检索结果-内蒙古大学图书馆

Balanced Class-Incremental 3d object classification and Retrieval

IEEE TRANSACTIONS ON KNOWLEdGE ANd dATA ENGINEERING 2024年第1期36卷 35-48页

作者： Liu, An-An Lu, Haochun Zhou, Heyu Li, Tianbao Kankanhalli, Mohan Tianjin Univ Sch Elect & Informat Engn Tianjin 300072 Peoples R China Hefei Comprehens Natl Sci Ctr Inst Artificial Intelligence Hefei 230088 Anhui Peoples R China Natl Univ Singapore Sch Comp Singapore 117543 Singapore

Most existing 3d object classification and retrieval algorithms rely on one-off supervised learning on closed 3d object sets and tend to provide rigid convolutional neural networks with little scalability. Such limitations substantially restrict their potential to learn newly emerged 3d object classes continually in the real world. Aiming to go beyond these limitations, we innovatively propose two new and challenging tasks: class-incremental 3d object classification (CI-3dOC) and class-incremental 3d object retrieval (CI-3dOR), the key to which is class-incremental 3d representation learning. It expects the network to update continually to learn new 3d class representations without forgetting the previously learned ones. To this end, we design a novel balanced distillation network (BdNet) that uses a dual supervision mechanism to balance between consolidating old knowledge (stability) and adapting to new 3d object classes (plasticity) carefully. On the one hand, we employ stability-based supervision to retain the stable and discriminative information of old classes that greatly benefit both classification and retrieval tasks. On the other hand, we use plasticity-based supervision to improve the network's generalization for learning new class 3d representations by transferring knowledge from a temporary teacher network to the current model. By properly handling the relationship between the two modules, we achieve a surprising performance improvement. Furthermore, considering there is no available dataset for evaluation, we build two 3d datasets, INOR-1 and INOR-2, to evaluate these two new tasks. Extensive experimental results demonstrate that our method can significantly outperform other state-of-the-art class-incremental learning methods. Even if we store 500-1000 fewer 3d objects than SOTA methods, BdNet still achieves comparable performance.

关键词： 3d representation learning 3d object classification 3d object retrieval class-incremental learning

来源：评论

学校读者我要写书评

暂无评论

Adaptive Multi-Hypergraph Convolutional Networks for 3d object classification

引用

IEEE TRANSACTIONS ON MULTIMEdIA 2023年 25卷 4842-4855页

作者： Nong, Liping Peng, Jie Zhang, Wenhui Lin, Jiming Qiu, Hongbing Wang, Junyi Xidian Univ Sch Telecommun Engn Xian 710071 Peoples R China Guangxi Normal Univ Coll Phys & Technol Guilin 541004 Peoples R China Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Guilin Univ Elect Technol Sch Informat & Commun Guilin 541004 Peoples R China

3d object classification is an important task in computer vision. In order to explore the high-order and multi-modal correlations among 3d data, we propose an adaptive multi-hypergraph convolutional networks (AMHCN) framework to enhance 3d object classification performance. The proposed network improves the current hypergraph neural networks in two aspects. Firstly, existing networks rely on hyperedge constrained neighborhoods for feature aggregation, which may introduce noise or ignore positive information outside the hyperedges. To this end, we develop the partially absorbing random walks (PARW) to hypergraph for capturing optimal vertex neighborhoods from hypergraph globally. Then, based on the PARW on hypergraph, we design a new hypergraph convolution operator to learn deep embeddings from the optimized high-order correlation, which enables effective information propagation among the most relevant vertices. Secondly, concerning the multi-modal representations in practice, the current multi-modal hypergraph learning models either treat all modalities equally or introduce abundant parameters to learn weights of different modalities. To overcome these shortcomings, we propose a simple but effective dynamic weighting strategy for combining multi-modal representations, in which the importance of each modality can be adjusted adaptively by the loss function. We apply the proposed model to 3d object classification, and the experimental results on two 3d benchmark datasets demonstrate that our method outperforms the state-of-the-art methods, testifying to the effectiveness of both our convolution method and multi-modality fusion strategy.

关键词： Hypergraph learning hypergraph convolution operator partially absorbing random walks multi-modal fusion 3d object classification

来源：评论

学校读者我要写书评

暂无评论

AdaBoost-Based 3d object classification from Surface and depth Map descriptors 24

AdaBoost-Based 3D Object Classification from Surface and Dep...

