Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{205185,
        author = {Dr. Deepti Varshney and Arpan Gujar and Anshul Patel and Sharvari Arun Utekar},
        title = {Deep Learning-Based Invisible Watermarking System: A CNN Encoder-Decoder Approach for Robust Logo Embedding and Extraction},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {5457-5471},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=205185},
        abstract = {Protecting media rights and preventing unapproved sharing is why digital watermarks matter. Digital watermarks help safeguard media rights and stop unauthorized distribution. A new method for setting up image watermarks uses deep learning tools. This approach employs deep learning to create the image watermark setup.
It works by placing the image in a system that reduces its size. Then, it adds a logo to the image. The logo remains hidden on the image. The system also includes a network that can locate the logo again, even if the image is damaged. The system continuously learns how to do this. It measures its performance by assessing the clarity of the image and the logo’s detectability.
Colors are also crucial for watermarks. Thus, the system follows a rule to ensure the colors remain accurate. This rule prevents unwanted changes to the image’s colors. The system was tested on numerous images, and the results showed that the digital watermarks were very difficult to see. The system performs well even when the image is damaged. It can still locate the logo when the image is blurry or noisy. The system is built using PyTorch, employing a framework for learning and a setup for processing requests. When it receives an image, processing takes only a fraction of a second. The system is also compact, making it suitable for production environments without issues. Digital watermarks are important, and this system offers a new method for creating them.},
        keywords = {Digital Watermarking, Deep Learning, Convolu-tional Neural Networks, Encoder-Decoder Architecture, Residual Learning, Image Authentication, Logo Extraction, Robustness, PSNR, SSIM.},
        month = {June},
        }