Image analysis – Applications
Reexamination Certificate
1999-01-22
2002-09-10
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
Reexamination Certificate
active
06449378
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a data processing apparatus and method of embedding or padding electronic water-mark information in moving image data and a storage medium storing the method.
2. Related Background Art
As techniques of embedding electronic water-mark information in moving image data, a method of embedding electronic water-mark information in movement vectors between frames, a method of embedding electronic water-mark information by combining two types of moving images obtained by two cameras for photographing the same object at slightly different angles, and the like are known.
The principle of moving image compression in a moving image compression scheme including an intraframe encoding means and an interframe encoding means, e.g., MPEG moving image compression, will be described first. A method of embedding electronic water-mark information in movement vectors between a plurality of frames of moving image data will be described next.
In MPEG moving image compression, the redundancy in the time axis direction is reduced by calculating the difference between a plurality of frames, and the obtained difference data is subjected to the DCT (Discrete Cosine Transform) and variable-length encoding processing to reduce the redundancy in the spatial direction, thereby realizing efficient encoding as a whole. Consider the redundancy in the time axis direction. Since the correlation between consecutive frames of a moving image is high, the redundancy can be reduced by calculating the difference between a target frame to be encoded and a frame temporally preceding or succeeding the target frame.
According to this compression scheme, one of the following pictures is assigned to each frame: an intra-coded picture (I-picture) obtained by the intra-encoding mode of encoding a frame based on only the image data in the frame to be encoded; a forward predictive-coded picture (P-picture) obtained by encoding the difference value between a target frame and a frame one frame ahead of the target frame; and a bidirectional predictive-coded picture (B-picture) obtained by the difference value between a target frame and a frame temporally preceding or succeeding the target frame or the difference value between interpolated frames generated from these two frames. Each frame is encoded in accordance with this assignment result, and three types of encoded data obtained in this manner are combined with each other in a predetermined sequence. The resultant data is then output.
In MPEG, it is recommended that 1 I-picture, 4 P-picture, and 10 B-picture constitute one unit (GOP) with the I-picture being assigned at the beginning of the 15 frames, and 2 B-pictures and 1 P-picture being repeatedly assigned afterward. By assigning I-pictures to a plurality of frames in a predetermined cycle, special reproduction such as reverse reproduction or partial reproduction in units of GOPs can be performed, and error propagation can be prevented.
Assume that a given target frame to be encoded is a new image that is not related to a temporally preceding frame. In this case, the difference value between the target frame and a succeeding frame may become smaller than that between the target frame and the preceding frame. For this reason, in MPEG, bidirectional predictive encoding is performed to realize more efficient compression.
In MPEG, so-called movement compensation is performed. In this operation, the differences between macroblocks at
ear the same positions in a target frame and a temporally preceding or succeeding frame in units of a predetermined number of blocks (macroblocks) consisting of 4 luminance data blocks and 2 color difference data blocks, each block serving as a processing unit in the above DCT and consisting of 8 pixels×8 pixels. By discriminating a macroblock exhibiting the minimum difference, a movement vector is detected, and the detected movement vector is output as part of encoded data.
A given macroblock is decoded as follows. Macroblock data at the same position in a temporally preceding or succeeding frame is extracted by using this movement vector, and the macroblock encoded in the above manner is decoded on the basis of this extracted data. In the above movement compensation, after a frame temporally preceding the target frame is temporarily encoded on the encoding side, a preceding frame that can be obtained by decoding the encoded frame is generated in advance on the decoding side. Then, movement compensation is performed by using macroblocks in this preceding frame and the target frame. With this operation, the target frame can be properly encoded by referring to the actual reproduced image on the decoding side.
According to the scheme of embedding electronic water-mark information in movement vectors, 1 bit of water-mark information can be embedded by changing one movement vector. When, for example, the value of 1 bit is to be set to “1”, this movement vector is moved to a visually unrecognizable degree. When the above value is to be set to “0”, the movement vector is not moved. This changing/unchanging processing is performed for movement vectors equal in number to the bits of water-mark information, thereby embedding one piece of water-mark information.
When electronic water-mark information is to be embedded in image data by using this method, movement vectors must always exist. It is therefore difficult to embed electronic water-mark information in a frame having an image without any movement vector, i.e., a frame having an image without any movement of the object.
The method of embedding water-mark information by combining two types of moving image data obtained by two cameras for photographing the same object at slightly different angles will be described next.
According to this method, two cameras are installed to photograph one object at slightly different angles. Since the angle defined by the two cameras with the object serving as a vertex is very small, the difference between the images obtained by the two cameras cannot be identified by the human eye.
In the following description, the two cameras will be referred to as cameras A and B. Each of the moving images obtained by the cameras A and B is divided into frames. The respective frames are represented by (a1, a2, . . . , an) and (b1, b2, . . . , bn). Frames are randomly selected from the respective moving images obtained by the cameras A and B and are combined with each other to generate an original moving image. With this operation, for example, the original moving image can be generated in the form of (a1, a2, b3, a4, b5, . . . , bn).
Subsequently, several bits constituting water-mark information are sequentially embedded one by one in each frame. More specifically, when 1 bit to be embedded in 1 frame of the original moving image, i.e., “a1”, is “1”, the frame “a1” of the original moving image is replaced with another frame “b1” of the moving image. When the value of this 1 bit is “0”, the frame is not be replaced. By performing this processing for frames equal in number to the bits of the water-mark information, one piece of water-mark information can be embedded in the moving image data.
When the water-mark information embedded by this processing is to be extracted and decoded, the respective images (a1, a2, . . . , an) and (b1, b2, . . . , bn) obtained by the two cameras A and B are required in addition to the original moving image. Therefore, a large storage area is required. In addition, since moving images obtained by the two cameras for photographing the same object at slightly different angles are required, this method cannot be applied to an existing moving image.
As described above, according to the method of embedding electronic water-mark information in movement vectors, no electronic water-mark information can be embedded in a moving image constituted by a plurality of frames whose contents hardly change. The method of embedding electronic water-mark information by combining two types of moving images obtained by photographing the s
Iwamura Keiichi
Nagasawa Kenichi
Yoshida Jun
Canon Kabushiki Kaisha
Fitzpatrick ,Cella, Harper & Scinto
Johns Andrew W.
Nakhjavan Shervin
LandOfFree
Data processing apparatus and method and storage medium does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data processing apparatus and method and storage medium, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data processing apparatus and method and storage medium will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2837138