WO2001043422A1

WO2001043422A1 - Information processing method and recorded medium

Info

Publication number: WO2001043422A1
Application number: PCT/JP1999/006838
Authority: WO
Inventors: Masahiko Katsuragi; Yasuhiro Akiyama; Kenta Morishima
Original assignee: Hitachi,Ltd
Priority date: 1999-12-07
Filing date: 1999-12-07
Publication date: 2001-06-14

Abstract

Digital audio ID information (3) having a meaning of ID information is embedded as invisible electronic watermark information in a sampled value region of digital original image information (1) or in a frequency region where the digital original information is subjected to frequency conversion, and the embedded digital audio ID information is made separable from the sampled value region or the frequency region by an information processing device (41), thereby realizing a function of inhibiting illegal copy of digital image information. Since digital audio ID information is adopted as electronic watermark information having a meaning of ID information, the information is recognized as a meaning even if, e.g., 50% of the information is lost. Further the watermark strength against disturbance is high. The strength against disturbance is desirably and simply selected according to a request by readily changing the audio parameters, such as the number of quantized bits and sampling frequency. ID information can be relatively easily recognized by using an electronic watermark even by a low-cost system.

Description

Description Information processing method and recording medium

The present invention relates to a technology for embedding a digital watermark in original image information for clarifying a copyright or the like of original image information created by a digital information content provider or the like, and to information processing for embedding a digital watermark. The present invention relates to a method, an information processing device, and a technology effective when applied to an information processing system for electronic approval or electronic transaction using information in which an electronic watermark is embedded. Background art

As an example of a document describing the digital watermarking technology, there are pages 199 to 124 of Nikkei Electronics 1997.2.24 (no. 683) published by Nikkei BP Company. Digital watermarking technology is a technology that embeds and hides some information in multimedia data such as images and sounds. If the invisible digital watermark information is embedded and continues to remain in some places such as an electronic work, the effect of suppressing unauthorized copying of the work can be expected.

Conventionally, when embedding a digital watermark in images, audio, text, etc., the ID (Identification) number encoded for each individual information copyright holder and each company that has the distribution right of the original information is encrypted. ID information that is one-to-one corresponding to information such as image data, audio data, and text is encrypted and embedded in the original information. Was. In each case, the data format of the embedded ID information itself The formula was embedded as image or text information. The key number is used to detect the ID number at that time. The key information is provided with information that can identify the location where the ID number is embedded.

As a technology in which the digital watermarking technology is applied from a viewpoint different from the suppression of the illegal copying, there is a video device described in Japanese Patent Application Laid-Open No. H10-290404 (the video device described in It is designed to embed subtitles and audio digital signals in multiple languages into digital video signals using digital watermarking technology, so that multiple languages can be used without changing the recording format. This makes it possible to record textual information and audio information, and at the time of playback, it is possible to select and reproduce any subtitles or audio digital signals extracted from the video signal. Is to efficiently embed subtitles or audio information related to video information using digital watermarking technology. And the audio information must be digital watermarks that can clearly restore and maintain the quality of the reproduced information, such as the voices and sounds of each performer can be distinguished from each other. This is completely different from digital watermarking technology as an ID for protection.

Further, Japanese Patent Application Laid-Open No. 11-166686 discloses an image information processing technique using information such as images, sounds, and text as embedded information. The embedded information is information to be superimposed on the original information as visible information, and is a technique for embedding information accompanying the original information. Therefore, the embedded information is one type of multimedia information that is added to the original information, and the embedded information is information that should be actively exposed to users. It is completely different from information embedding technology. For example, the technology When applied to children's picture books, when a landscape image of a zoo is used as the embedded image information, for example, a lion's cry is embedded in a lion's picture, and information about the ecology of a penguin is embedded in a penguin's picture. By associating the embedded image information with the embedded image information relatively easily associated with each other, it becomes easy to search for the embedded information desired by the user from a plurality of embedded information.