引用

16th International Conference on Machine Learning and Computing (ICMLC)

作者： dou, Wentao Sun, Jiaqi Niu, dongmei Peng, Jingliang Jinan Univ Sch Informat Sci & Engn Shandong Prov Key Lab Network Based Intelligent C Guangzhou Peoples R China

ISBN: (纸本)9798400709234

In this work, we aim to advance the traditional approach to 3d object classification for its lighter computation and memory costs than the deep learning based approach. Specifically, we propose a novel algorithm that uses multiple handcrafted descriptors to characterize the 3d surface and its multi-view 2d projections, pairs each descriptor with a base classifier and applies AdaBoost to derive the final classifier. In the proposed algorithm, simple yet effective descriptors are selected or designed for lightweight construction, and shape analyses are made on multiple scales of the 2d projections and integrated for the final classification. Experiments show that the proposed algorithm achieves significantly higher accuracy and efficiency than the benchmark traditional methods. Experiments also show that, compared with the state-of-the-art benchmark deep learning based methods, the proposed algorithm wins sharply in efficiency though sacrificing the accuracy by several percent.

关键词： 3d object classification AdaBoost SVM multi-view multi-scale

来源：评论

学校读者我要写书评

暂无评论

Selective Multi-View deep Model for 3d object classification

Selective Multi-View Deep Model for 3D Object Classification

引用

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

作者： Alzahrani, Mona Usman, Muhammad Anwar, Saeed Helmy, Tarek KFUPM Dept Informat & Comp Sci Dhahran Saudi Arabia Jouf Univ Coll Comp & Informat Sci Sakaka Saudi Arabia KFUPM SDAIA KFUPM Joint Res Ctr Artificial Intelligence Dhahran Saudi Arabia KFUPM Ctr Intelligent Secure Syst Dhahran Saudi Arabia

ISBN: (纸本)9798350365474

3d object classification has emerged as a practical technology with applications in various domains, such as medical image analysis, automated driving, intelligent robots, and crowd surveillance. Among the different approaches, multi-view representations for 3d object classification have shown the most promising results, achieving state-of-theart performance. However, there are certain limitations in current view-based 3d object classification methods. One observation is that using all captured views for classification can confuse the classifier and lead to misleading results for certain classes. Additionally, some views may contain more discriminative information for object classification than others. These observations motivate the development of smarter and more efficient selective multi-view classification models. In this work, we propose a Selective MultiView deep Model that extracts multi-view images from 3d data representations and selects the most influential view by assigning importance scores using the cosine similarity method based on visual features detected by a pre-trained CNN. The proposed method is evaluated on the ModelNet40 dataset for the task of 3d classification. The results demonstrate that the proposed model achieves an overall accuracy of 88.13% using only a single view when employing a shading technique for rendering the views, pre-trained ResNet152 as the backbone CNN for feature extraction, and a Fully Connected Network (FCN) as the classifier.

关键词： 3d object classification 3d object Recognition Multi-View Conventional Neural Network Multi-View object classification

来源：评论

学校读者我要写书评

暂无评论

diagnose Label Errors for 3d object classification

Diagnose Label Errors for 3D Object Classification

引用

2023 International Conference on Electronics, Information, and Communication, ICEIC 2023

作者： Wang, Jiayue Wang, Ren Kim, Tae Sung Kim, Jin-Sung Lee, Hyuk-Jae Seoul National University Electrical and Computer Engineering Seoul Korea Republic of Sun Moon University Electronic Engineering Asan Korea Republic of

ISBN: (纸本)9798350320213

3d object classification is widely used in many real-life scenarios and has recently become a popular research area. Meanwhile, 3d datasets tend to be larger and more complex, increasing the difficulty of labeling. Since label errors in test data affect the accuracy of model evaluation and analysis, this study proposes a multi-view post-training label refinery protocol to clean the label errors in test data. Experimental results on ModelNet40 show that 4.58% of test samples are selected as label-error candidates and 29.2% of them are relabeled. On this basis, this study analyzes the weakness of ModelNet40 and provides suggestions for future dataset construction. © 2023 IEEE.

