Fast and Secured Visual Content Hiding in Lossy Compressed Images and Video Streams

This paper describes a reversible visual data hiding scheme. From general Red-Green-Blue image or video footage, a meaningless appearing version of it is created with an optional encrypted set of keys. The encrypted image or video frame appears as random cues, and it is not visually recognisable by general viewers. They can be saved as lossy compressed JPEG image files or lossy compressed MP4 video files. These can be transferred over the public domains such as DropBox, Pixabay, YouTube, Dailymotion, Vimeo, or other public sharing servers. These providers allow anyone broadcast (upstream) images or video feeds, and the public domains can view (downstream) these easily. The main concern is the privacy of the uploaded contents. YouTube, for instance, is relatively simple to set up a free account, and the cost is almost zero (no dedicated server, no maintenance and no storage needed). However, before allowing to download, the images/videos usually are lossy compressed. The quality is reduced for smaller storage and faster transferring rate; although, they may be ‘visually lossless‘. At the receiving end, the image or video frame can be decrypted, and we can achieve its perceivable version (close to the original up-streaming data). This technique allows general users to take advantages of free public image/video hosting services. This novel visual content hiding method gives them a worry-free tool to upload sensitive or insensitive contents without bothering about the leaks of the data (even the hosts can not visualise them correctly without a key). In this paper, the detail of the proposed method is discussed. Results from lossless and various lossy images/videos sequences are examined.