Transformation Process for Visible Watermarks

Chapter 5 Active Copyright Protection of MPEG Videos

5.4 Proposed Method for Transforming Invisible Watermarks Actively to Visible

5.4.2 Transformation Process for Visible Watermarks

At the beginning of the proposed process for transforming invisible watermarks into visible ones, users can see clear MPEG videos on a public web page.

VLD

Frame Type?

Apply Transformation Process

I P or B

VLC Watermarked

Video

Recovered Video

Figure 5. 8 A flowchart of the transformation process for visible watermarks.

Everyone can see these videos without paying anything, but if someone wants to download them, the proposed system will extract the hidden active agents and transform the embedded invisible watermarks into visible ones with random intensity values in grayscale. A flowchart of the current process is shown in Figure 5.8.

u s e r

Server-Site Preparing Process

Web Page

Server-Site Program

MPEG Videos

Extract Active Agents Transform

Invisible Watermarks into

Visible Ones

Request for downloading MPEG videos

Active Video Player Watermarked MPEG Videos

Figure 5. 9 An illustration of the proposed method for transforming invisible

After transforming invisible watermarks into visible ones, visible watermarks can immediately show the ownership of the MPEG videos. Because these visible watermarks are with random intensity values in grayscale, they are not easy to remove.

An illustration of the proposed method is shown in Figure 5.9.

5.5 Experimental Results

In our experiments, two applications for active copyright protection of MPEG videos are elaborated. One is a transformations of visible watermarks into invisible ones actively by an active agent, and another is an inverse process of the previous one.

In the first experiment, while a user browses a web page, the proposed system presented in Figure 5.10(a) shows a video with a visible watermark and an active video player. If the user wants to see a clear video displayed in Figure 5.10(b), an authentic key should be provided to the system. Otherwise, a degraded video still shows on the screen.

In the second experiment, an MTV video and a description are displayed on a public web page in Figure 5.11(a). While a user browses this web page, a clear video with an invisible watermark is presented. If the user calls a pop-up menu shown in Figure 5.11(b) and wants to download the video, the proposed system extracts an active agent and transform the invisible watermarks into visible ones actively. After the downloading process is completed as shown in Figure 5.11(d), the video with a visible watermarks shown on the local computer, as shown in Figure 5.11(e).

(a)

(b)

Figure 5.10 (a) A video with a visible watermark. (b) A clear video with an invisible watermark after an authentic key is provided.

User key field

Input button

(a)

(b)

Figure 5.11 (a) A clear video with an invisible watermark. (b) A saving button on a pop-up menu. (c) A dialog which can select the saving location of the video. (d) The downloaded video on the desktop. (e) A video with a visible watermark in the local computer. (continued)

Pop-up menu

(c)

(d)

Dialog

Downloaded video

(e)

5.6 Summary and Discussions

In this chapter, two methods for protecting the copyright of MPEG videos have been proposed and tested. Both of them can display the ownership of MPEG videos.

In fact, the proposed method in Section 5.3 may be regarded as the inverse process of the proposed method in Section 5.4.

The application environment of the experiments for transforming invisible watermarks into visible ones is simulated as a real Internet Explorer program and illicit users does not easily notice anything different. After illegal downloading of the MPEG videos, users will get watermarked ones. Based on our experimental results, the proposed methods may be seen to be useful for real applications.

Chapter 6 Active and Passive Large-Volume Covert Communication by Cover Images with Secret Authentication Capability on Cellular Phones

6.1 Introduction

With the advance of mobile computing technologies, more and more electronic devices, such as personal digital assistants, cellular phones, and notebooks, support the ability of executing programming languages. Since these mobile devices can do so, various applications can be developed on them now.

Due to the popularity of using cellular phones and the progress of technology developments of them, the platform of a cellular phone may be adopted to perform data hiding applications now, as done in this study. And the applications can also be implemented on other electronic mobile devices which support JAVA programming languages.

In Section 6.2, an active covert communication method which carry out the method on the platform of a personal computer is proposed. In Section 6.2, a passive covert communication method which supports transmitting large-volume secret messages by multiple cover images is proposed. Finally, some experimental results

and a summary are given.

6.2 Proposed Active Covert

Communication Method for Cover Images on Cellular Phones

The original idea of the proposed application of active covert communication on cellular phones comes from that of the active covert communication method implemented on personal computers, in which it is desired to extract secret messages from cover media without the need of installing data hiding programs. The proposed method is that each cellular phone supporting JAVA programming languages can get secret messages from cover media on web pages.