According to the study of the digital watermark technique used as the ID by the present inventors, the following points have been further clarified. Conventionally, digital watermark information used as an ID is an image, text, or information obtained by encrypting them. Here, the image information in which the electronic watermark information having the meaning as the ID information is embedded is, for example, noise addition during transmission during transmission and reception of information in a network business, etc. Various disturbances occur in embedded watermark information such as information loss during transmission, encryption processing, code compression processing, decryption processing, and code decompression processing. Therefore, digital watermark information must be resistant to such disturbances. This is because the purpose is to protect copyright, and when watermark information is extracted, it is necessary that the meaning of the ID information can be finally confirmed accurately. However, there is a natural limit to increasing the strength of digital watermark information. Because the electronic watermark information is added to the original information, it is desirable that embedding does not affect the original information as much as possible. In order to increase the strength of digital watermark information and minimize the influence on the original information, the more complicated and heavy the watermark processing becomes, the higher the price of the information processing system becomes. Etc. becomes difficult.

An object of the present invention is to reduce the amount of embedding information relative to original image information as compared with the conventional invisible electronic embedding technology, An information processing method that realizes an electronic watermark as ID information that is resistant to disturbance and information loss, a recording medium and a transmission medium for a program that contributes to the realization of the information processing method, and an information processing method It is to provide an apparatus and an information processing system.

Another object of the present invention is to provide an information processing method capable of relatively easily recognizing ID information by a digital watermark even in a low-cost system, a recording medium of a program contributing to the realization of the information processing method, and transmission. It is to provide a medium, and an information processing device and an information processing system.

Still another object of the present invention is to provide an information processing apparatus that has a small amount of embedded information relative to original image information, and is strong against disturbance to the embedded original information and against information loss. An object of the present invention is to provide a recording medium storing a program capable of easily realizing embedding of ID information by a digital watermark and reproduction of the embedded ID information.

The above and other objects and novel features of the present invention will become apparent from the following description of the present specification and the accompanying drawings. Disclosure of the invention

[1] The information processing method from the viewpoint of embedding digital watermark information uses the ID information as the ID information in the digitized original image information, that is, in the sample value area (sample value area) of the digitized original image information. It is characterized by embedding separable digital voice ID information as invisible digital watermark information. An information processing method according to a more detailed aspect of the above information processing method includes an embedding method for determining a plurality of embedding positions for embedding digital watermark information in digital original image information by referring to the original digital image information. A position determination process, an order determination process for determining an embedding order for the determined embedding position, and an ID information is stored in the embedding position in accordance with the embedding order. And embedding digital voice ID information having meaning as information as invisible digital watermark information.

According to the above-described means, the digital audio ID information as invisible digital watermark information generally has a larger number of quantization bits, a smaller sampling period, and a larger number of audio data, the larger the audio ID watermark. Strength can be increased. As a feature of the voice ID, it is only necessary that the meaning of the voice when the watermark is finally extracted can be recognized (meaning understanding) by a person (checker) or by a machine, and the voice quality (sound quality, noise level, gender identification) , Individuality identification, etc.) are good enough. During playback, even if there is a lack of information of several 10% in the digital audio ID information, it can be recognized as meaning. Therefore, the amount of embedded information relative to the original image information is small, and it is possible to realize a digital watermark as ID information that is strong against disturbance to the embedded original image information and also against information loss. . In addition, by easily changing the audio parameters such as the number of quantization bits and the sampling frequency, it is desirable to obtain the strength against disturbance according to the information amount of the original image information or as necessary. It is easy to choose. These make it possible to realize ID information recognition by digital watermarking relatively easily even in a low-cost system.

