Data Hiding for Video-in-Video

1
Data Hiding for Video-in-Video
Mitchell D. Swanson BinZhu Ahmed
H.Tewfik Department of Electrical and Computer
Engineering, University of Minnesota
STUDENT Chia-Feng Yeh SUID 259357749
E-Mail cyeh04_at_ecs.syr.edu
2
Abstract

• How to hide data video-in-video
• The secondary video into a digital video
  stream by
• directly modifying the pixels in in the video
  frames.
• This technique requires no separate channel
  or bit
• interleaving to transmit the extra
  information
• The data is invisible embedded using a
  perception-
• based and quantization algorithm.

3
Introduction

• The hidden data may include real-time video and
  speech, text, hypertext, image data, etc. The
  hidden data remains intact as the host video is
  compressed, stored, transmitted, etc. In general,
  the embedded data may be extracted from the video
  stream automatically, or when a user selects a
  particular spatial/temporal region.To ensure high
  visual quality, the embedded data is perceptually
  invisible within the host media.

4
Introduction(cont.)

• Note that data hiding embeds information directly
  into the pixels in each frame. There are several
  advantages over storing the data in a file header
  a separate file.
• The storing the data in a file header a separate
  files disadvantage
• 1.Header data may be easily lost when the
  format of
• the video is changed or the video is
  cropped
• 2.Separate files require care when
  transmitting the
• video.

5
Introduction(cont.)

• 3.These approaches also require additional
  storage.
• Hiding data directly into the video data resolves
  these problem.Furthermore, by hiding information
  in specific spatial or temporal regions, the
  hidden data remains in these regions if they are
  removed and pasted into another file or document.
• Our approach also provides users with an
  additional flexibility. The hidden data can be
  made accessible to different levels of users,
  depending on the wisher of the author.

6
Introduction(cont.)

• In particular, our projection based data hiding
  scheme can use private author key, allowing only
  privileged users to access the data. If the
  author wishes global access to the data, keys
  available to the general public may be used.
• Our video dependent scheme employs a masking
  model to guarantee the hidden data is invisible
  and to maximize the strength of the hidden data.

7
Video Data Hiding

• Our data hiding approach is based on two steps
• 1.linear projection, a common linear
  algebra
• operation
• 2.quantization and perturbation
• The video data hiding algorithm is shown in
  Fig1.

8
Video Data Hiding(cont.)
9
Result

• How many data can we embed in a video sequence?
• For a frame of size RC, and embedding N bits
  per nm block, we can embed N R/n C/m bits per
  frame. At 30 frames per second, we can embed
• N R/n C/m 30 bit per second.
• A powerful application of this algorithm is
  embedding video-in-video. Unlike the
  Picture-in-Picture technology present in many
  current television sets, video-in-video does not
  use separate data streams for each video sequence.

10
Result(cont.)

• Only the host video stream is transmitted. The
  embedded video stream may be extracted and
  displayed in a window within the host video,
  streamed to second display device, or streamed to
  a file.
• Speech is also easily supported by the video data
  hiding algorithm.