Filters








53 Hits in 4.6 sec

PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection [article]

Shaoshuai Shi, Li Jiang, Jiajun Deng, Zhe Wang, Chaoxu Guo, Jianping Shi, Xiaogang Wang, Hongsheng Li
2022 arXiv   pre-print
In this paper, we propose the Point-Voxel Region-based Convolution Neural Networks (PV-RCNNs) for 3D object detection from point clouds.  ...  These two steps deeply integrate the 3D voxel CNN with the PointNet-based set abstraction for extracting discriminative features.  ...  PV-RCNN-v2 with local vector representation for local feature aggregation.  ... 
arXiv:2102.00463v2 fatcat:s36oc3krvne7pgmqy4pn2tlc64

PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection [article]

Shaoshuai Shi, Chaoxu Guo, Li Jiang, Zhe Wang, Jianping Shi, Xiaogang Wang, Hongsheng Li
2021 arXiv   pre-print
We present a novel and high-performance 3D object detection framework, named PointVoxel-RCNN (PV-RCNN), for accurate 3D object detection from point clouds.  ...  Extensive experiments on both the KITTI dataset and the Waymo Open dataset show that our proposed PV-RCNN surpasses state-of-the-art 3D detection methods with remarkable margins by using only point clouds  ...  PV-RCNN for Point Cloud Object Detection In this paper, we propose the PointVoxel-RCNN (PV-RCNN), which is a two-stage 3D detection framework aiming at more accurate 3D object detection from point clouds  ... 
arXiv:1912.13192v2 fatcat:l4j5mera5vdg3oivhz55on7sfa

PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection

Shaoshuai Shi, Chaoxu Guo, Li Jiang, Zhe Wang, Jianping Shi, Xiaogang Wang, Hongsheng Li
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We present a novel and high-performance 3D object detection framework, named PointVoxel-RCNN (PV-RCNN), for accurate 3D object detection from point clouds.  ...  Our proposed method deeply integrates both 3D voxel Convolutional Neural Network (CNN) and PointNet-based set abstraction to learn more discriminative point cloud features.  ...  PV-RCNN for Point Cloud Object Detection In this paper, we propose the PV-RCNN, a two-stage 3D detection framework aiming at more accurate 3D object detection from point clouds.  ... 
doi:10.1109/cvpr42600.2020.01054 dblp:conf/cvpr/ShiGJ0SWL20 fatcat:2ocag5ptzvdbrist7rm2if37fu

M3DeTR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers [article]

Tianrui Guan, Jun Wang, Shiyi Lan, Rohan Chandra, Zuxuan Wu, Larry Davis, Dinesh Manocha
2021 arXiv   pre-print
We present a novel architecture for 3D object detection, M3DeTR, which combines different point cloud representations (raw, voxels, bird-eye view) with different feature scales based on multi-scale feature  ...  In particular, our approach ranks 1st on the well-known KITTI 3D Detection Benchmark for both car and cyclist classes, and ranks 1st on Waymo Open Dataset with single frame point cloud input.  ...  Adopted from PointNet++ [36] and PV-RCNN [39] , Set Abstraction and Voxel Set Abstraction (VSA) module take raw point coordinates P and the 3D voxel-based features f voxel , respectively, to generate  ... 
arXiv:2104.11896v3 fatcat:dytvkxn6bjaozljmvsfvqz4tuy

3D Object Detection Combining Semantic and Geometric Features from Point Clouds [article]

Hao Peng, Guofeng Tong, Zheng Li, Yaqi Wang, Yuyuan Shao
2021 arXiv   pre-print
However, most current methods use a voxel-based detection head with anchors for final classification and localization.  ...  The combination of the two is an effective solution for 3D object detection from point clouds.  ...  PV-RCNN [12] deeply integrates both 3D voxel Convolutional Neural Network (CNN) and PointNet-based set abstraction to learn more discriminative point cloud features.  ... 
arXiv:2110.04704v1 fatcat:zd5vai3dpbdqnku4ef5vcdlwy4

AFE-RCNN: Adaptive Feature Enhancement RCNN for 3D Object Detection

Feng Shuang, Hanzhang Huang, Yong Li, Rui Qu, Pei Li
2022 Remote Sensing  
To perform precise and effective 3D object detection, it is necessary to improve the feature representation ability to extract more feature information of the object points.  ...  Therefore, we propose an adaptive feature enhanced 3D object detection network based on point clouds (AFE-RCNN). AFE-RCNN is a point-voxel integrated network.  ...  Acknowledgments: The authors thanks Jianchuan Qin, Xiuning Liu and Qipeng Pu for helping with the draft writing and checking. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/rs14051176 fatcat:reynlyararcadglvnhlu4ku53u

JPV-Net: Joint Point-Voxel Representations for Accurate 3D Object Detection

Nan Song, Tianyuan Jiang, Jian Yao
2022 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Voxel and point representations are widely applied in recent 3D object detection tasks from LiDAR point clouds.  ...  Voxel representations contribute to efficiently and rapidly locating objects, whereas point representations are capable of describing intra-object spatial relationship for detection refinement.  ...  Acknowledgements This work was partially supported by the Shenzhen Central Guiding the Local Science and Technology Development Program (No. 2021Szvup100).  ... 
doi:10.1609/aaai.v36i2.20125 fatcat:7r7fkhuvlrdzxcmzpajlzjlwbi

SA-Det3D: Self-Attention Based Context-Aware 3D Object Detection [article]