The information processing method which focuses on the frequency conversion value region (frequency domain) of the digitized original image information is performed by a frequency conversion process of the digitized original image information and a predetermined frequency obtained by the conversion process. An embedding process of embedding digit voice ID information having meaning as ID information in the component as invisible digital watermark information, and an inverse conversion process of performing an inverse conversion for the frequency conversion on the information after the embedding process. , Is included. An information processing method according to a more detailed aspect of this information processing method includes a conversion process of performing frequency conversion on digital original image information and a conversion process obtained by the conversion process. An embedding position determining process for determining, as an embedding position, a plurality of embedding frequency components for embedding digital watermark information from each frequency component, and an order determining process for determining an embedding order for the determined plurality of embedding positions. A process of embedding digital audio ID information having meaning as ID information at the embedding position as invisible digital watermark information in accordance with the embedding order; and performing an inverse process for the frequency conversion on the information having passed the embedding process. And inverse conversion processing for performing the conversion.

Embedding audio ID information in the frequency domain slightly increases the amount of information processing compared to embedding in the sample value domain, but can easily increase the watermark strength.

A parameter calculation process for calculating a digital voice configuration parameter for defining the number of quantization bits, a sampling frequency, and a voice time of the digital voice ID information based on the digital original image information. May be included. Even during playback, it is possible to directly obtain the required daisy-voice audio configuration parameters from the original image information. Instead of this, it is also possible to prepare in advance a plurality of kinds of parameters for the digital voice configuration in a table or the like, and to select the parameters in accordance with the required watermarking strength.

The digital voice ID information includes, for example, one of the number of quantization bits selected from 3 bits to 16 bits and a voice parameter of a sampling frequency selected from a range of 3 to 40 kHz. Even relatively low-accuracy speech information, such as that specified by an overnight, is sufficient for ID recognition. Similarly, the digitized voice ID information is quantized and formed by a 1-bit oversampling method, and is voice information specified by voice parameters having a sampling frequency of 3 to 40 kHz. There is no inconvenience. Since the digital audio ID information is not limited to embedding along the position pattern of the original image like embedding the image ID, It is easy to embed multiple. In other words, the embedding position is determined depending on the position pattern of the image such that the image ID information is embedded as noise in a so-called redundant portion that is not important for human perception. In the case of digital audio ID information, the amount of information is small and inherently invisible information, so it is easy to embed multiple times. At this time, if the embedding position of the current voice ID information is determined so that the center of gravity differs from the centroid of a plurality of embedding positions that have already been embedded, reproduction of the ID information becomes easy.

[2] The information processing method from the viewpoint of the reproduction of digital watermark information is based on the case where digital audio ID information is embedded in the sample value area of the original digital image information, as ID information in the original digital image information. Separation processing for separating the digital audio ID information from the information embedded with the separable digital audio ID information as invisible digital watermark information, and the separated digital audio ID information And a sound reproduction process for reproducing the sound. The audio playback is performed using the audio configuration parameters for the separated and reconstructed digital audio ID information, and the data order during reproduction follows the embedding order of the digital audio ID information.

When digitized voice ID information is embedded in the frequency domain of the digitized original image information that has been frequency-converted, the information processing method for reproduction is based on the fact that an ID is assigned to a predetermined frequency component of the digitized original image information. Frequency conversion processing for performing frequency conversion on information in which digital audio ID information having meaning as information is embedded as invisible digital watermark information, and the digital audio ID based on information obtained by the frequency conversion processing. It includes a separation process for separating information and a sound reproduction process for reproducing the separated digital audio ID information.

Since it is sufficient to recognize digital voice ID information as digital watermark information, it is relatively difficult to recognize ID information by digital watermark even in a low-cost system. It can be easily realized.

[3] The invention according to another aspect provides a program for implementing the information processing method on an information processing device such as a computer device.

The recording medium of the program is a medium such as a CD-OM for statically recording the program. The transmission medium is a communication medium that dynamically transmits the program for electronically, electromagnetically, or optically distributing or distributing the program via a network connected by a wired line or a wireless line ₍ When embedding the ID information in the sample value area, the recording medium of the program uses the process of embedding the digital audio ID information, which has meaning as ID information, in the original digital image information in a separable manner as invisible digital watermark information. And a process for separating the digitized voice ID information from the embedded information described above, and a program for causing the information processing device to execute the program is recorded in the information processing device in a readable manner.