关键词： 3d dataset analysis 3d object classification noisy labels

来源：评论

学校读者我要写书评

暂无评论

Point Cloud-Based 3d object classification With Non Local Attention and Lightweight Convolution Neural Networks

引用

IEEE ACCESS 2024年 12卷 158530-158545页

作者： Karthik, R. Inamdar, Rohan Sundarr, S. Kavin Cho, Jaehyuk Veerappampalayam Easwaramoorthy, Sathishkumar Vellore Inst Technol CCPS Chennai 600127 India Vellore Inst Technol Sch Comp Sci & Engn SCOPE Chennai 600127 India Jeonbuk Natl Univ Dept Software Engn Div Elect & Informat Engn Jeonju Si 54896 Jeonrabug Do South Korea Sunway Univ Sch Engn & Technol Selangor 47500 Malaysia

Three-dimensional (3d) object classification is crucial in various applications, including autonomous driving, robotics, and augmented reality, where precise detection of objects in 3d space is required. Traditional techniques for 3d object classification encompass voxel-based representations, multi-view projections, and point cloud analysis. However, these approaches can be computationally intensive, and might not capture complete structure of the underlying object. This research introduces a novel lightweight model for 3d object classification based on point clouds that achieves higher accuracy while requiring minimal computing overhead. The proposed system incorporates adaptive sampling, an Attentive-Convolution Feature block, a dualpath feature processing block, and a non-local attention mechanism to enhance feature representation and classification performance. To our knowledge, this is the first study to propose a lightweight TnetLight module as a transformation network that aligns and converts input point clouds, capturing subtle geometric variations and highlighting differences between objects to enhance shape recognition and differentiation accuracy. Additionally, the model integrates an Attentive-Convolution Feature extraction block that combines local geometric features with global contextual information, enhancing the network's ability to capture both detailed and broader characteristics for better differentiation between objects with slight variations. The network was evaluated on the ModelNet10 dataset, achieving an accuracy of 94.4%.

关键词： Three-dimensional displays Feature extraction Shape Point cloud compression Solid modeling object recognition Convolutional neural networks Accuracy Adaptation models Convolution 3d object classification lightweight CNN point cloud non-local attention dualpath block adaptive sampling

来源：评论

学校读者我要写书评

暂无评论

L3dOC: Lifelong 3d object classification

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2021年 30卷 7486-7498页

作者： Liu, Yuyang Cong, Yang Sun, Gan Zhang, Tao dong, Jiahua Liu, Hongsen Chinese Acad Sci Shenyang Inst Automat State Key Lab Robot Shenyang 110016 Peoples R China Chinese Acad Sci Inst Robot Shenyang 110169 Peoples R China Chinese Acad Sci Inst Intelligent Mfg Shenyang 110169 Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China JD Com Inc Beijing 100176 Peoples R China

3d object classification has been widely applied in both academic and industrial scenarios. However, most state-of-the-art algorithms rely on a fixed object classification task set, which cannot tackle the scenario when a new 3d object classification task is coming. Meanwhile, the existing lifelong learning models can easily destroy the learned tasks performance, due to the unordered, large-scale, and irregular 3d geometry data. To address these challenges, we propose a Lifelong 3d object classification (i.e., L3dOC) model, which can consecutively learn new 3d object classification tasks via imitating "human learning". More specifically, the core idea of our model is to capture and store the cross-task common knowledge of 3d geometry data in a 3d neural network, named as point-knowledge, through employing layer-wise point-knowledge factorization architecture. Afterwards, a task-relevant knowledge distillation mechanism is employed to connect the current task to previous relevant tasks and effectively prevent catastrophic forgetting. It consists of a point-knowledge distillation module and a transforming-space distillation module, which transfers the accumulated point-knowledge from previous tasks and soft-transfers the compact factorized representations of the transforming-space, respectively. To our best knowledge, the proposed L3dOC algorithm is the first attempt to perform deep learning on 3d object classification tasks in a lifelong learning way. Extensive experiments on several point cloud benchmarks illustrate the superiority of our L3dOC model over the state-of-the-art lifelong learning methods.

关键词： 3d object classification lifelong learning point-knowledge task-relevant knowledge distillation

来源：评论

学校读者我要写书评

暂无评论

Hierarchical multi-view context modelling for 3d object classification and retrieval

引用

INFORMATION SCIENCES 2021年 547卷 984-995页

作者： Liu, An-An Zhou, Heyu Nie, Weizhi Liu, Zhenguang Liu, Wu Xie, Hongtao Mao, Zhendong Li, Xuanya Song, dan Tianjin Univ Sch Elect & Informat Engn Tianjin 300072 Peoples R China Zhejiang Gongshang Univ Hangzhou 310018 Peoples R China AI Res JD Beijing 100105 Peoples R China Univ Sci & Technol China Sch Informat Sci & Technol Hefei 230026 Peoples R China Baidu Inc Beijing 100105 Peoples R China

