3D Point Cloud Classification Research Articles

HBIM MODELLING PROCESS FROM 3D POINT CLOUDS BY APPLYING ARTIFICIAL INTELLIGENCE ALGORITHMS IN CULTURAL HERITAGE

In the context of Cultural Heritage (CH), the widespread adoption of 3D point cloud technology, coupled with Artificial Intelligence (AI) algorithms, plays a pivotal role. These technologies facilitate the creation of as-built models by integrating Building Information Modelling (BIM) strategies, enhancing collaboration within the Architecture, Engineering, and Construction (AEC) sector. Leveraging computer vision, robotics, and remote sensing, 3D points clouds provide rich data. However, manual segmentation and classification are labor-intensive and error prone. Consequently, researchers increasingly turn to machine learning (ML) and deep learning (DL) techniques for automating these tasks. The transition from manual reconstruction to automated procedures is crucial. Despite progress, gaps remain, particularly in incorporating 3D point cloud segmentation into Historical Building Information Modelling (HBIM). The lack of conclusive evidence regarding automated derivation of parametric attributes from segmentation outcomes underscores the need for further exploration. Addressing this gap is essential for cultural asset documentation, conservation, and upkeep. By automating the segmentation and classification of 3D point clouds, efficient communication via a shared database becomes feasible. The article aims to review studies on semantically parsing and classifying 3D point clouds using AI algorithms, particularly within complex cultural heritage geometries, shedding light on potential benefits and barriers.

Hyperbolic prototype rectification for few-shot 3D point cloud classification

Few-shot point cloud classification is a challenging task in 3D computer vision and has received widespread attention from researchers. Most of the works on deep learning models rely heavily on Euclidean spatial metrics. However, point cloud objects often have complex non-Euclidean geometric structures, with underlying inter/intra-class hierarchical structures, which are difficult to capture by current Euclidean-based deep learning models. Moreover, due to the lack of training samples, many few-shot learning methods often suffer from the overfitting problem. Given the Hyperbolic metric of non-Euclidean geometry offering hierarchical structural prior, as we assume to be able to assist FSL task, we propose Hyperbolic Prototype Rectification (HPR) for few-shot point cloud classification, without requiring extra learnable parameter. Firstly, point clouds are embedded into hyperbolic space to better describe hierarchical similarity relationships in data. Secondly, the HPR utilizes hyperbolic spatial and distributional information to enhance the feature representation and improve the generalization capability, with more appropriate hyperbolic prototypes. The few-shot classification experiments and further ablation studies conducted on widely used point cloud datasets demonstrate the effectiveness of our method. On the real-world ScanObjectNN(-PB) datasets, the average classification accuracy outperforms the SOTA method by 2.08%(0.66%), respectively, indicating that the proposed HPR has great generalization capability and strong robustness against perturbed data. Our code is available at: https://github.com/Jonathan-UCAS/HPR.

A novel method for the classification of 3D point clouds based on the improved PointNet++

Abstract In deep learning, point clouds are used as the primary input format for 3D data, which can provide detailed geometric information about objects in the original 3D space. PointNet++ is a deep learning network that uses point cloud data as an input format, which avoids the losses associated with the previous conversion of point cloud into 3D voxelization and a collection of 2D images. Although PointNet++ can directly process point cloud data in various ways, due to the disordered, irregular, and unevenly distributed nature of point cloud data, the effect of extracting point cloud features could be better. The large amount of point cloud data also leads to the training model falling into the local optimal solution, which affects the training results. In recent years, some effective methods and strategies have emerged to address these problems. In this thesis, three methods are proposed based on the PointNet++ network: feature similarity-based attention pooling, small kernel convolution, and diverse branch block method to improve the performance of the PointNet++ network. Experiments show that the improvement methods proposed in this paper effectively improve the feature extraction accuracy, which improves the accuracy of the PointNet++ network for classification on the ModelNet40_Normal_Resampled dataset, with an overall improvement of 1% compared with PointNet++.

