Multi-Feature Fusion Based Auto-Encoder for Perceptual Authentication Hashing

In recent decades, numerous perceptual authentication hashing schemes have been proposed for image content authentication. However, existing schemes fail to provide satisfactory robustness and discrimination capability when facing complex manipulations in real-world scenarios. In this work, we introduce a multi-feature fusion based auto-encoder for perceptual authentication hashing. Specifically, we initially introduce a joint encoder of Convolutional Auto-Encoder (CAE) and Variational Auto-Encoder (VAE) into the field of perceptual authentication hashing. We then integrate them using a feature enhancement module, ultimately transforming it into a hash sequence for image authentication. Experimental results on a large-scale dataset demonstrate that our scheme exhibits superior performance compared to state-of-the-art schemes.