Prarthana Bhattacharyya, Chengjie Huang, Krzysztof Czarnecki
2021 arXiv   pre-print
In this paper, we propose two variants of self-attention for contextual modeling in 3D object detection by augmenting convolutional features with self-attention features.  ...  Existing point-cloud based 3D object detectors use convolution-like operators to process information in a local neighbourhood with fixed-weight kernels and aggregate global context hierarchically.  ...  For objects with very few points, FSA can increase the 3D AP for PointPillars by 2.8% and PV-RCNN by 1.5%.  ... 
arXiv:2101.02672v5 fatcat:re3fvyv4zbco3gsux3llksf7nu

3D-FCT: Simultaneous 3D Object Detection and Tracking Using Feature Correlation [article]

Naman Sharma, Hocksoon Lim
2021 arXiv   pre-print
3D object detection using LiDAR data remains a key task for applications like autonomous driving and robotics.  ...  Calculating correlation across keypoints only allows for real-time object detection. We further extend the multi-task objective to include a tracking regression loss.  ...  A voxel set abstraction module subsequently extracts keypoint specific features from each level of the 3D voxel CNN.  ... 
arXiv:2110.02531v1 fatcat:tnh6ef2i6za2hmunvwptlb3g5i

From Voxel to Point: IoU-guided 3D Object Detection for Point Cloud with Voxel-to-Point Decoder [article]

Jiale Li and Hang Dai and Ling Shao and Yong Ding
2021 arXiv   pre-print
In this paper, we present an Intersection-over-Union (IoU) guided two-stage 3D object detector with a voxel-to-point decoder.  ...  We use a 3D Region of Interest (RoI) alignment to crop and align the features with the proposal boxes for accurately perceiving the object position.  ...  PV-RCNN [33] proposes a Voxel Set Abstraction operation to aggregate the voxel-wise features in the backbone to some sampled keypoints.  ... 
arXiv:2108.03648v1 fatcat:tk7goldrsfh7lideinejmxu4oq

A Survey of Robust 3D Object Detection Methods in Point Clouds [article]

Walter Zimmer, Emec Ercelik, Xingcheng Zhou, Xavier Jair Diaz Ortiz, Alois Knoll
2022 arXiv   pre-print
Finally, we mention the current challenges in 3D object detection in LiDAR point clouds and list some open issues.  ...  The purpose of this work is to review the state-of-the-art LiDAR-based 3D object detection methods, datasets, and challenges.  ...  The authors would like to express their gratitude to the funding agency and to the numerous students at TUM for technical editing, language editing, and proofreading.  ... 
arXiv:2204.00106v1 fatcat:a36gqfxiljbgfjiszufusgkcau

ProposalContrast: Unsupervised Pre-training for LiDAR-based 3D Object Detection [article]

Junbo Yin, Dingfu Zhou, Liangjun Zhang, Jin Fang, Cheng-Zhong Xu, Jianbing Shen, Wenguan Wang
2022 arXiv   pre-print
Considering region-level representations are more suitable for 3D object detection, we devise a new unsupervised point cloud pre-training framework, called ProposalContrast, that learns robust 3D representations  ...  Specifically, with an exhaustive set of region proposals sampled from each point cloud, geometric point relations within each proposal are modeled for creating expressive proposal representations.  ...  We pre-train the backbones of PointRCNN [31] and PV-RCNN [30] on Waymo and transfer to KITTI 3D object detection with different label configurations.  ... 
arXiv:2207.12654v1 fatcat:5cnghswwe5bchfmj5epjqfkmta

Improving 3D Object Detection with Channel-wise Transformer [article]

Hualian Sheng and Sijia Cai and Yuan Liu and Bing Deng and Jianqiang Huang and Xian-Sheng Hua and Min-Jian Zhao
2021 arXiv   pre-print
Previous works on refining 3D proposals have relied on human-designed components such as keypoints sampling, set abstraction and multi-scale feature fusion to produce powerful 3D object representations  ...  Though 3D object detection from point clouds has achieved rapid progress in recent years, the lack of flexible and high-performance proposal refinement remains a great hurdle for existing state-of-the-art  ...  Related Work Point Cloud Representations for 3d Object Detection. Recently, there has been a lot of progress on learning effective representations for the raw LiDAR point clouds.  ... 
arXiv:2108.10723v2 fatcat:4gzdvu5i3bdtxhwc2vu6nkijm4

BADet: Boundary-Aware 3D Object Detection from Point Clouds [article]

Rui Qian, Xin Lai, Xirong Li
2022 arXiv   pre-print
In this paper, we propose BADet for 3D object detection from point clouds.  ...  Besides, we devise a lightweight Region Feature Aggregation Module to fully exploit voxel-wise, pixel-wise, and point-wise features with expanding receptive fields for more informative RoI-wise representations  ...  PV-RCNN [6] leverages set abstraction operation among voxels instead of raw point clouds to achieve flexible receptive fields for fine-grained patterns while maintains computational efficiency.  ... 
arXiv:2104.10330v5 fatcat:62h4geokwnfvfjnbig4akrjzti

SIENet: Spatial Information Enhancement Network for 3D Object Detection from Point Cloud [article]

Ziyu Li, Yuncong Yao, Zhibin Quan, Wankou Yang, Jin Xie
2021 arXiv   pre-print
features for further box refinement.  ...  The predicted spatial shapes are complete and dense point sets, thus the extracted structure information contains more semantic representation.  ...  E-mail: liziyu@seu.edu.cn though point clouds could provide adequate information for neural networks to localize objects in 3D scenes, there are still many challenges in feature extraction and representation  ... 
arXiv:2103.15396v2 fatcat:7oar5xdsvrh2hk5gre2zsvxow4
« Previous Showing results 1 — 15 out of 53 results