Robust, fast and flexible symmetry plane detection based on differentiable symmetry measure

Hruda, Lukáš; Kolingerová, Ivana; Váša, Libor

doi:10.1007/s00371-020-02034-w

Cited by 12 publications

(43 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparison Results: In Figure 4, we plot the recall vs precision curves for the algorithms proposed by Cicconet et al [22], Ecins et al [21], Speciale et al [33], Hruda [38], and Nagar and Raman [52]. We observe that the value of the F-score (0.93) for the proposed approach is the highest among all methods.…”

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 95%

“…This leads an increased computational complexity. Hruda et al proposed an efficient method where they proposed a differential measure for reflection symmetry in 3D point clouds [37], [38]. However, the performance of this method does not perform well for non-uniformly sampled point clouds.…”

Section: Related Workmentioning

confidence: 99%

“…Funk et al published a dataset of 1354 objects represented as 3D point clouds exhibiting single reflection symmetry. We compare the accuracy of detection of plane of symmetry with the accuracies that of the state-of-the-art methods proposed by Cicconet et al [22], Ecins et al [21], Speciale et al [33], Hruda [38], and Nagar and Raman [52]. We use the F-score as a performance measure metric proposed by Funk et al [1].…”

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 99%

“…Hence, the distance threshold criterion is true for all the approaches. (M3), Hruda [38] (M4), Nagar and Raman [52] (M5), and our method on the benchmark dataset [1].…”

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 99%

“…4. Precision (pr) vs Recall (rc) curve for the methods proposed by Cicconet et al [22] (M1), Ecins et al [21] (M2), Speciale et al [33](M3), Hruda[38] (M4), Nagar and Raman[52] (M5), and our method on the benchmark dataset[1].…”

mentioning

confidence: 99%

See 4 more Smart Citations

Robust Estimation of Reflection Symmetry in Noisy and Partial 3D Point Clouds

Nagar¹

2021

Preprint

View full text Add to dashboard Cite

Detecting the reflection symmetry plane of an object represented by a 3D point cloud is a fundamental problem in 3D computer vision and geometry processing due to its various applications such as compression, object detection, robotic grasping, 3D surface reconstruction, etc. There exist several efficient approaches for solving this problem for clean 3D point clouds. However, this problem becomes difficult to solve in the presence of outliers and missing parts due to occlusions while scanning the objects through 3D scanners. The existing methods try to overcome these challenges mostly by voting-based techniques but fail in challenging settings. In this work, we propose a statistical estimator for the plane of reflection symmetry that is robust to outliers and missing parts. We pose the problem of finding the optimal estimator as an optimization problem on a 2-sphere that quickly converges to the global solution. We further propose a 3D point descriptor that is invariant to 3D reflection symmetry using the spectral properties of the geodesic distance matrix constructed from the neighbors of a point. This helps us in decoupling the chicken-and-egg problem of finding optimal symmetry plane and correspondences between the reflective symmetric points. We show that the proposed approach achieves the state-of-the-art performance on the benchmarks dataset.

show abstract

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 95%

Section: Related Workmentioning

confidence: 99%

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 99%

“…Hence, the distance threshold criterion is true for all the approaches. (M3), Hruda [38] (M4), Nagar and Raman [52] (M5), and our method on the benchmark dataset [1].…”

Section: A Evaluation Of Reflection Symmetry Planementioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations