Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Engineered Biomimetic Nanomicelles Target Inflammation in Sepsis-Associated Acute Lung Injury by Scavenging ROS and Reprogramming Macrophages.

International journal of nanomedicine·2026
Same author

A glucose-enriched lung pre-metastatic niche triggered by matrix stiffness-tuned exosomal miRNAs in hepatocellular carcinoma.

Nature communications·2025
Same author

Toward autonomous vehicles: A survey on cooperative vehicle-infrastructure system.

iScience·2024
Same author

Author Correction: Early gastric cancer detection and lesion segmentation based on deep learning and gastroscopic images.

Scientific reports·2024
Same author

Early gastric cancer detection and lesion segmentation based on deep learning and gastroscopic images.

Scientific reports·2024
Same author

ACnerf: enhancement of neural radiance field by alignment and correction of pose to reconstruct new views from a single x-ray.

Physics in medicine and biology·2024

Related Experiment Video

Updated: Jul 9, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.9K

SMIFormer: Learning Spatial Feature Representation for 3D Object Detection from 4D Imaging Radar via Multi-View

Weigang Shi1, Ziming Zhu2, Kezhi Zhang2

  • 1School of Automotive Studies, Tongji University, Shanghai 201804, China.

Sensors (Basel, Switzerland)
|December 9, 2023
PubMed
Summary
This summary is machine-generated.

This study introduces SMIFormer, a novel network for 4D imaging radar, enhancing autonomous driving by fusing multi-view data (bird's-eye, front, side) to overcome point cloud sparseness and noise for improved 3D object detection.

Keywords:
3D object detection4D imaging radarautonomous drivingdeep learningmulti-view feature interactionpoint cloudvoxel feature decoupling

More Related Videos

Author Spotlight: Insights into the Analysis of Human Interaction with 3D Virtual Objects
06:36

Author Spotlight: Insights into the Analysis of Human Interaction with 3D Virtual Objects

Published on: October 18, 2024

998
Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

633

Related Experiment Videos

Last Updated: Jul 9, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.9K
Author Spotlight: Insights into the Analysis of Human Interaction with 3D Virtual Objects
06:36

Author Spotlight: Insights into the Analysis of Human Interaction with 3D Virtual Objects

Published on: October 18, 2024

998
Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis
05:41

Author Spotlight: Integrating Ultrasound Imaging with Biochemical Markers for Thyroid Disease Diagnosis

Published on: February 9, 2024

633

Area of Science:

  • Robotics and Autonomous Systems
  • Computer Vision
  • Sensor Fusion

Background:

  • 4D millimeter wave (mmWave) imaging radar offers cost-effective and weather-robust sensing for autonomous driving.
  • Challenges include sparse and noisy point clouds, limiting practical application.
  • Existing methods struggle with insufficient feature representation from single-view data.

Purpose of the Study:

  • To introduce SMIFormer, a multi-view feature fusion network for 4D radar.
  • To address the limitations of sparse and noisy point clouds in 4D radar data.
  • To improve 3D object detection performance in autonomous driving systems.

Main Methods:

  • Developed SMIFormer, a network framework for single-modal 4D radar input.
  • Decoupled 3D scenes into bird's-eye view (BEV), front view (FV), and side view (SV).
  • Proposed multi-view feature interaction (MVI) for intra-view and cross-view feature integration.

Main Results:

  • Evaluated SMIFormer on the View-of-Delft (VoD) dataset.
  • Achieved mean Average Precision (mAP) of 48.77% in the fully annotated area.
  • Reached an mAP of 71.13% in the driving corridor area.

Conclusions:

  • SMIFormer effectively models 3D scenes by integrating multi-view radar data.
  • The multi-view approach overcomes single-view limitations caused by sparse point clouds.
  • 4D radar demonstrates significant potential for advancing 3D object detection in autonomous driving.