In consideration of embedding the ID information in the frequency domain, the recording medium includes a first conversion process for frequency-converting the original digital image information and a predetermined frequency component obtained in the first conversion process as ID information. Embedding processing for embedding meaningful digital audio ID information as invisible digital watermark information; inverting processing for performing inverse conversion on the frequency conversion of the information that has undergone the embedding processing; A program for causing an information processing device to execute a second conversion process for converting to a frequency component and a separation process for separating the digital audio ID information from the information obtained in the second conversion process is executed by the information processing device. It is recorded so that it can be read.

When embedding the ID information in the sample value area, the transmission medium of the program embeds the digitized audio ID information having meaning as the ID information in the original digitized image information so as to be separable as invisible digital watermark information. Features and And a function of separating the digitized voice ID information from the embedded information in an information processing apparatus.

To consider embedding the ID information in the frequency domain, the transmission medium, the first conversion processing function for frequency conversion of the original image information, and the predetermined frequency component obtained by the first conversion processing function have the ID An embedding processing function for embedding digital audio ID information having meaning as information as invisible digital watermark information, and an inverse conversion processing for performing an inverse conversion of the frequency conversion on the information obtained by the embedding processing function. A second conversion processing function for converting the inversely converted information into frequency components, and a separation processing function for separating the digitized voice ID information from the information obtained by the second conversion processing function. A program for realizing the processing device is transmitted.

The program may include a description of a processing procedure for controlling or realizing the processing of reproducing the separated digital audio ID information. According to the recording medium and the transmission medium of the program, the amount of embedded information relative to the original image information is small, and it is strong against disturbance to the embedded original information and also against information loss. The information processing device can easily realize the embedding of the ID information by the electronic watermark and the separation of the embedded ID information.

[4] A recording medium according to another aspect of the present invention records information in which digital audio ID information is embedded as digital watermark information in original image information. In other words, a recording medium according to this aspect is a recording medium in which image information is recorded in an information processing apparatus in a readable manner, and the image information is a digital audio ID information having a meaning as an ID, and invisible digital watermark information. The digital sound embedded and embedded in the sample value area of the original digital image information or in the frequency-converted frequency area of the original digital image information as Voice ID information is information that can be separated from the sample value area or the frequency domain by the information processing device. More specifically, the image information is information in which digital audio ID information having meaning as ID information is separably embedded as invisible digital watermark information in original digital image information. Alternatively, in the image information, digital audio ID information having a meaning as ID information is embedded as invisible electronic watermark information in a predetermined frequency component of the original digital image information, and the digital file is transmitted through frequency conversion. The voice ID information is separable information.

Further, a transmission medium according to another aspect of the present invention transmits the image information in which the digital audio ID information is separably embedded in the original digital image information to an information processing apparatus.

An information processing device such as a convenience store device for inputting the image information recorded on the recording medium or the image information transmitted via the transmission medium may be provided with voice ID information as digital watermark information as necessary. This makes it possible to relatively easily realize a process of separating the contents and confirming the contents of the information processing device even with a simple cost.

[5] The present invention from the viewpoint of an information processing system that processes or uses information in which audio ID information is embedded in original image information takes into account an information processing system for electronic approval, electronic commerce, information distribution, and the like.

In particular, when electronic approval is considered, the information processing system may use digital voice ID information having meaning as ID in the sample value area of the original digital image information or as the invisible electronic watermark information. A first information processing apparatus for outputting image information embedded in a frequency domain subjected to frequency conversion so as to be separable, and a second information processing apparatus for inputting the image information and using the input image information as electronic approval information And an information processing device. Considering the electronic signature used in electronic approval and electronic commerce, the information processing system converts the digitized voice ID information, which has meaning as ID, into the sample value area of the digitized original image information as invisible electronic watermark information. Alternatively, a first information processing apparatus that outputs image information embedded in a frequency domain obtained by frequency-converting the digital original image information in a separable manner, and the image information that is input after inputting the image information, And a second information processing device that uses the information as electronic signature information.

