Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
139 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

A Task-driven Network for Mesh Classification and Semantic Part Segmentation (2306.05246v3)

Published 8 Jun 2023 in cs.CV and cs.GR

Abstract: With the rapid development of geometric deep learning techniques, many mesh-based convolutional operators have been proposed to bridge irregular mesh structures and popular backbone networks. In this paper, we show that while convolutions are helpful, a simple architecture based exclusively on multi-layer perceptrons (MLPs) is competent enough to deal with mesh classification and semantic segmentation. Our new network architecture, named Mesh-MLP, takes mesh vertices equipped with the heat kernel signature (HKS) and dihedral angles as the input, replaces the convolution module of a ResNet with Multi-layer Perceptron (MLP), and utilizes layer normalization (LN) to perform the normalization of the layers. The all-MLP architecture operates in an end-to-end fashion and does not include a pooling module. Extensive experimental results on the mesh classification/segmentation tasks validate the effectiveness of the all-MLP architecture.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (66)
  1. Scape: shape completion and animation of people. ACM Trans. Graph. 24 (2005), 408–416.
  2. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. ArXiv abs/2104.13478 (2021).
  3. Layer normalization. ArXiv abs/1607.06450 (2016).
  4. Learning class-specific descriptors for deformable shapes using localized spectral convolutional networks. Computer Graphics Forum 34 (2015).
  5. Faust: Dataset and evaluation for 3d mesh registration. 2014 IEEE Conference on Computer Vision and Pattern Recognition (2014), 3794–3801.
  6. An image is worth 16x16 words: Transformers for image recognition at scale. ArXiv abs/2010.11929 (2021).
  7. Deep geometric functional maps: Robust feature learning for shape correspondence. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2020), pp. 8592–8601.
  8. Laplacian2mesh: Laplacian-based mesh understanding. IEEE Transactions on Visualization and Computer Graphics (2023), 1–13.
  9. Gwcnn: A metric alignment layer for deep shape analysis. Computer Graphics Forum 36 (2017).
  10. Meshnet: Mesh neural network for 3d shape representation. In Proceedings of the AAAI Conference on Artificial Intelligence (2019), vol. 33, pp. 8279–8286.
  11. Shape retrieval contest 2007: Watertight models track. SHREC competition 8, 7 (2007), 7.
  12. Pct: Point cloud transformer. Comput. Vis. Media 7 (2021), 187–199.
  13. 3d-coded: 3d correspondences by deep deformation. In Proceedings of the european conference on computer vision (ECCV) (2018), pp. 230–246.
  14. Meshcnn: a network with an edge. ACM Transactions on Graphics (TOG) 38 (2019), 1 – 12.
  15. Curvanet: Geometric deep learning based on directional curvature for 3d shape analysis. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (2020).
  16. Subdivision-based mesh convolution networks. ACM Trans. Graph. 41, 3 (mar 2022).
  17. Deep residual learning for image recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2016), 770–778.
  18. Ioffe S.: Batch renormalization: Towards reducing minibatch dependence in batch-normalized models. In NIPS (2017).
  19. Klokov R., Lempitsky V. S.: Escape from cells: Deep kd-networks for the recognition of 3d point cloud models. 2017 IEEE International Conference on Computer Vision (ICCV) (2017), 863–872.
  20. A multi-view recurrent neural network for 3d mesh segmentation. Comput. Graph. 66 (2017), 103–112.
  21. Pointcnn: Convolution on x-transformed points. In NeurIPS (2018).
  22. So-net: Self-organizing network for point cloud analysis. In Proceedings of the IEEE conference on computer vision and pattern recognition (2018), pp. 9397–9406.
  23. Shrec ’11 track: Shape retrieval on non-rigid 3d watertight meshes. In 3DOR@Eurographics (2011).
  24. Loop C. T.: Smooth subdivision surfaces based on triangles. In Masters Thesis, Department of Mathematics, University of Utah (January 1987) (1987).
  25. Deep functional maps: Structured prediction for dense shape correspondence. In Proceedings of the IEEE international conference on computer vision (2017), pp. 5659–5667.
  26. Lahav A., Tal A.: Meshwalker: Deep mesh understanding by random walks. ACM Trans. Graph. 39 (2020), 263:1–263:13.
  27. Convolutional neural networks on surfaces via seamless toric covers. ACM Transactions on Graphics (TOG) 36 (2017), 1 – 10.
  28. Field convolutions for surface cnns. 2021 IEEE/CVF International Conference on Computer Vision (ICCV) (2021), 9981–9991.
  29. Primal-dual mesh convolutional neural networks. ArXiv abs/2010.12455 (2020).
