
Research Article
Impact of Resolution and Resize Process on Face Recognition Accuracy with GAN-Generated Images
@INPROCEEDINGS{10.4108/eai.21-11-2024.2354613, author={Yu Xing}, title={Impact of Resolution and Resize Process on Face Recognition Accuracy with GAN-Generated Images}, proceedings={Proceedings of the 2nd International Conference on Machine Learning and Automation, CONF-MLA 2024, November 21, 2024, Adana, Turkey}, publisher={EAI}, proceedings_a={CONF-MLA}, year={2025}, month={3}, keywords={generative adversarial networks facenet image post-processing}, doi={10.4108/eai.21-11-2024.2354613} }
- Yu Xing
Year: 2025
Impact of Resolution and Resize Process on Face Recognition Accuracy with GAN-Generated Images
CONF-MLA
EAI
DOI: 10.4108/eai.21-11-2024.2354613
Abstract
This paper discusses the challenges facial recognition systems face in recognizing images generated by generative adversarial networks (GANs) and then proposes solutions by simple image processing methods. The study highlights that facial recognition models behave completely differently when dealing with low-resolution, super-resolution, or resized facial images and that the resizing method specifically affects the success rate of GANs. Changing the resolution also seems to affect the attack's success rate slightly.
Copyright © 2024–2025 EAI