When information distribution is particularly taken into account, the information processing system uses the digital audio ID information having a meaning as an ID as the invisible digital watermark information in the sample value area of the original digital image information or the digital original image information. It includes an information processing device that stores image information embedded in a frequency domain whose frequency has been converted so as to be separable, and outputs the stored image information in response to a distribution request.

According to the information processing system described above, information obtained by embedding the voice ID information in the original image information is processed or used in electronic approval, electronic commerce, information distribution, and the like. You can expect a function to prevent unauthorized copying.

[6] From the viewpoint of the information processing apparatus used in the information processing method and the information processing system, the information processing apparatus includes an input unit and an arithmetic control unit, and the input unit includes a digital source image. Information is input, and the arithmetic control means can embed separable digital audio ID information having meaning as ID information into the input digital original image information as invisible digital watermark information.

Also, an information processing apparatus according to another aspect includes a storage unit and an operation control unit, wherein the storage unit stores digital original image information, and the operation control unit includes a digital storage unit stored in the storage unit. Frequency conversion to original image information And embeds digital audio ID information having meaning as ID information as invisible digital watermark information in a predetermined frequency component obtained by the frequency conversion, and applies the information after the embedding to the frequency conversion. It is possible to perform an inverse transformation.

An information processing apparatus according to a more detailed aspect includes a storage unit and an operation control unit, wherein the storage unit stores digital original image information, and the operation control unit includes digital original image information stored in the storage unit. , Determining a plurality of embedding positions for embedding digital watermark information in the digital original image information, determining an embedding order for the determined embedding position, and as the ID information at the embedding position according to the embedding order. It is possible to embed the digital voice ID information with the meaning of as invisible digital watermark information.

An information processing apparatus according to another aspect of the information processing apparatus according to another aspect includes a storage unit and an operation control unit, wherein the storage unit stores digital original image information, and the operation control unit includes the storage unit. The original digital image information stored in the means is frequency-converted, and a plurality of embedding frequency components for embedding digital watermark information from among the frequency components obtained by the frequency conversion are determined as embedding positions. The embedding order for a plurality of embedding positions is determined, and digitized voice ID information having meaning as ID information is embedded as invisible electronic watermark information in the embedding position according to the embedding order, and the embedding process is performed. It is possible to perform an inverse transformation to the frequency transformation on the information that has passed through.

At this time, from the viewpoint of embedding the digital watermark information in a multiplex manner, the operation control means determines the embedding position when the digital audio ID information is already embedded as an invisible digital watermark. Has a center of gravity that is different from the center of gravity of multiple embedding positions that have already been embedded. Thus, the embedding position of the current voice ID information may be determined.

The audio parameters of the audio ID information may be calculated from the state of the original image information or selected from a table. The information processing apparatus adopting the latter includes an input unit, a storage unit, and an arithmetic control unit, and the storage unit converts the frequency of the sample value area of the original image information or the original image information. The embedding control information for embedding the invisible digital watermark information in the frequency domain and the order of the positions are predetermined, and the voice configuration parameters of the digital audio ID information having meaning as the digital watermark information are set in advance. The determined voice parameter overnight information is stored. The input means inputs digital original image information. The arithmetic control means selects required voice parameters and embedding control information from the storage means, and selects the selected audio parameters in a sample value area of digital original image information or a frequency-converted frequency area of digital original image information. And embedding the digitized voice ID information according to the selected voice parameter in the position and order specified by the selected embedding control information.

Further, encryption may be adopted for embedding the digital watermark information. That is, the information processing device includes an input unit, a storage unit, and an arithmetic control unit, and the storage unit stores a sample value region of the digitized original image information or a frequency-converted frequency region of the digitized original image information. The embedding control information for embedding the invisible digital watermark information in the video and the order of the positions are encrypted in advance and the embedding control information, and the voice configuration parameters of the digital voice ID information meaningful as the digital watermark information are specified in advance. Stores the encrypted audio parameter information. The input means inputs digital original image information. The arithmetic control means selects required encrypted voice parameters and encrypted embedded control information from the storage means, decodes the selected encrypted voice parameters and encrypted embedded control information, and decodes the digitized original image information. Sample value area or data Embedding digital audio ID information in accordance with the decoded audio parameter information in the frequency domain of the digital original image information in which the frequency has been converted, in the position and order specified by the decoded embedding control information It is.