In Sections 6.2.1 and 6.2.2, both the embedding process and the extraction process for secret messages hidden in cover media are introduced.

6.2.1 Process for Embedding Secret Messages

The secret message embedding process is implemented on the platform of a server-site personal computer. An illustration of the process is presented in Figure 6.1.

First, an authentication signal has to be generated by the secret messages and the user key. The method of calculating authentication signals is the same as that proposed in Chapter 3. After generating an authentication signal, it can be hidden behind the secret messages and the proposed system can utilize a user key to randomize it. Because the size of the display screen on the cellular phone is usually small, that of the cover image should be small enough so as to be displayed appropriately on the cellular phone screen. For the sake of the limited data hiding

capacity of the cover image, a 2-LSB data hiding technique is adopted in this chapter.

Every bit of randomized secret messages with an authentication signal is embedded in the cover image by modifying the two least significant bits of each pixel.

Cover Image (C)

Secret Messages

(M)

Calculate and Embed Authentication

Signals

Transform New Secret Messages (M’)

into Binary Formula

Embed New Secret Messages

Key Randomize

Cover Image with Secret Messages

(C’)

Figure 6. 1 An illustration of the process for embedding secret messages.

6.2.2 Process for Extracting Secret Messages

The extraction process is implemented on the platform of a cellular phone. An illustration of the process is shown in Figure 6.2. It is the inverse process the

2-LSB data hiding technique. After extracting all the hidden data, an input user key is used to de-randomize them and obtain an extracted authentication signal. Then, the extracted secret messages have to be used to regenerate an authentication signal. After comparing the extracted authentication signal and the regenerated one, if these two signals are the same, the extracted secret messages are considered as correct and kept;

otherwise, they will be discarded.

Recalculate

Figure 6. 2 An illustration of the process for extracting secret messages.

A flowchart of the proposed method is shown in Figure 6.3. After preparing a cover image with secret messages, they are put into a JAVA program on a web page for people to download. When a user downloads the JAVA program, the cover image

is displayed on the screen of a cellular phone. If the user wants to extract the secret messages from the cover image, a button on the cellular phone has to be pressed and an authentic user key should be provided for correct extraction.

Key

Server-Site Preparing Process

Web Page

Java Program

Cover Images

Extract Active Agents Extract and Authenticate Secret Messages Request for downloading secret messages

from cover images

Active Image Player Secret Messages

Correct Secret Messages?

Discard N

Figure 6. 3 A flowchart of the proposed active covert communication method for

6.3 Proposed Passive Large-Volume

Covert Communication Method for Cover Images on Cellular Phones

Using cellular phones to exchange secret messages is a kind of insecure behavior.

No matter utilizing the voice conversation or sending short messages, both of these two methods let secret messages exposed under the public transmission environment.

A passive covert-communication method for cover images on cellular phones is proposed to take this security problem into consideration. In order to transmit large-volume secret messages, both a dividing process and a combination one of the secret messages are proposed.

In Section 6.3.1, a method for dividing and combining large-volume secret messages is introduced. In Section 6.3.2 and Section 6.3.3, both an embedding process and an extraction process are presented.

6.3.1 Method for Dividing and Combining Large-Volume Secret Messages

Because the size of cover images is usually small, the corresponding data hiding capacity is usually not very large. To solve this problem, a method for dividing and combining large-volume secret messages is proposed.

An illustration of the dividing method is shown in Figure 6.4. According to the data hiding capacity of each cover image, secret messages are divided into several segments. Each segment has its own authentication signal and is randomized by an input user key.

Cover Images

Figure 6. 4 An illustration of the process for dividing secret messages into segments.

An illustration of the combination method is shown in Figure 6.5. Each segment must be de-randomized by an input user key and then an authentication signal would be regenerated. After verifying the fidelity of the secret messages of each segment, the proposed system can combine them according to their original sequence. If there is any incorrect segment, the receiver should request for retransmissions of correct segments.

Large-Volume

Figure 6. 5 An illustration of the process for combining segments of secret messages.

6.3.2 Process for Embedding Secret Messages in

Multiple Cover Images

We all know that keying in large-volume messages on a cellular phone is not an easy job. With the use of wireless transmission devices of infrared rays or Bluetooth, a user can prepare all messages on a personal computer easily in advance and then use these wireless devices to transmit the messages into a cellular phone.

