PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
January 17, 2026Discover Artificial IntelligenceOpen Access

HOGE: integrating feature descriptor and transfer learning for masked face recognition

View Full Paper
Ask AI
Bookmark
Share

Authors

MYMing Chun YoSCSiew Chin ChongLCLee-Ying Chong

Discussion

Loading...

Member takes

Overview

Novel approach improves masked face recognition, suggesting effective techniques for occlusion handling.

Key Points

  • This research aims to enhance masked face recognition by integrating feature descriptors with deep learning techniques.
  • Introduced HOGE, which uses HOG images as input for a modified EfficientNetV2-S model.
  • Implemented an additional convolution layer for processing greyscale masked face images.
  • Evaluated the performance using two benchmark datasets: LFW-SMFRD and RMFRD.
  • Achieved 97.41% accuracy on the LFW-SMFRD dataset.
  • Achieved 99.38% accuracy on the RMFRD dataset.
  • Demonstrated effective recognition of masked face images across both datasets.

Cite This Study

Yo et al. (2026) studied this question.

synapsesocial.com/papers/696b2616d2a12237a9349699https://doi.org/10.1007/s44163-025-00819-3
View Full Paper
Ask AI
Bookmark
Share