3DFFL: privacy-preserving Federated Few-Shot Learning for 3D point clouds in autonomous vehicles

This paper presents a comprehensive study of 3D point cloud Federated Few-Shot Learning (3DFFL), focusing on addressing challenges such as limited data availability and privacy concerns in point cloud classification for applications such as autonomous vehicles. We introduce a novel approach that integrates Federated Learning with Few-Shot Learning techniques, with a special emphasis on optimizing network architectures for 3D point cloud data. Our method capitalizes on the strengths of PointNet++ for feature extraction and ProtoNet for classification, all within a federated learning framework to ensure data privacy and collaborative learning. Significantly, the approach is augmented with the use of attention and SoftMax layers, enhancing the feature extraction and classification processes. Extensive experiments on the ModelNet40, ShapeNet, and ScanOnjectNN datasets validate our method’s accuracy and adaptability in handling 3D point cloud classification, especially in privacy-sensitive and collaborative scenarios. This study not only demonstrates the potential of integrating attention mechanisms and SoftMax layers in 3DFFL but also lays a robust foundation for future advancements in this evolving field, particularly in technologies dependent on 3D data processing.

Granformer: A granular transformer net with linear complexity

Recently, transformer models have demonstrated excellent performance across various intelligent applications owing to their ability to understand global context through self-attention mechanism. However, the extensively investigated multiplicative-based attention mechanism is inadequate for capturing relationship-based feature representations, as the dot product cannot depict the intricate semantic information between objects. Moreover, it has a great computational burden with a complexity of O(n2), due to the global feature representation capability achieved by calculating the relationship between each token within the entire feature sequence. To solve the current problem, this paper proposes a granular transformer framework with linear complexity, wherein diverse granulation functions can be employed to supersede the prevailing multiplicative relationships, and an innovative linearization methodology in the form of matrix factorization is designed to reduce the computational burden. Relying on the intricate semantics information embedded within granular structures, the capacity for feature extraction is significantly more comprehensive. Then, a novel matrix factorization methodology is developed for the linearity of granulation-based attention, accomplished by implementing separate deformable convolution sampling and using an approximate iterative algorithm based on cubic equations to calculate the Moore–Penrose inverse. The mathematical proof that our method is approximate with the complete granulation-based attention matrix is investigated in detail. Finally, the performance of Granformer, an innovative reconfiguration of plug-and-play transformer block, is evaluated on representative intelligent applications, including 3D point cloud classification, emotion recognition and sentiment analysis, and object detection. The experimental results suggest that our methodologies outperform the state-of-the-art models.

Fuzzy preference matroids rough sets for approximate guided representation in transformer

Recently, the transformer has exhibited remarkable performance across various applications, primarily owing to its exceptional capability in capturing global information through the attention mechanism. Nevertheless, the dot-product within attention results in inaccurate and confusing guidance in the presence of uncertain features, subsequently affecting robustness. Hence, introducing an approximate guided method to handle uncertain features is essential for achieving enhanced objectivity. In this study, we employed fuzzy preference relations to construct a matroids structure, introducing it into rough sets to form the fuzzy preference matroids rough sets. Detailed mathematical properties and axiomatic proofs are presented. The closed sets within the matroids form a subspace of the lower approximation proving that the proposed satisfy the approximation ability of rough sets. Further, an approximate guided representation method with a lower approximation of the matroids has been developed. It is integrated into a plug-and-play transformer block that can be flexibly deployed across various tasks. Experimental validation has demonstrated the outstanding performance of this method across 3D point cloud classification, named entity recognition, multimodal emotion recognition, and image classification. Specifically, on the Weibo NER and the IEMOCAP datasets, attain state-of-the-art (SOTA) F1-scores. The resource code is validated at https://github.com/WinnieSunning/FPMRST.

ModelNet-O: A large-scale synthetic dataset for occlusion-aware point cloud classification

Recently, 3D point cloud classification has made significant progress with the help of many datasets. However, these datasets do not reflect the incomplete nature of real-world point clouds caused by occlusion, which limits the practical application of current methods. To bridge this gap, we propose ModelNet-O, a large-scale synthetic dataset of 123,041 samples that emulates real-world point clouds with self-occlusion caused by scanning from monocular cameras. ModelNet-O is 10 times larger than existing datasets and offers more challenging cases to evaluate the robustness of existing methods. Our observation on ModelNet-O reveals that well-designed sparse structures can preserve structural information of point clouds under occlusion, motivating us to propose a robust point cloud processing method that leverages a critical point sampling (CPS) strategy in a multi-level manner. We term our method PointMLS. Through extensive experiments, we demonstrate that our PointMLS achieves state-of-the-art results on ModelNet-O and competitive results on regular datasets such as ModelNet40 and ScanObjectNN, and we also demonstrate its robustness and effectiveness. Code available: https://github.com/fanglaosi/ModelNet-O_PointMLS.