An information processing apparatus from the viewpoint of reproducing ID information includes input means and arithmetic control means, and the input means includes a digital audio ID having a meaning as ID information in the original digital image information. Information in which information is separably embedded as invisible digital watermark information is input, and the arithmetic control means can separate the digital audio ID information from the input information.

An information processing apparatus according to a more detailed aspect from the viewpoint of reproduction of ID information includes input means and arithmetic control means, and the input means has a predetermined frequency component of digital original image information having a meaning as ID information. Enter the information in which the digital voice ID information is embedded as invisible digital watermark information. The arithmetic control unit can perform frequency conversion on the input information, and can separate the digitized audio ID information from the information obtained by the frequency conversion.

The arithmetic control means may further perform a noise removal filtering process on the separated digitized audio ID information. The voice ringing process makes it easier to recognize the voice ID. The arithmetic control unit may further perform a speech recognition process on the digital audio data D that has been subjected to the noise removal filtering process. Automatic recognition of voice ID becomes possible. BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 is an explanatory diagram showing an example of a method of embedding digital audio ID information as a digital watermark in a sample value area of digital original image information. FIG. 2 is an explanatory diagram showing an example of a method for embedding digital audio ID information as a digital watermark in the frequency domain of digital original image information. FIG. 3 is an explanatory view exemplifying a rule that specifies the order of embedding digital voice ID information.

FIG. 4 is a flowchart showing in detail a processing procedure for embedding the audio ID information by the method of FIG.

FIG. 5 is an explanatory diagram illustrating a method of separating the digital audio ID information and the original image information from the image information in which the digital audio ID information is embedded by the method of FIG.

FIG. 6 is an explanatory diagram exemplifying a method of separating the digital audio ID information and the original image information from the image information in which the digital audio ID information is embedded by the method of FIG.

FIG. 7 is a flowchart showing in detail the processing procedure for separating audio ID information and original image information by the method of FIG.

FIG. 8 is a block diagram showing a basic configuration example of an information processing system assuming a case where a digital watermark using audio ID information is applied to electronic approval and electronic commerce.

FIG. 9 is an explanatory diagram showing in detail the flow of processing by the information processing apparatus in FIG.

FIG. 10 is an explanatory diagram showing emphasizing an embedding place when a digital watermark based on audio ID information is multiplexed into image information.

FIG. 11 is an explanatory diagram showing an example of embedding an electronic watermark by voice ID information in an electronic approval document and electronic identification documents.

FIG. 12 is a block diagram illustrating an outline of an electronic commerce or electronic approval system using voice ID information.

FIG. 13 is a block diagram showing an example of an information processing device that uses voice ID information for digital watermarking, and an example of an information processing network including the information processing device. BEST MODE FOR CARRYING OUT THE INVENTION

FIG. 1 shows an example of a method of embedding a digital audio ID as a digital watermark in a sample value area of digital original image information. Fig. 2 shows an example of a method for embedding digital audio ID information as a digital watermark in the frequency domain of digital original image information.

When the digital audio ID information 3 is embedded in the original image information as a digital watermark, the digital audio ID information 3 is embedded in the sample value area, which is the raw original image information 1 itself, as shown in Fig. 1, and as shown in Fig. 2. As described above, the raw original image information 1 may be embedded in a frequency domain obtained by performing a frequency transformation (orthogonal transformation) process 6 such as FFT (Fast Fourier Transform). In FIG. 1 and FIG. 2, a plurality of types of digital voice ID information 3 having different voice parameters are prepared in advance, and are selected by the selection means 104, and are used for the watermark embedding processes 2 and 7. Provided. The digital audio ID information 3 is digital audio information generated by A / D conversion of analog audio for indicating copyright ownership or the like, or digital audio information generated by digital audio synthesis. Information.

