior work has focused on computing efficient maps between pairs of shapes, and has shown a quantifiable benefit of joint map synchronization, where a collection of shapes are used to improve (denoise) the pairwise maps for consistency and correctness. However, these existing map synchronization techniques place very strong assumptions on the input shapes collection such as all the input shapes fall into the same category and/or the majority of the input pairwise maps are correct. In this paper, we present a multiple map synchronization approach that takes a heterogeneous shape collection as input and simultaneously outputs consistent dense pairwise shape maps. We achieve our goal by using a novel tensor-based representation for map synchronization, which is efficient and robust than all prior matrix-based representations. We demonstrate the usefulness of this approach across a wide range of geometric shape datasets and the applications in shape clustering and shape co-segmentation.

show abstract

Section: Experimental Evaluationmentioning

confidence: 88%

Section: Discussion Limitations and Future Workmentioning

confidence: 99%

Section: Experimental Evaluationmentioning

confidence: 99%

Section: Figmentioning

confidence: 99%

See 2 more Smart Citations

Tensor maps for synchronizing heterogeneous shape collections

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…These methods formulate transformation synchronization as low-rank matrix recovery, where the input relative transformations are considered noisy measurements of this low-rank matrix. In the literature, people have proposed convex optimization [47,23,24,13], non-convex optimization [11,53,33,26], and spectral techniques [31,25,39,38,42,44,7,2,9] for solving various low-rank matrix recovery formulations. Com-pared with the first category of methods, the second category of methods is computationally more efficient.…”

Section: Related Workmentioning

confidence: 99%

Learning Transformation Synchronization

Huang

Liang

Zhou

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Self Cite

View full text Add to dashboard Cite

Reconstructing the 3D model of a physical object typically requires us to align the depth scans obtained from different camera poses into the same coordinate system. Solutions to this global alignment problem usually proceed in two steps. The first step estimates relative transformations between pairs of scans using an off-the-shelf technique. Due to limited information presented between pairs of scans, the resulting relative transformations are generally noisy. The second step then jointly optimizes the relative transformations among all input depth scans. A natural constraint used in this step is the cycle-consistency constraint, which allows us to prune incorrect relative transformations by detecting inconsistent cycles. The performance of such approaches, however, heavily relies on the quality of the input relative transformations. Instead of merely using the relative transformations as the input to perform transformation synchronization, we propose to use a neural network to learn the weights associated with each relative transformation. Our approach alternates between transformation synchronization using weighted relative transformations and predicting new weights of the input relative transformations using a neural network. We demonstrate the usefulness of this approach across a wide range of datasets.

show abstract