SiamesePointNet: A Siamese Point Network Architecture for Learning 3D Shape Descriptor

Jian-hui, Zhou; Wang, M. J.; Mao, Wendong; Gong, Minglun; Liu, X. P.

doi:10.1111/cgf.13804

Cited by 13 publications

(4 citation statements)

References 36 publications

(60 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, they require a large Graphic Process Unit (GPU) memory and are sensitive to rotation variations. Some representative algorithms in this category are PPFNet [105], SiamesePointNet [106] and deep closest point (DCP) [107]. These methods offer robust and accurate registration using a simple RANdom SAmple Consensus (RANSAC) iterative algorithm.…”

Section: D Point Cloudsmentioning

confidence: 99%

Deep learning based computer vision under the prism of 3D point clouds: a systematic review

Tychola,

Vrochidou,

Papakostas

2024

Vis Comput

View full text Add to dashboard Cite

Point clouds consist of 3D data points and are among the most considerable data formats for 3D representations. Their popularity is due to their broad application areas, such as robotics and autonomous driving, and their employment in basic 3D vision tasks such as segmentation, classification, and detection. However, processing point clouds is challenging compared to other visual forms such as images, mainly due to their unstructured nature. Deep learning (DL) has been established as a powerful tool for data processing, reporting remarkable performance enhancements compared to traditional methods for all basic 2D vision tasks. However new challenges are emerging when it comes to processing unstructured 3D point clouds. This work aims to guide future research by providing a systematic review of DL on 3D point clouds, holistically covering all 3D vision tasks. 3D technologies of point cloud formation are reviewed and compared to each other. The application of DL methods for point cloud processing is discussed, and state-of-the-art models’ performances are compared focusing on challenges and solutions. Moreover, in this work the most popular 3D point cloud benchmark datasets are summarized based on their task-oriented applications, aiming to highlight existing constraints and to comparatively evaluate them. Future research directions and upcoming trends are also highlighted.

show abstract

Section: D Point Cloudsmentioning

confidence: 99%

Deep learning based computer vision under the prism of 3D point clouds: a systematic review

Tychola,

Vrochidou,

Papakostas

2024

Vis Comput

View full text Add to dashboard Cite

show abstract

“…The whole network is optimized by using the difference between the input and output using Chamfer loss. Similarly, SiamesePointNet [118] produces the descriptor of interest points by a hierarchical encoder-decoder architecture.…”

Section: B Learning On Point Cloudmentioning

confidence: 99%

A comprehensive survey on point cloud registration

Huang

Mei

Zhang

et al. 2021

Preprint

View full text Add to dashboard Cite

Registration is a transformation estimation problem between two point clouds, which has a unique and critical role in numerous computer vision applications. The developments of optimization-based methods and deep learning methods have improved registration robustness and efficiency. Recently, the combinations of optimization-based and deep learning methods have further improved performance. However, the connections between optimization-based and deep learning methods are still unclear. Moreover, with the recent development of 3D sensors and 3D reconstruction techniques, a new research direction emerges to align cross-source point clouds. This survey conducts a comprehensive survey, including both same-source and cross-source registration methods, and summarize the connections between optimization-based and deep learning methods, to provide further research insight. This survey also builds a new benchmark to evaluate the state-of-the-art registration algorithms in solving cross-source challenges. Besides, this survey summarizes the benchmark data sets and discusses point cloud registration applications across various domains. Finally, this survey proposes potential research directions in this rapidly growing field.

show abstract

“…Correspondences can also be computed through finding a canonical embedding of the input. This idea has been developed for 2D images [8,54] as well as 3D data [62]. The latter works, as many others in this domain, take advantage of point-based architectures such as PointNet [44] and its extensions [45,2,55] that provide a powerful way to learn signatures for point clouds, and that have been mainly exploited for shape classification but not yet for smooth and consistent dense non-rigid shape correspondence.…”

Section: Functional Mapsmentioning

confidence: 99%

Correspondence Learning via Linearly-invariant Embedding

Marin¹,

Rakotosaona²,

Melzi³

et al. 2020

Preprint

View full text Add to dashboard Cite

In this paper, we propose a fully differentiable pipeline for estimating accurate dense correspondences between 3D point clouds. The proposed pipeline is an extension and a generalization of the functional maps framework. However, instead of using the Laplace-Beltrami eigenfunctions as done in virtually all previous works in this domain, we demonstrate that learning the basis from data can both improve robustness and lead to better accuracy in challenging settings. We interpret the basis as a learned embedding into a higher dimensional space. Following the functional map paradigm the optimal transformation in this embedding space must be linear and we propose a separate architecture aimed at estimating the transformation by learning optimal descriptor functions. This leads to the first end-to-end trainable functional map-based correspondence approach in which both the basis and the descriptors are learned from data. Interestingly, we also observe that learning a canonical embedding leads to worse results, suggesting that leaving an extra linear degree of freedom to the embedding network gives it more robustness, thereby also shedding light onto the success of previous methods. Finally, we demonstrate that our approach achieves state-of-the-art results in challenging non-rigid 3D point cloud correspondence applications. * denotes equal contribution.34th Conference on Neural Information Processing Systems (NeurIPS 2020),

show abstract

SiamesePointNet: A Siamese Point Network Architecture for Learning 3D Shape Descriptor

Cited by 13 publications

References 36 publications

Deep learning based computer vision under the prism of 3D point clouds: a systematic review

Deep learning based computer vision under the prism of 3D point clouds: a systematic review

A comprehensive survey on point cloud registration

Correspondence Learning via Linearly-invariant Embedding

Contact Info

Product

Resources

About