In the case of the method shown in FIG. 2, the information that has undergone the embedding process is subjected to an inverse conversion process 8 with respect to the frequency conversion process. In FIGS. 1 and 2, what is indicated by 5 and 9 is information formed by embedding digital audio ID information 3 in the original image information as a digital watermark.

In the embedding processes 2 and 7, the optimal embedding position of the voice ID information can be determined from the original digital image information according to a conventional general digital watermark embedding position determining algorithm. At that time, the original information is kept as small as possible. For example, in the case of the watermark embedding process 2, the masking effect that low-luminance pixels near high-luminance pixels are difficult to see is used, and digital audio Embed ID information. In the case of the watermark embedding process 7, for example, the digital audio ID information is embedded in a portion having many high-frequency components which is difficult to understand even when the watermark information is embedded.

Here, as the parameters of the digitized voice configuration of voice ID information 3, the number of quantization bits of voice a (i), sampling period (frequency) Δΐ, voice data — number of evenings (time) η, etc. Can be mentioned. These parameters need to be determined according to the watermark embedding strength required for audio ID information. In general, the greater the number of quantization bits, the smaller the sampling period, and the greater the number of audio data, the stronger the watermark strength of audio ID information can be. In other words, the ability to maintain the original form of information against disturbances such as noise and information loss is high.

As a feature of the digital voice ID information, when the digital voice ID information as an embedded digital watermark is extracted, the meaning of the voice is recognized by a person (checker) or by a machine (semantic understanding). You can do it. Even if the sound quality such as sound quality, noise level, gender identification, and individuality identification is poor, it is sufficient if the meaning can be understood. For example, as an example of a digital audio configuration parameter, if the number of quantization bits is 4 bits, the sampling frequency is 4 kHz, and the digital audio time is 1 second to 2 seconds, the digital audio It is sufficient that the watermark data amount of the ID information is 500 to 100 points at the sample points. In the examples of Fig. 1 and Fig. 2, several voice parameters were varied according to the required embedding strength and the content of the embedding information (whether or not the content was easy to understand). The digital voice ID information 3 is prepared in advance, and is selected by the selection means 4 as appropriate. That is, the embedding voice parameters are determined in the watermark embedding processes 2 and 7, and the digital voice ID information is selected by the selection means 4 in accordance with the parameters. All voice parameters of selectable digital voice ID information It need not be selectable. It is only necessary to prepare several types of digital audio ID information in which the number n of audio data and the sampling period At are fixed.

In the embedding processes 2 and 7, the embedding order of the digitized audio ID information, that is, the information reproduction order, is determined in addition to the information embedding position. As the embedding order, as shown in FIG. 3, a horizontal zigzag (A), a vertical zigzag (B), a horizontal zigzag (C), an irregular random number (D) and the like can be selected. For example, when (C) in FIG. 3 is adopted as the data order, the digital original image information is scanned sequentially in the direction of the arrow in FIG. To be more specific, the data of the data sequence that constitutes the digital voice ID information (data block units) is embedded.

After the watermark embedding processes 2 and 7, in the case of FIG. 1, the information itself obtained from the embedding process 2 becomes new information 5 in which the voice ID information is embedded, and in the case of FIG. 2, the frequency conversion process 8 The information after this is the new information 9 in which the voice ID information is embedded. In the examples of FIGS. 1 and 2, the new information 5 and 9 also include the information of the digital watermark embedding position as key data. Although not specifically shown, the parameters of the voice ID information configuration such as the number of quantization bits and the data of the embedding order may be included together with the key data. If the parameters of the voice ID information configuration are calculated according to the state of the original image information, and the embedding positions are also calculated based on the parameters of the voice ID information configuration according to the state of the original image information, the reproduction is performed. It is not necessary to include such information in the new information 5 and 9. This is because it can be newly obtained by calculation again during playback. If such calculation is not possible, the embedding position and parameters of the voice ID information configuration may be included in the above information 5 and 9 as key data, but the external certificate authority or the like stores the key data. And If available, key information need not be included in information 5 or 9 at all. In addition, when the voice configuration parameters, embedding position, and embedding order are fixedly determined according to a certain algorithm, such information need not be included in the key data.

