PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
April 17, 2026PeerJ Computer ScienceOpen Access

YOLO-augment strategy with diffusion-based inpainting for enhanced traffic sign detection

View Full Paper
Ask AI
Bookmark
Share

Authors

CSChayanon Sub-r-paPPPraphan PavarangkoonSHSu-Wen Huang

Discussion

Loading...

Member takes

Overview

Novel dataset augmentation enhances traffic sign detection in autonomous driving, indicating improved training data quality.

Key Points

  • This research aims to develop a dataset augmentation method to enhance traffic sign detection for autonomous driving systems.
  • Utilizes Stable Diffusion inpainting to generate realistic synthetic traffic signs.
  • Introduces an object-size-based crop (OSB-crop) technique for mask adjustments.
  • Evaluates augmentation quality using the Fréchet Inception Distance (FID).
  • Achieves average FID scores of 195.85 for the DFG-T10 subset and 247.077 for DFG-B10 subset.
  • Demonstrates enhanced realism in traffic signs, especially for minority classes.
  • Qualitative analyses show effective integration of synthetic signs into real-world scenes.

Cite This Study

Sub-r-pa et al. (2026) studied this question.

synapsesocial.com/papers/69e1ce605cdc762e9d8576b2https://doi.org/10.7717/peerj-cs.3778
View Full Paper
Ask AI
Bookmark
Share