Multipurpose Watermarking Approach for Copyright and Integrity of Steganographic Autoencoder Models

With the great achievements of deep learning technology, neural network models have emerged as a new type intellectual property. Neural models’ design and training require considerable computational resources time. Watermarking is potential solution for achieving copyright protection integrity without excessively compromising accuracy stability. In this work, we develop multipurpose watermarking method securing steganographic autoencoder referred to “HiDDen.” This model used hide different kinds watermark messages in digital images. Copyright information embedded with imperceptibly modified parameters, verified by embedding Hash value generated from parameters. Experimental results show that proposed can reliably identify ownership localize tampered parts Furthermore, robustness are perfectly preserved.