With the growing popularity of the Internet, digital images are being used and transferred more frequently. While this facilitates easy access to information, it also raises concerns about security and intellectual property rights violations through illegal use, copying, and digital content theft. To address these issues, watermarks are commonly used in digital images to maintain security by proving ownership. Watermarks can be either text or an image placed overtly or covertly in an image, and they are expected to be challenging to remove. In this paper titled "Convolutional Neural Network-Based Image Watermarking using Discrete Wavelet Transform," Alireza Tavakoli et al. propose a novel approach for watermarking using a combination of convolutional neural networks (CNNs) and wavelet transforms. The proposed network is designed for embedding and extracting watermarks and is independent of the host image resolution. It can accept all kinds of watermarks while maintaining performance with only 11 layers. The performance of the watermarking network is measured based on two terms: the similarity between the extracted watermark and the original one, and the similarity between the host image and the watermarked one. By leveraging CNNs and wavelet transforms, this approach aims to enhance both robustness against attacks such as affine transformation and JPEG compression as well as adaptability to different resolutions of host images. The authors also provide some additional context by referencing related works in the field. Xiangui Kang proposed a blind DWT-DFT composite image watermarking algorithm that is robust against affine transformation and JPEG compression. Another approach called ReDMark utilizes an end-to-end autoencoder with a simulated DCT layer in its deep network to achieve robustness against JPEG attacks. Jae-Eun Lee proposed a robust watermarking neural network that includes an attack simulation component and adjusts its performance based on the resolution of both the host image and watermark. Overall, this paper presents a comprehensive approach for digital image watermarking using a combination of CNNs and wavelet transforms which offers robustness, adaptability, as well as performance while addressing security concerns protecting intellectual property rights in today's digital age.
The Internet is becoming more popular, so people are using and sharing digital pictures more. Some people are worried about security and protecting the rights of the people who made the pictures. Watermarks are marks that show who owns a picture and help keep it safe. Watermarks can be words or pictures that are put on a picture to show who owns it. Alireza Tavakoli and others have come up with a new way to put watermarks on pictures using special computer programs. This new way works well with different sizes of pictures and can protect them from being changed or copied.
Convolutional Neural Network-Based Image Watermarking using Discrete Wavelet Transform
With the growing popularity of the Internet, digital images are being used and transferred more frequently. While this facilitates easy access to information, it also raises concerns about security and intellectual property rights violations through illegal use, copying, and digital content theft. To address these issues, watermarks are commonly used in digital images to maintain security by proving ownership. Watermarks can be either text or an image placed overtly or covertly in an image, and they are expected to be challenging to remove.
In this paper titled "Convolutional Neural Network-Based Image Watermarking using Discrete Wavelet Transform," Alireza Tavakoli et al. propose a novel approach for watermarking using a combination of convolutional neural networks (CNNs) and wavelet transforms. The proposed network is designed for embedding and extracting watermarks and is independent of the host image resolution. It can accept all kinds of watermarks while maintaining performance with only 11 layers.
Performance Measurement
The performance of the watermarking network is measured based on two terms: the similarity between the extracted watermark and the original one, and the similarity between the host image and the watermarked one. By leveraging CNNs and wavelet transforms, this approach aims to enhance both robustness against attacks such as affine transformation and JPEG compression as well as adaptability to different resolutions of host images.
Related Works
The authors also provide some additional context by referencing related works in the field. Xiangui Kang proposed a blind DWT-DFT composite image watermarking algorithm that is robust against affine transformation and JPEG compression. Another approach called ReDMark utilizes an end-to-end autoencoder with a simulated DCT layer in its deep network to achieve robustness against JPEG attacks. Jae-Eun Lee proposed a robust watermarking neural network that includes an attack simulation component and adjusts its performance based on the resolution of both the host image
With the growing popularity of the Internet, digital images are used and transferred more frequently. Although this phenomenon facilitates easy access to information, it also creates security concerns and violates intellectual property rights by allowing illegal use, copying, and digital content theft. Using watermarks in digital images is one of the most common ways to maintain security. Watermarking is proving and declaring ownership of an image by adding a digital watermark to the original image. Watermarks can be either text or an image placed overtly or covertly in an image and are expected to be challenging to remove. This paper proposes a combination of convolutional neural networks (CNNs) and wavelet transforms to obtain a watermarking network for embedding and extracting watermarks. The network is independent of the host image resolution, can accept all kinds of watermarks, and has only 11 layers while keeping performance. Performance is measured by two terms; the similarity between the extracted watermark and the original one and the similarity between the host image and the watermarked one.
