Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 25, 2026Scientific ReportsOpen Access

WTCFNet for industrial defect detection using wavelet transform and cross layer feature fusion

View Full Paper
Ask AI
Bookmark
Share

Authors

HCHao ChenYRYu-Bo Ren

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates improved defect detection in industrial products, suggesting enhanced quality control.

Key Points

  • The aim is to enhance the detection of industrial surface defects by improving feature representation and recognition under complex backgrounds.
  • Developed WTCF-Net utilizing wavelet transform to separate image features and enhance defect representation.
  • Created an Interactive Residual Module for optimal aggregation of multi-dimensional features.
  • Introduced a Cross-level Feature Aggregation Network for better interaction among different scales.
  • Model outperformed baseline with mAP increases of 5.9%, 1.3%, and 1.4% on NEU-DET, PCB and DeepPCB datasets respectively.
  • Achieved a detection speed of 53 frames per second, indicating high efficiency.

Cite This Study

Chen et al. (2026) studied this question.

synapsesocial.com/papers/6a13e78b0e02ee3982d32299https://doi.org/10.1038/s41598-025-27745-9
View Full Paper
Ask AI
Bookmark
Share