Pillar-Based Object Detection for Autonomous Driving

Wang, Yue; Fathi, Alireza; Kundu, Anjan; Ross, David A.; Pantofaru, Caroline; Funkhouser, Thomas; Solomon, Justin

doi:10.1007/978-3-030-58542-6_2

Cited by 150 publications

(87 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Table 13, our proposed PV-RCNN-v2 framework outperforms previous state-of-the-art method [24] significantly with +1.74% mAP LEVEL 1 gain and +1.79% mAP LEVEL 2 gain for the most important vehicle detection. Table 13 also demonstrates that our proposed PV-RCNN-v2 framework also consistently outperforms all previous methods in terms of pedestrian and cyclist detection, including very recent Pillar-based method [64] and Part-A2-Net [24]. Compared with our preliminary work PV-RCNN-v1, our latest PV-RCNN-v2 framework achieves remarkably better mAP/mAPH on all difficulty levels for the detection of all three categories, while also increasing the processing speed from 3.3 FPS to 10 FPS for the 3D detection of 150m × 150m such a large area, which validates the effectiveness and efficiency of our proposed PV-RCNN-v2.…”

Section: D Detection On the Waymo Open Datasetmentioning

confidence: 75%

“…[62], [63] utilize multiple detection heads while [24] explores the object part locations for improving the performance. In addition, [64], [65] predicts bounding box parameters following the anchor-free paradigm. These grid-based methods are generally efficient for accurate 3D proposal generation but the receptive fields are constraint by the kernel size of 2D/3D convolutions.…”

Section: D Object Detection With Point Cloudsmentioning

confidence: 99%

See 1 more Smart Citation

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

Shi

Guo²,

Jiang³

et al. 2020

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

1,430

1,045

View full text Add to dashboard Cite

3D object detection is receiving increasing attention from both industry and academia thanks to its wide applications in various fields. In this paper, we propose the Point-Voxel Region based Convolution Neural Networks (PV-RCNNs) for accurate 3D detection from point clouds. First, we propose a novel 3D object detector, PV-RCNN-v1, which employs the voxel-to-keypoint scene encoding and keypoint-to-grid RoI feature abstraction two novel steps. These two steps deeply incorporate both 3D voxel CNN and PointNet-based set abstraction for learning discriminative point-cloud features. Second, we propose a more advanced framework, PV-RCNN-v2, for more efficient and accurate 3D detection. It consists of two major improvements, where the first one is the sectorized proposal-centric strategy for efficiently producing more representative and uniformly distributed keypoints, and the second one is the VectorPool aggregation to replace set abstraction for better aggregating local point-cloud features with much less resource consumption. With these two major modifications, our PV-RCNN-v2 runs more than twice as fast as the v1 version while still achieving better performance on the large-scale Waymo Open Dataset with 150m × 150m detection range. Extensive experiments demonstrate that our proposed PV-RCNNs significantly outperform previous state-of-the-art 3D detection methods on both the Waymo Open Dataset and the highly-competitive KITTI benchmark.

show abstract

Section: D Detection On the Waymo Open Datasetmentioning

confidence: 75%

Section: D Object Detection With Point Cloudsmentioning

confidence: 99%

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

Shi

Guo²,

Jiang³

et al. 2020

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

1,430

1,045

View full text Add to dashboard Cite

show abstract

“…CNNs have been applied to rendered images [34] or depth images [32] transformed from 3D data. Projection-based methods have also been proposed for object detection [15], [41], [53], scene segmentation [16], [47], registration [17], and 3D reconstruction [21], [28]. Furthermore, projection onto higherdimensional lattices was suggested for effective feature aggregation [33].…”

Section: Related Workmentioning

confidence: 99%

Projection-Based Point Convolution for Efficient Point Cloud Segmentation

Ahn

Yang

et al. 2022

IEEE Access

View full text Add to dashboard Cite

Understanding point cloud has recently gained huge interests following the development of 3D scanning devices and the accumulation of large-scale 3D data. Most point cloud processing algorithms can be classified as either point-based or voxel-based methods, both of which have severe limitations in processing time or memory, or both. To overcome these limitations, we propose Projection-based Point Convolution (PPConv), a point convolutional module that uses 2D convolutions and multi-layer perceptrons (MLPs) as its components. In PPConv, point features are processed through two branches: point branch and projection branch. Point branch consists of MLPs, while projection branch transforms point features into a 2D feature map and then apply 2D convolutions. As PPConv does not use point-based or voxelbased convolutions, it has advantages in fast point cloud processing. When combined with a learnable projection and effective feature fusion strategy, PPConv achieves superior efficiency compared to stateof-the-art methods, even with a simple architecture based on PointNet++. We demonstrate the efficiency of PPConv in terms of the trade-off between inference time and segmentation performance. The experimental results on S3DIS and ShapeNetPart show that PPConv is the most efficient method among the compared ones.

show abstract

“…It extracts the features from the pillar to assess the object of interest. This approach assesses the features rather than relaying on fixed encoders, and It tunes the model pillars instead of voxels [58]. In [59] a novel point pillar network has been designed based on quantization and pruning techniques for object detection, and they have used KITTI, Waymo, and NuScenes sets.…”

Section: Pillar-basedmentioning

confidence: 99%

Deep Learning Inspired Object Consolidation Approaches Using LiDAR Data for Autonomous Driving: A Review

Mekala

Park

Dhiman³

et al. 2021

Arch Computat Methods Eng

View full text Add to dashboard Cite

Autonomous Driving Vehicle (ADV) services have become a prominent motif in intelligent vehicle technology by adapting deep learning features. Automated driverless services are a hercules task due to the dynamic driving environment and the performance is deliberately reliant on the quality of data fusion from sensors. Therefore, considering advanced 3D LiDAR sensors is essential to measure the surrounding with 360 • coverage. However, accomplishing maximum autonomy is the main challenge because of debilitated and complex driving environmenst. Deep Learning (DL) based models potentially impact surrounding measures for object detection, classification, and tracking. This study describes the importance of DL-LiDAR strategies to formulate ADV research challenges followed by a comprehensive analysis of Semantic Segmentation, Data Fusion, Data Representation, Feature Extraction, Dynamic Object Detection, and Autonomous Driving-Multi-Objective Tracking mechanisms. Unlike existing review papers, we examine various deep learning influenced approaches and describe their precarious measurements. Also, we analyze the impact of systematic methods based on open-source data sets for objectx (x=classification, localization, tracking) consolidation. Most of the summaries confine scalability, and feasibility of DL techniques in ADV to accomplish an intelligent driving system without human interference. Open research problems are discussed in the last section of the article.

show abstract

Pillar-Based Object Detection for Autonomous Driving

Cited by 150 publications

References 36 publications

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

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

Projection-Based Point Convolution for Efficient Point Cloud Segmentation

Deep Learning Inspired Object Consolidation Approaches Using LiDAR Data for Autonomous Driving: A Review

Contact Info

Product

Resources

About