SBIR Phase I: Novel photographic steganography to reduce digital piracy on live streaming platforms

点击翻译按钮获取中文摘要

The broader impact of this Small Business Innovation Research (SBIR) Phase I project will be to detect screen-to-camera video piracy, through first-in-kind development of a robust video piracy detection method that uniquely applies camera-display messaging to copyright protection, while maintaining network quality standards. Existing watermarking solutions cannot reliably handle video-in-video piracy, where users upload pirated media captured using a smartphone camera; either they fail to detect video piracy or exhibit injurious hypersensitivity. The proposed solution balances these two concerns by embedding unique and unextractable watermarks into video content, thereby improving the ability for platforms to accurately match uploaded content with messages in the copyright-protected database. The resulting technology will allow for the detection of copyright information in camera-captured images and video, so that distributors can automatically filter against the illegal sharing of copyright-protected material, dramatically reducing liability—and costly copyright infringement fines—for social media companies. This Small Business Innovation Research (SBIR) Phase I project seeks to develop an innovative camera-display message system that uses photographic steganography to embed covert messages in displayed video, addressing the need to detect screen-to-camera video piracy. Although photographic steganography has been applied to still images, video message embedding is complicated by the presence of perceptible artifacts that degrade visual quality. This project proposes to use novel spatio-temporal embedding that adapts to video content to reduce visual artifacts and support message recovery. This method circumvents the artifact-based constraints of traditional grid systems and will support the standard viewing quality. The following Phase I objectives will establish technical and commercial feasibility: 1) develop a video compression model that can handle multiple compression variations; 2) develop an optimal adaptive spatio-temporal embedding approach that reduces observed visual artifacts while maintaining a high accuracy rate; 3) design and implement a paired comparison protocol perceptual metric to evaluate video quality after message embedding; 4) develop a real-time robust screen detection and segmentation algorithm.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.