Recent advances in 3d sensors and 3d modelling software have led to big 3d data. 3d object classification and retrieval are becoming important but challenging tasks. One critical problem for them is how to learn the discriminative multi-view visual characteristics. To address it, we proposes a hierarchical multi-view context modelling method (HMVCM). It consists of four key modules. First, the module of view-level context learning is designed to learn visual context features with respect to individual views and their neighbours. This module can imitate the human need to look back and forth to identify and compare the discriminative parts of individual 3d objects based on a joint convolutional neural network (CNN) and bidirectional long short-term memory (Bi-LSTM) network. Then, a multi-view grouping module is introduced to split views into several groups based on their visual appearance. A raw group-level representation can be obtained by the weighted sum of the view-level descriptors. Furthermore, we employ the Bi-LSTM to exploit the context among adjacent groups to generate group-wise context features. Finally, all group-wise context features are fused into a compact 3d object descriptor according to their significance. Extensive experiments on ModelNet10, ModelNet40 and ShapeNetCore55 demonstrate the superiority of the proposed method. (C) 2020 Elsevier Inc. All rights reserved.

关键词： Multi-view learning 3d object retrieval 3d object classification

来源：评论

学校读者我要写书评

暂无评论

Hypergraph wavelet neural networks for 3d object classification

引用

NEUROCOMPUTING 2021年 463卷 580-595页

作者： Nong, Liping Wang, Junyi Lin, Jiming Qiu, Hongbing Zheng, Lin Zhang, Wenhui Xidian Univ Sch Telecommun Engn Xian 710071 Peoples R China Guangxi Normal Univ Coll Phys & Technol Guilin 541004 Peoples R China Guilin Univ Elect Technol Sch Informat & Commun Guilin 541004 Peoples R China Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China

Recently, hypergraph learning has shown great potential in a variety of classification tasks. However, existing hypergraph neural networks lack flexibility in modeling and extracting high-order relationships among data. To solve this problem, we propose a novel framework called hypergraph wavelet neural networks (HGWNN) to explore the high-order correlation in 3d data. Firstly, considering the non-uniformity of most data sets in the real world, we propose a "data-driven" hypergraph construction scheme, which is more efficient than some commonly used hypergraph construction methods. Secondly, in order to efficiently learn deep embeddings from the constructed hypergraph, we propose a hypergraph wavelet convolution operator. It enables efficient information aggregation by fully exploiting the localization property of wavelets. This convolution operator is suitable for both non-uniform and uniform hyper graphs. Finally, we design a new hypergraph regularizer based on the sparse prior of wavelet coefficients to promote local smoothness and avoid network overfitting. We have conducted experiments on object classification tasks on two 3d benchmark datasets: the National Taiwan University (NTU) 3d model data set and the ModelNet40 dataset. Experimental results demonstrate the effectiveness of the proposed method compared with the state-of-the-art methods. (c) 2021 Elsevier B.V. All rights reserved.

关键词： 3d object classification Hypergraph Hypergraph wavelet transform Hypergraph wavelet convolution Sparse prior

来源：评论

学校读者我要写书评

暂无评论

Pointwise CNN for 3d object classification on Point Cloud

引用

JOURNAL OF INFORMATION PROCESSING SYSTEMS 2021年第4期17卷 787-800页

作者： Song, Wei Liu, Zishu Tian, Yifei Fong, Simon North China Univ Technol Sch Informat Sci & Technol Beijing Peoples R China Univ Macau Dept Comp & Informat Sci Macau Peoples R China

Three-dimensional (3d) object classification tasks using point clouds are widely used in 3d modeling, face recognition, and robotic missions. However, processing raw point clouds directly is problematic for a traditional convolutional network due to the irregular data format of point clouds. This paper proposes a pointwise convolution neural network (CNN) structure that can process point cloud data directly without preprocessing. First, a 2d convolutional layer is introduced to percept coordinate information of each point. Then, multiple 2d convolutional layers and a global max pooling layer are applied to extract global features. Finally, based on the extracted features, fully connected layers predict the class labels of objects. We evaluated the proposed pointwise CNN structure on the ModelNet10 dataset. The proposed structure obtained higher accuracy compared to the existing methods. Experiments using the ModelNet10 dataset also prove that the difference in the point number of point clouds does not significantly influence on the proposed pointwise CNN structure.

关键词： Point Clouds Pointwise CNN 3d object classification

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：