ESTATE: A Large Dataset of Under-Represented Urban Objects for 3D Point Cloud Classification

Abstract. Cityscapes contain a variety of objects, each with a particular role in urban administration and development. With the rapid growth and implementation of 3D imaging technology, urban areas are increasingly surveyed with high-resolution point clouds. This technical advancement extensively improves our ability to capture and analyse urban environments and their small objects. Deep learning algorithms for point cloud data have shown considerable capacity in 3D object classification but still face problems with generally under-represented objects (such as light poles or chimneys). This paper introduces the ESTATE dataset (https://github.com/3DOM-FBK/ESTATE), which combines available datasets of various sensors, densities, regions, and object types. It includes 13 classes featuring intensity and/or colour attributes. Tests using ESTATE demonstrate that the dataset improves the classification performance of deep learning techniques and could be a game-changer to advance in the 3D classification of urban objects.

Neural Radiance Selector: Find the best 2D representations of 3D data for CLIP based 3D tasks

Representing the world in 3D space provides vivid texture and depth information. However, 3D datasets currently do not match the scale of 2D datasets. There is a growing trend in representing 3D data as multi-view 2D images and using large-scale 2D models, to solve 3D tasks. In this work, we present the Neural Radiance Selector, a method that automatically selects the optimal 2D representations of 3D data. Instead of indiscriminately sampling multi-view 2D images, we define the optimal 2D views as those capable of reconstructing the entire 3D scene with a conditional neural radiance field. We propose two distinct methods for 3D point cloud data and 3D implicit models to achieve faster inference. We demonstrate the efficacy of our methods in various 3D tasks, including zero-shot 3D point cloud classification, 3D implicit model classification, and language-guided NeRF editing.

A comprehensive overview of deep learning techniques for 3D point cloud classification and semantic segmentation

A comprehensive overview of deep learning techniques for 3D point cloud classification and semantic segmentation

Context-based local-global fusion network for 3D point cloud classification and segmentation

3D point clouds have gained much research attention because of their ability to represent the spatial information of real-world environments in a detailed manner. Despite recent progress in point cloud processing with deep neural networks, most of them either implement sophisticated local feature aggregation methods or imitate 2D convolution operations in the range of K nearest neighbors with limited local context information. These methods may struggle to distinguish between similar geometric shapes within the local region of K nearest neighbors, such as doors and walls. To address this issue, we propose a novel local–global fusion network that captures the diverse local geometric shapes with global structure information. The proposed local–global fusion network comprises two main modules. Firstly, we have developed an effective approach for local context learning using incremental dilated KNN (IDKNN) as the neighbor selecting mechanism to enlarge the receptive field and incorporate more reliable points for local geometric shape learning. Secondly, a three-direction region-wise spatial attention (TRSA) algorithm has been developed to explore the global contextual dependencies. For global context learning, we first split the entire 3D space into regions with equal numbers of points, and, then, intra-region context features are extracted to learn the inter-region relations from three orthogonal directions, taking global structural knowledge into account. By fusing the local context information and global contextual dependencies, we establish a Local-Global Fusion Network, end-to-end framework, called LGFNet. Extensive experimental results on several benchmark datasets clearly demonstrate our approach can achieve state-of-the-art (SOTA) performance on point cloud classification, part segmentation, and indoor semantic segmentation. In addition, TRSA and IKDNN can be easily used in a plug-and-play fashion with various existing SOTA networks to substantially improve their performance. Our code is available at https://github.com/jasonwjw/IDKNN

