Anybody concerned with graphic design in whatsoever manner must be aware of the watermark put on images. This is done in order to protect the copyright violation of images. Visible watermarks are often used by stock photography websites. They are thin lines and shadows and are difficult to remove, but that is not impossible. Manually, removing watermarks is a tedious process and it may require a sophisticated editing tool. However, this is not impossible tool. But is it a reliable way to protect the image copyright.
It is to be noted that watermarks are added in a consistent manner to many images and this consistency can be used to invert the watermarking process. This can be done automatically without any difficulty. Google has recently reflected how they can be easily removed with the help of an artificial intelligence algorithm.
How it’s possible
Layering of images with the same watermark gives a rough pattern of the watermark. From here, a multi-optimization problem is used to find the watermark’s decomposition in the image and separate the watermark (the foreground) from the rest of the image (the background).
This makes the task of accurately removing the watermark easily from images containing it even if changing the opacity or position fails to bring any result. That is certainly not an ideal or good situation for image creators or those contributing to stock images. The good news is that Google is also offering how to stop doing this.
Warping the watermark will defeat the whole purpose of watermark-removing algorithm. The difference in the logo is virtually not visible to naked eye. Even applying the algorithm to images leaves traces of watermark. That is because the algorithm can’t estimate the warp field of the watermark used. Apart from that, there are many types of randomization that can be used to deal with watermark removal.
Google Research Blog says, “This optimization is able to produce very accurate estimations of the watermark components already from hundreds of images, and can deal with most watermarks used in practice, including ones containing thin structures, shadows or color gradients (as long as the watermarks are semi-transparent). Once the watermark pattern is recovered, it can be efficiently removed from any image marked by it.