Cover Images

Figure 6. 6 An illustration of the process for embedding secret messages into multiple

After dividing large-volume secret messages into segments, the proposed system utilizes a 2-LSB data hiding technique to embed each segment in the corresponding cover image. An illustration of the process is shown in Figure 6.6.

6.3.3 Process for Extracting Secret Messages from Multiple Cover Images

An illustration of the process of extracting secret messages from multiple cover images is shown in Figure 6.7.

Large-Volume

Figure 6. 7 An illustration of the process for extracting secret messages from multiple

cover images.

On the receiver site, after collecting all the cover images from the sender site, the proposed system can use an input user key to authenticate each segment of the secret messages and do the combination manipulation.

Large-Volume

Figure 6. 8 A flowchart of the proposed passive large-volume covert communication method for cover images on cellular phones.

A flowchart of the proposed method is shown in Figure 6.8. The sender only needs to input the telephone number of the receiver’s cellular phone and every cover image will be encapsulated as many short message packets and transmitted to the receiver through the wireless transmission of the short messaging system (SMS). At the receiver site, the proposed system can combine each packet to form a complete image and utilizes an input user key to extract and recover the hidden secret messages.

6.4 Experimental Results

In our experiments, two applications, namely, active and passive covert communication, via cover images are elaborated.

(a) (b) Figure 6.9 An experimental result. (a) A browser in the cellular phone. (b) A public

web page. (c) Downloading the JAVA program. (d) An icon of the program. (e) The execution screen of the program. (f) A success extraction with a user key 123. (g) A failed extraction with a user key 12.

(continued)

(e) (f) Figure 6.9 An experimental result. (a) A browser in the cellular phone. (b) A public

(continued)

(g) Figure 6.9 An experimental result. (a) A browser in the cellular phone. (b) A public

(continued)

In the first experiment, while a user utilizes a cellular phone to browse a web page, a website address must be inputted in a browser. Then, a cover image and some descriptions are displayed on the screen. The user can press the button on the cellular phone and download a JAVA program presented in Figure 6.9(c). After the installation is completed, an icon is shown on the main screen. If the user executes the downloaded program, the cover image can be displayed on the screen. While the selection button presented in Figure 6.9(e) is pressed, the user can input a user key to extract secret messages hidden in the cover image. If the provided key is wrong, the extraction process fails, as presented in Figure 6.9(g).

(a) (b)

Taking a picture. (c) A captured image. (d) Keying in secret messages.

(e) The completed embedding process. (f) A cover image in the database. (g) The transmission system on the sender site. (h) The receiving system on the receiver site. (i) Loading of the received cover image. (j) The completed extraction process. (k) A success extraction with a user key 123. (l) A failed extraction with a user key 12.

(continued)

(e) (f)

(g) (h) Figure 6.10 An experimental result. (a) Three icons of the proposed system. (b)

Taking a picture. (c) A captured image. (d) Keying in secret messages.

(continued)

(i) (j)

(k) (l) Figure 6.10 An experimental result. (a) Three icons of the proposed system. (b)

Taking a picture. (c) A captured image. (d) Keying in secret messages.

(continued)

In the second experiment, when a user wants to deliver secret messages to others making use of the proposed system, a picture as a cover image can be taken by the camera built in the cellular phone, as presented in Figure 6.10(b). The user has to key in a user key and the secret messages, and the proposed system can embed them in the cover image. After the embedding process is completed, the user can utilize the wireless network transmission system presented in Figure 6.10(e) to transmit the cover image to others. On the receiver site, a corresponding receiving system should be executed to get the cover image. After the receiving process is completed, the receiver also has to key in an authentic user key to the proposed system to extract the secret messages, as presented in Figure 6.10(i).

6.5 Summary and Discussions

In this chapter, two types of data hiding applications, including active covert communication and passive covert communication, have been proposed and tested. A user can utilize the proposed system and a cellular phone to achieve the purpose of covert communication for data hiding applications.

There are some advantages of using cellular phones as the executing platform against personal computers, as illustrated in the following.

1. The size of a cellular phone is much smaller than a personal computer.

2. The acquisition of a cellular phone is much easier than a personal computer.

3. The ability of wireless connections of a cellular phone is much more convenient than personal computers.

But there are also some benefits of using personal computers as the executing platform against cellular phones illustrated as follows.

