|
This article is cited in 2 scientific papers (total in 2 papers)
IMAGE PROCESSING, PATTERN RECOGNITION
An improved video watermarking algorithm with extraction using a mobile device camera
O. O. Evsutin, A. S. Melman, D. I. Podbolotov, A. G. Stankevich National Research University Higher School of Economics, Moscow
Abstract:
The use of screen-capture-resistant digital watermarks is a promising way to store information invisibly in a video stream for later retrieval by the user using a smartphone camera. However, the development of algorithms that implement this scenario is associated with the problem of balancing between the imperceptibility of embedding and robustness. A serious problem is the extraction of watermarks using a mobile device. Most people use the vertical positioning of the smartphone when shooting, which excludes the possibility of only marked video sequences entering the frame. The extraction algorithm first finds the screen area in the image and then extracts the watermark under various distortion conditions. This study proposes an approach to improve the efficiency of the algorithm for embedding digital watermarks into video data based on rectangular patterns, which provides resistance to screen shooting. The proposed approach to increasing the embedding imperceptibility provided an increase in the PSNR and SSIM values by 17.18% and 7.90%, respectively. The use of a neural network at the extraction stage reduced the BER value by 64.64%.
Keywords:
digital watermark, video, neural network, screen capture
Received: 27.04.2023 Accepted: 07.07.2023
Citation:
O. O. Evsutin, A. S. Melman, D. I. Podbolotov, A. G. Stankevich, “An improved video watermarking algorithm with extraction using a mobile device camera”, Computer Optics, 47:6 (2023), 972–979
Linking options:
https://www.mathnet.ru/eng/co1201 https://www.mathnet.ru/eng/co/v47/i6/p972
|
Statistics & downloads: |
Abstract page: | 4 | Full-text PDF : | 6 |
|