Overview of Proposed Methods

1.3.1 Definitions of Terms

The definitions of related terms used in this study are described in this section in the following.

1. Cover media: Cover media, such as an image, a text-type document, an HTML document, or a video, is a file in which a message may be embedded.

2. Stego-video: A stego-video is a video in which a message has been embedded.

3. Public video: A public video is a video which is published by an author and can be downloaded by any user arbitrarily.

have been embedded.

5. Watermarked video: A watermarked video is a video in which visible or invisible watermarks have been embedded.

6. Captured image: A captured image is an image which is taken by a camera built in a cellular phone.

7. Protected image: A protected image is an image in which authentication signals have been embedded.

1.3.2 Brief Descriptions of Proposed Methods

1.3.2.1 Proposed Active Covert Communication Method for MPEG Videos

A steganographic method is proposed in this study for covert communication, which exploits the use of MPEG videos exhibited on public web pages to hide secret messages and transmits them to the receiver site via video downloading. First, MPEG videos are processed with variable-length decoding and the frequency-domain data of each frame are obtained. Then, secret messages can be embedded in MPEG videos with the use of every luminance block and motion vectors of macroblocks in each frame. After the embedding process is complete, these frames will be processed with variable length coding to generate stego-videos.

When a user wants to acquire the message from others, without installing a data extraction program, the user can browse a public web page and download stego-videos with secret messages.

1.3.2.2 Proposed Active Authentication Method for

MPEG Videos

A data hiding method for authentication is proposed in this study, which uses active agents to verify suspicious MPEG videos on web pages. Original videos are first processed with variable length decoding and the frequency-domain data of each frame can be obtained. Then, the index of the GOP of the video and the number of inter-coded frames in the GOP are embedded in the frames. After the embedding process is completed, these frames are processed with variable length coding to regenerate protected videos.

When a user wants to get the message from a protected MPEG video on a web page, without installing a video authentication program, the user may require the proposed system to verify the fidelity and the integrity of the suspicious videos and generate an authentication report.

1.3.2.3 Proposed Active Copyright Protection Method for MPEG videos

Two methods using digital watermarking techniques are proposed for copyright protection in this study, which use active agents to transform embedded watermarks in MPEG videos. Original videos are first processed with variable length decoding and the frequency domain data of each frame are obtained. The input watermark is embedded in MPEG videos with the use of the DC value of each luminance block of the macroblocks in each frame. After the embedding process is completed, these frames are processed with variable length coding to regenerate a watermarked video.

When an illicit user wants to download the watermarked video which is put on a

visible one and claim immediately the ownership of the MPEG video. Another reverse application situation is that multimedia providers may put a video with a removable visible watermark on a public web page for people to preview. An authorized user with an authentic key can access the proposed system and clear the visible watermark on the video displayed on the public web page. On the contrary, an unauthorized user can just see the watermarked video.

1.3.2.4 Proposed Active and Passive Covert Communication Method for Cover Images on Cellular Phones

Two steganographic methods are proposed in this study for covert communica-tion, which utilize cover images to embed secret messages and transmit them to the receiver site via the wireless network provided by telecommunication companies.

The first proposed method is one useful for active covert communication, in which secret messages are embedded in cover images using a 2-LSB data hiding technique and cover images are encapsulated with a JAVA program before being put on public web pages. Then, users can browse the public web page and download cover images with secret messages. In the other proposed method which is useful for passive covert communication, large-volume secret messages are first divided into several segments according to the data hiding capacity of each cover image and embedded in them using a 2-LSB data hiding technique. After the embedding process is completed, users just have to key in the receiver’s cellular phone number to the proposed system and the cover images will be transmitted to the receiver site via the wireless network. Both of these two methods have the capability of verifying the

fidelity of the secret messages.

1.3.2.5 Proposed Authentication Method for Captured Images on Cellular Phones

A data hiding method for authentication is proposed in this study, which can generate authentication signals to produce protected images. First, a captured image is divided into several 4×4 blocks and an input user key together with certain related information of the current block are used to generate a random integer as an authentication signal. With the use of a 2-LSB data hiding technique, an integer with the length of 32 bits can be embedded completely in a 4×4 block. After the embedding process is completed, the resulting protected images can be transmitted to a receiver site and the receiver can utilize the proposed system to verify the fidelity and the integrity of the received images.