Deep Watermarking Based on Swin Transformer for Deep Model Protection
This study improves existing protection strategies for image processing models by embedding invisible watermarks into model outputs to verify the sources of images. Most current methods rely on CNN-based architectures, which are limited by their local perception capabilities and struggle to effectiv...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-05-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/10/5250 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|