PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
March 17, 2026Results in EngineeringOpen Access

ViFusion: Enhanced 3D Object Detection via Virtual Point Augmentation and Dynamic Multi-Source Feature Fusion

View Full Paper
Ask AI
Bookmark
Share

Authors

SSShaojing SongZXZhicheng XiangMIMuhammad Ilyas

Discussion

Loading...

Member takes

Overview

ViFusion demonstrates improved 3D detection accuracy in autonomous driving, indicating its significance in technology advancement.

Key Points

  • The aim is to enhance 3D object detection performance through efficient virtual point generation and feature fusion methods.
  • Developed ViFusion framework combining VPA and DMF.
  • VPA generates high-quality virtual points for object detection.
  • DMF facilitates dynamic multi-source feature fusion with optimized resource use.
  • Combined VPA and Ground Truth Augmentation to refine data quality.
  • Tested performance on the KITTI benchmark.
  • Achieved up to 5.84% improvement in Average Precision (AP) on Pointpillars for car detection.
  • Reached 76.65% AP for bicycles using Casa-V integrated with ViFusion.
  • Demonstrated efficient performance increase with reduced computational resources.

Cite This Study

Song et al. (2026) studied this question.

synapsesocial.com/papers/69b8ef6ddeb47d591b8c57f1https://doi.org/10.1016/j.rineng.2026.109983
View Full Paper
Ask AI
Bookmark
Share