Point-Sim: A Lightweight Network for 3D Point Cloud Classification

Analyzing point clouds with neural networks is a current research hotspot. In order to analyze the 3D geometric features of point clouds, most neural networks improve the network performance by adding local geometric operators and trainable parameters. However, deep learning usually requires a large amount of computational resources for training and inference, which poses challenges to hardware devices and energy consumption. Therefore, some researches have started to try to use a nonparametric approach to extract features. Point-NN combines nonparametric modules to build a nonparametric network for 3D point cloud analysis, and the nonparametric components include operations such as trigonometric embedding, farthest point sampling (FPS), k-nearest neighbor (k-NN), and pooling. However, Point-NN has some blindness in feature embedding using the trigonometric function during feature extraction. To eliminate this blindness as much as possible, we utilize a nonparametric energy function-based attention mechanism (ResSimAM). The embedded features are enhanced by calculating the energy of the features by the energy function, and then the ResSimAM is used to enhance the weights of the embedded features by the energy to enhance the features without adding any parameters to the original network; Point-NN needs to compute the similarity between each feature at the naive feature similarity matching stage; however, the magnitude difference of the features in vector space during the feature extraction stage may affect the final matching result. We use the Squash operation to squeeze the features. This nonlinear operation can make the features squeeze to a certain range without changing the original direction in the vector space, thus eliminating the effect of feature magnitude, and we can ultimately better complete the naive feature matching in the vector space. We inserted these modules into the network and build a nonparametric network, Point-Sim, which performs well in 3D classification tasks. Based on this, we extend the lightweight neural network Point-SimP by adding some trainable parameters for the point cloud classification task, which requires only 0.8 M parameters for high performance analysis. Experimental results demonstrate the effectiveness of our proposed algorithm in the point cloud shape classification task. The corresponding results on ModelNet40 and ScanObjectNN are 83.9% and 66.3% for 0 M parameters—without any training—and 93.3% and 86.6% for 0.8 M parameters. The Point-SimP reaches a test speed of 962 samples per second on the ModelNet40 dataset. The experimental results show that our proposed method effectively improves the performance on point cloud classification networks.

Deep Ordinal Classification in Forest Areas Using Light Detection and Ranging Point Clouds.

Recent advances in Deep Learning and aerial Light Detection And Ranging (LiDAR) have offered the possibility of refining the classification and segmentation of 3D point clouds to contribute to the monitoring of complex environments. In this context, the present study focuses on developing an ordinal classification model in forest areas where LiDAR point clouds can be classified into four distinct ordinal classes: ground, low vegetation, medium vegetation, and high vegetation. To do so, an effective soft labeling technique based on a novel proposed generalized exponential function (CE-GE) is applied to the PointNet network architecture. Statistical analyses based on Kolmogorov-Smirnov and Student's t-test reveal that the CE-GE method achieves the best results for all the evaluation metrics compared to other methodologies. Regarding the confusion matrices of the best alternative conceived and the standard categorical cross-entropy method, the smoothed ordinal classification obtains a more consistent classification compared to the nominal approach. Thus, the proposed methodology significantly improves the point-by-point classification of PointNet, reducing the errors in distinguishing between the middle classes (low vegetation and medium vegetation).

Point-to-Spike Residual Learning for Energy-Efficient 3D Point Cloud Classification

Spiking neural networks (SNNs) have revolutionized neural learning and are making remarkable strides in image analysis and robot control tasks with ultra-low power consumption advantages. Inspired by this success, we investigate the application of spiking neural networks to 3D point cloud processing. We present a point-to-spike residual learning network for point cloud classification, which operates on points with binary spikes rather than floating-point numbers. Specifically, we first design a spatial-aware kernel point spiking neuron to relate spiking generation to point position in 3D space. On this basis, we then design a 3D spiking residual block for effective feature learning based on spike sequences. By stacking the 3D spiking residual blocks, we build the point-to-spike residual classification network, which achieves low computation cost and low accuracy loss on two benchmark datasets, ModelNet40 and ScanObjectNN. Moreover, the classifier strikes a good balance between classification accuracy and biological characteristics, allowing us to explore the deployment of 3D processing to neuromorphic chips for developing energy-efficient 3D robotic perception systems.