1. The computing power of a personal computer is much stronger than that of a

cellular phone.

2. The physical memory of a personal computer is much larger than a cellular phone.

3. The network transmission rate of a personal computer is much faster than a cellular phone.

In fact, the selection of the executing platform depends on applications which we want.

Chapter 7 Image Transmission with

Authentication Capability on Cellular Phones

7.1 Introduction

Since the transmission of cover images on the platform of a cellular phone is exposed on the public wireless network environment, illicit users may intercept these images and edit them for deceiving receivers or misrepresentation. Thus, verifying the validity and the integrity of the transmitted images is necessary.

In Section 7.2, the proposed authentication method for captured images on cellular phones is introduced. This method describes how to generate authentication signals for images and two processes for embedding and extracting of them are also included here. Finally, some experimental results and discussions are given.

7.2 Proposed Authentication Method for Captured Images on Cellular Phones

Because most of modern cellular phones have built-in cameras, they have the

ability of taking pictures. Combined with the wireless network transmission system, cellular phones can help users investigate confidential cases.

Here is an application example. When an employee is investigating classified cases and has to take important pictures for evidences, the employee can utilize a camera built on a cellular phone to accomplish this job. After collecting essential pictures, the employee can transmit them to superiors through the public wireless network environment. After the superiors receive these pictures, they do not know whether the received pictures are genuine or not and can use the proposed system to authenticate them.

In Section 7.2.1, an authentication signal generation method is proposed. After generating authentication signals, two processes for embedding and extracting them are needed, and they are described in Sections 7.2.2 and 7.2.3, respectively.

7.2.1 Method for Generating Authentication Signals

In order to verify the fidelity of captured images, a 4×4 block in an image is taken as an authentication unit. While a suspicious image is being authenticated, the proposed system can check an authentication signal hidden in each 4×4 block. An illustration of a 4×4 block is shown in Figure 7.1 and the corresponding detailed algorithm is described in the following.

Algorithm 1: Generate an authentication signal for a 4×4 block.

Input: a 4×4 block B, a user key K, and each pixel value of a 4×4 block Pi, i = 1, 2, …, 16.

Output: an authentication signal S.

Steps:

2. Calculate the average pixel value of P_i’ and denote it by M.

3. Utilize K and M to generate a random integer R with the length of 32 bits as an authentication signal S.

… … … …

… … …

Captured image Let a 4x4 block be an authentication unit

Figure 7. 1 An illustration of a 4×4 block on a captured image.

7.2.2 Process for Embedding Authentication Signals

After capturing images using a camera built on a cellular phone, a user should take these images and a user key as input to the proposed system. The system uses the method for generating authentication signals mentioned in Section 7.2.1 and a 2-LSB data hiding technique to hide an authentication signal into a 4×4 block. For each block, it has 16 pixels and the system can employ the two least significant bits of each pixel to embed a random integer with the length of 32 bits in the red, green, and blue channels, respectively. A flowchart of the process is shown in Figure 7.2.

… … … …

… … …

Captured Images

Key

Captured Images with Authentication Signals Generate

Authentication Signal

Take Pictures

Embed Authentication Signal

… … …

Captured image Let a 4x4 block be an

authentication unit

Figure 7. 2 An illustration of the process for embedding authentication signals in a captured image.

7.2.3 Process for Extracting Authentication Signals

While the receiver gets a suspicious image, the system can utilize an input user key and the corresponding mean value of each block to regenerate an authentication

signal. If an extracted authentication signal and a regenerated one is the same, the current block can be considered as authentic, otherwise, it is thought unauthentic. The extraction process is an inverse of the embedding one and both of them utilize the 2-LSB data hiding technique. A flowchart of the process is shown in Figure 7.3

… … … … Let a 4x4 block be an authentication unit

Figure 7. 3 An illustration of the process for extracting authentication signals from a

captured image.

An illustration of the proposed method is shown in Figure 7.4. Both the sender and the receiver should take the same user key as input for correct verification of suspicious captured images. At the sender site, the receiver’s cellular phone number should be input to the proposed system and all data are transmitted to the receiver site through the public wireless network.

Captured

Figure 7. 4 An illustration of the proposed authentication method for captured images

在文檔中主動式資訊隱藏與應用之研究 (頁 81-0)