TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Papers/FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection

FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection

Danila Rukhovich, Anna Vorontsova, Anton Konushin

2021-12-01object-detection3D Object DetectionObject Detection
PaperPDFCodeCode(official)

Abstract

Recently, promising applications in robotics and augmented reality have attracted considerable attention to 3D object detection from point clouds. In this paper, we present FCAF3D - a first-in-class fully convolutional anchor-free indoor 3D object detection method. It is a simple yet effective method that uses a voxel representation of a point cloud and processes voxels with sparse convolutions. FCAF3D can handle large-scale scenes with minimal runtime through a single fully convolutional feed-forward pass. Existing 3D object detection methods make prior assumptions on the geometry of objects, and we argue that it limits their generalization ability. To get rid of any prior assumptions, we propose a novel parametrization of oriented bounding boxes that allows obtaining better results in a purely data-driven way. The proposed method achieves state-of-the-art 3D object detection results in terms of mAP@0.5 on ScanNet V2 (+4.5), SUN RGB-D (+3.5), and S3DIS (+20.5) datasets. The code and models are available at https://github.com/samsunglabs/fcaf3d.

Results

TaskDatasetMetricValueModel
Object DetectionSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
Object DetectionSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
Object Detection3RScanmAP@0.2560.1FCAF3D
Object Detection3RScanmAP@0.542.6FCAF3D
Object DetectionScanNet++mAP@0.2522.3FCAF3D
Object DetectionScanNet++mAP@0.511.4FCAF3D
Object DetectionS3DISmAP@0.2566.7FCAF3D
Object DetectionS3DISmAP@0.545.9FCAF3D
Object DetectionMultiScanmAP@0.2553.8FCAF3D
Object DetectionMultiScanmAP@0.540.7FCAF3D
Object DetectionScanNetV2mAP@0.2571.5FCAF3D
Object DetectionScanNetV2mAP@0.557.3FCAF3D
3DSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
3DSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
3D3RScanmAP@0.2560.1FCAF3D
3D3RScanmAP@0.542.6FCAF3D
3DScanNet++mAP@0.2522.3FCAF3D
3DScanNet++mAP@0.511.4FCAF3D
3DS3DISmAP@0.2566.7FCAF3D
3DS3DISmAP@0.545.9FCAF3D
3DMultiScanmAP@0.2553.8FCAF3D
3DMultiScanmAP@0.540.7FCAF3D
3DScanNetV2mAP@0.2571.5FCAF3D
3DScanNetV2mAP@0.557.3FCAF3D
3D Object DetectionSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
3D Object DetectionSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
3D Object Detection3RScanmAP@0.2560.1FCAF3D
3D Object Detection3RScanmAP@0.542.6FCAF3D
3D Object DetectionScanNet++mAP@0.2522.3FCAF3D
3D Object DetectionScanNet++mAP@0.511.4FCAF3D
3D Object DetectionS3DISmAP@0.2566.7FCAF3D
3D Object DetectionS3DISmAP@0.545.9FCAF3D
3D Object DetectionMultiScanmAP@0.2553.8FCAF3D
3D Object DetectionMultiScanmAP@0.540.7FCAF3D
3D Object DetectionScanNetV2mAP@0.2571.5FCAF3D
3D Object DetectionScanNetV2mAP@0.557.3FCAF3D
2D ClassificationSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
2D ClassificationSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
2D Classification3RScanmAP@0.2560.1FCAF3D
2D Classification3RScanmAP@0.542.6FCAF3D
2D ClassificationScanNet++mAP@0.2522.3FCAF3D
2D ClassificationScanNet++mAP@0.511.4FCAF3D
2D ClassificationS3DISmAP@0.2566.7FCAF3D
2D ClassificationS3DISmAP@0.545.9FCAF3D
2D ClassificationMultiScanmAP@0.2553.8FCAF3D
2D ClassificationMultiScanmAP@0.540.7FCAF3D
2D ClassificationScanNetV2mAP@0.2571.5FCAF3D
2D ClassificationScanNetV2mAP@0.557.3FCAF3D
2D Object DetectionSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
2D Object DetectionSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
2D Object Detection3RScanmAP@0.2560.1FCAF3D
2D Object Detection3RScanmAP@0.542.6FCAF3D
2D Object DetectionScanNet++mAP@0.2522.3FCAF3D
2D Object DetectionScanNet++mAP@0.511.4FCAF3D
2D Object DetectionS3DISmAP@0.2566.7FCAF3D
2D Object DetectionS3DISmAP@0.545.9FCAF3D
2D Object DetectionMultiScanmAP@0.2553.8FCAF3D
2D Object DetectionMultiScanmAP@0.540.7FCAF3D
2D Object DetectionScanNetV2mAP@0.2571.5FCAF3D
2D Object DetectionScanNetV2mAP@0.557.3FCAF3D
16kSUN-RGBD valmAP@0.2564.2FCAF3D (Geo only)
16kSUN-RGBD valmAP@0.548.9FCAF3D (Geo only)
16k3RScanmAP@0.2560.1FCAF3D
16k3RScanmAP@0.542.6FCAF3D
16kScanNet++mAP@0.2522.3FCAF3D
16kScanNet++mAP@0.511.4FCAF3D
16kS3DISmAP@0.2566.7FCAF3D
16kS3DISmAP@0.545.9FCAF3D
16kMultiScanmAP@0.2553.8FCAF3D
16kMultiScanmAP@0.540.7FCAF3D
16kScanNetV2mAP@0.2571.5FCAF3D
16kScanNetV2mAP@0.557.3FCAF3D

Related Papers

A Real-Time System for Egocentric Hand-Object Interaction Detection in Industrial Domains2025-07-17RS-TinyNet: Stage-wise Feature Fusion Network for Detecting Tiny Objects in Remote Sensing Images2025-07-17Decoupled PROB: Decoupled Query Initialization Tasks and Objectness-Class Learning for Open World Object Detection2025-07-17Dual LiDAR-Based Traffic Movement Count Estimation at a Signalized Intersection: Deployment, Data Collection, and Preliminary Analysis2025-07-17Vision-based Perception for Autonomous Vehicles in Obstacle Avoidance Scenarios2025-07-16Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping2025-07-15ECORE: Energy-Conscious Optimized Routing for Deep Learning Models at the Edge2025-07-08Beyond One Shot, Beyond One Perspective: Cross-View and Long-Horizon Distillation for Better LiDAR Representations2025-07-07