PointCVaR: Risk-Optimized Outlier Removal for Robust 3D Point Cloud Classification

With the growth of 3D sensing technology, the deep learning system for 3D point clouds has become increasingly important, especially in applications such as autonomous vehicles where safety is a primary concern. However, there are growing concerns about the reliability of these systems when they encounter noisy point clouds, either occurring naturally or introduced with malicious intent. This paper highlights the challenges of point cloud classification posed by various forms of noise, from simple background noise to malicious adversarial/backdoor attacks that can intentionally skew model predictions. While there's an urgent need for optimized point cloud denoising, current point outlier removal approaches, an essential step for denoising, rely heavily on handcrafted strategies and are not adapted for higher-level tasks, such as classification. To address this issue, we introduce an innovative point outlier cleansing method that harnesses the power of downstream classification models. Using gradient-based attribution analysis, we define a novel concept: point risk. Drawing inspiration from tail risk minimization in finance, we recast the outlier removal process as an optimization problem, named PointCVaR. Extensive experiments show that our proposed technique not only robustly filters diverse point cloud outliers but also consistently and significantly enhances existing robust methods for point cloud classification. A notable feature of our approach is its effectiveness in defending against the latest threat of backdoor attacks in point clouds.

3D meta-classification: A meta-learning approach for selecting 3D point-cloud classification algorithm

Algorithm selection technology, aimed at choosing the most suitable algorithm for a given machine learning task, has achieved significant success in the domain of 2D vision. However, few studies have explored its application to the 3D point cloud domain. The rapid proliferation and development of 3D point cloud classification algorithms underscore the urgent need for exploration in selecting these algorithms. In this paper, we propose a novel meta-learning-based 3D classification approach, termed 3D meta-classification, to address this gap. The approach operates at both the base-level and meta-level phases. At the base level, candidate 3D classification algorithms constantly classify various 3D datasets, recording their classification performance as empirical knowledge. As for the meta-level phase, it specifically tailors the 3D meta-knowledge generator, 3D meta-feature extractor, and 3D meta-database constructor to establish the 3D meta-database, capturing the relationship between the 3D meta-features and empirical knowledge. Leveraging the 3D meta-database, the 3D classification meta-learner trains the meta-model and predicts suitable algorithms for new incoming 3D datasets. Extensive experiments are conducted to select the best-performing classification algorithm for specific 3D datasets from ModelNet40. The results demonstrate the effectiveness of the proposed 3D meta-classification model; the accuracies of one-from-two and one-from-three algorithm selection tasks reach over 98% and 80%, respectively.

A 3D Point Cloud Classification Method Based on Adaptive Graph Convolution and Global Attention.

In recent years, there has been significant growth in the ubiquity and popularity of three-dimensional (3D) point clouds, with an increasing focus on the classification of 3D point clouds. To extract richer features from point clouds, many researchers have turned their attention to various point set regions and channels within irregular point clouds. However, this approach has limited capability in attending to crucial regions of interest in 3D point clouds and may overlook valuable information from neighboring features during feature aggregation. Therefore, this paper proposes a novel 3D point cloud classification method based on global attention and adaptive graph convolution (Att-AdaptNet). The method consists of two main branches: the first branch computes attention masks for each point, while the second branch employs adaptive graph convolution to extract global features from the point set. It dynamically learns features based on point interactions, generating adaptive kernels to effectively and precisely capture diverse relationships among points from different semantic parts. Experimental results demonstrate that the proposed model achieves 93.8% in overall accuracy and 90.8% in average accuracy on the ModeNet40 dataset.

3D point cloud classification and segmentation based on dual attention and weighted dynamic graph convolution

3D point cloud classification and segmentation based on dual attention and weighted dynamic graph convolution

Point Cloud Classification Using Content-Based Transformer via Clustering in Feature Space

Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention, but ignore their content and fail to establish relationships between distant but relevant points. To overcome the limitation of local spatial attention, we propose a point content-based Transformer architecture, called PointConT for short. It exploits the locality of points in the feature space (content-based), which clusters the sampled points with similar features into the same class and computes the self-attention within each class, thus enabling an effective trade-off between capturing long-range dependencies and computational complexity. We further introduce an Inception feature aggregator for point cloud classification, which uses parallel structures to aggregate high-frequency and low-frequency information in each branch separately. Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification. Especially, our method exhibits 90.3% Top-1 accuracy on the hardest setting of ScanObjectNN. Source code of this paper is available at https://github.com/yahuiliu99/PointConT.

Adaptive Multiview Graph Convolutional Network for 3D Point Cloud Classification and Segmentation

Adaptive Multiview Graph Convolutional Network for 3D Point Cloud Classification and Segmentation