  30. Zoomout: spectral upsampling for efficient shape correspondence. ACM Transactions on Graphics (TOG) 38, 6 (2019), 1–14.
  31. Vnect: Real-time 3d human pose estimation with a single rgb camera. ACM Trans. Graph. 36 (2017), 44:1–44:14.
  32. Deep learning-based semantic vessel graph extraction for intracranial aneurysm rupture risk management. International Journal of Computer Assisted Radiology and Surgery 18 (2023), 517 – 525.
  33. Functional maps: a flexible representation of maps between shapes. ACM Transactions on Graphics (ToG) 31, 4 (2012), 1–11.
  34. Poulenard A., Ovsjanikov M.: Multi-directional geodesic neural networks via equivariant convolution. ACM Transactions on Graphics (TOG) 37, 6 (2018), 1–14.
  35. Pointnet: Deep learning on point sets for 3d classification and segmentation. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017), 77–85.
  36. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. In NIPS (2017).
  37. You only look once: Unified, real-time object detection. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2016), 779–788.
  38. U-net: Convolutional networks for biomedical image segmentation. In MICCAI (2015).
  39. Part-based mesh segmentation: a survey. In Computer Graphics Forum, vol. 37, Wiley Online Library, pp. 235–274.
  40. Continuous and orientation-preserving correspondences via functional maps. ACM Transactions on Graphics (TOG) 37 (2018), 1 – 16.
  41. Unsupervised deep learning for structured shape matching. In Proceedings of the IEEE/CVF International Conference on Computer Vision (2019), pp. 1617–1627.
  42. Octnet: Learning deep 3d representations at high resolutions. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017), 6620–6629.
  43. Diffusionnet: Discretization agnostic learning on surfaces. ACM Trans. Graph. 41, 3 (mar 2022).
  44. Deep learning 3d shape surfaces using geometry images. In ECCV (2016).
  45. Deeppano: Deep panoramic representation for 3-d shape recognition. IEEE Signal Processing Letters 22 (2015), 2339–2343.
  46. Multi-view convolutional neural networks for 3d shape recognition. 2015 IEEE International Conference on Computer Vision (ICCV) (2015), 945–953.
  47. Smirnov D., Solomon J. M.: Hodgenet: Learning spectral geometry on triangle meshes. ACM Trans. Graph. 40 (2021), 166:1–166:11.
  48. Instance normalization: The missing ingredient for fast stylization. ArXiv abs/1607.08022 (2016).
  49. Articulated mesh animation from multi-view silhouettes. ACM SIGGRAPH 2008 papers (2008).
  50. Active co-analysis of a set of shapes. ACM Transactions on Graphics (TOG) 31 (2012), 1 – 10.
  51. ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders. arXiv e-prints (Jan. 2023), arXiv:2301.00808.
  52. Cnns on surfaces using rotation-equivariant features. ACM Transactions on Graphics (TOG) 39 (2020), 92:1 – 92:12.
  53. Wu Y., He K.: Group Normalization. arXiv e-prints (Mar. 2018), arXiv:1803.08494.
  54. O-cnn. ACM Transactions on Graphics (TOG) 36 (2017), 1 – 11.
  55. Pointconv: Deep convolutional networks on 3d point clouds. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition (2019), pp. 9621–9630.
  56. 3d shapenets: A deep representation for volumetric shapes. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2015), 1912–1920.
  57. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (TOG) 38 (2019), 1 – 12.
  58. Adaptive O-CNN: A Patch-based Deep Representation of 3D Shapes. ACM Transactions on Graphics (SIGGRAPH Asia) 37, 6 (2018).
  59. Directionally convolutional networks for 3d shape segmentation. In 2017 IEEE International Conference on Computer Vision (ICCV) (2017), pp. 2717–2726.
  60. Paconv: Position adaptive convolution with dynamic kernel assembling on point clouds. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2021), pp. 3173–3182.
  61. Spidercnn: Deep learning on point sets with parameterized convolutional filters. In Proceedings of the European conference on computer vision (ECCV) (2018), pp. 87–102.
  62. A survey on deep geometry learning: From a representation perspective. Computational Visual Media 6 (2020), 113–133.
  63. Geometry sharing network for 3d point cloud classification and segmentation. In Proceedings of the AAAI Conference on Artificial Intelligence (2020), vol. 34, pp. 12500–12507.
  64. Intra: 3d intracranial aneurysm dataset for deep learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2020), pp. 2656–2666.
  65. 3d medical point transformer: Introducing convolution to attention networks for medical point cloud analysis. arXiv preprint arXiv:2112.04863 (2021).
  66. Adaptive graph convolution for point cloud analysis. In Proceedings of the IEEE/CVF international conference on computer vision (2021), pp. 4965–4974.
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now