Comparing the information processing methods in Fig. 1 and Fig. 2, in Fig. 1, there are no frequency conversion processing 6 and its inverse conversion processing 8, so the number of operation processing steps is small and the processing is light. It is convenient for real-time processing such as video. The method shown in Fig. 2 is more complicated than the method shown in Fig. 1 in terms of frequency conversion and inverse conversion, and it is more complicated than the method in Fig. 1 in terms of preventing falsification of watermark information. The frequency conversion is not limited to FFT, but may be DCT (Discrete cosine transform) or the like.

FIG. 4 shows a detailed processing procedure for embedding the audio ID information by the method shown in FIG. The processing shown in the figure can be executed by, for example, an input device such as a convenience store main unit, a display, a keyboard, and an information processing device such as a display device having an interface circuit capable of transmitting information to the outside. Processing.

When audio ID information embedding processing is instructed, digital audio ID information for a watermark to be embedded in the original image information is set (11). The setting of the voice ID information performed here is to create the digital voice ID information primitive. For example, audio information that indicates copyright ownership is generated as analog audio information, and this is converted into digital audio ID information using multiple types of audio parameters. Next, the voice ID information parameters are determined with reference to the original image information, and the corresponding digital voice ID information for embedding is determined (12). For example, depending on the data density and color tone of the original image, how much error can be embedded and how much watermark strength is required In consideration of the above, etc., the required voice composition parameters are determined. Further, the position and the order of embedding the digital audio ID information as the watermark information are determined from the original image information according to a predetermined algorithm such as using the above-mentioned mask effect (13). After that, the process of embedding the digital audio ID information for the watermark to the original image information to the end while changing the position according to the above order is repeated (14, 15) <The method of Fig. 2 is realized. In this case, although not shown, the original image information is frequency-converted before step 14. After all embedding is completed (after No in step 15), inverse conversion of frequency conversion is performed on the information in which the voice ID information is embedded. Their processing is different from that in Fig. 4 ο

FIG. 5 shows a method of separating the digital audio ID information and the original image information from the image information in which the digital audio ID information is embedded by the method of FIG.

The watermark extraction process that separates the digital audio ID information from the original image information from the image information in which the digital audio ID information is embedded 21 is based on the image information 5 from the digital information 5 according to the embedding position and the embedding order when embedding the image. The position where the audio ID information is extracted is obtained, and the original image information is separated from the digital audio ID information.

The embedding position and embedding order of the digital voice ID information are, for example, encrypted at the time of watermark embedding and stored as key data, for example, in information 5, so if the key data is taken out and used according to certain rules The removal position and the removal order are required. As a result, the separated digital voice ID information is reconstructed in a reproducible state. The reconstructed digital voice ID information can be played back or recognized in accordance with the voice configuration parameters. When the embedding position and the order of embedding are determined using the original image information at the time of embedding, even if the information indicating the embedding position and order is not directly included in the image information 5, the embedding position is determined again from the original image information. And the order may be calculated, and the voice ID information may be reconstructed in accordance with the calculated position and order. (At this time, the number of quantization bits of the sound a (i), the sampling period (frequency) t, the number of sound data (Time) Speech parameters such as n can be obtained from the restoration information after removing the embedding information from the embedding position, according to the same algorithm as that determined at the time of embedding, or a predetermined rule is determined in advance. If so, you just have to follow it.

In Fig. 5, what is indicated by 2 2 is the separated digital audio I

The information indicated by D information and 23 are the restoration information of the separated original image information ₍ Fig. 6 shows the digital audio data from the image information in which the digital audio ID information is embedded by the method of Fig. 2). In this case, a method for separating the ID information from the original image information is shown.In this case, the frequency conversion processing 26 equivalent to the frequency conversion processing 6 such as FFT (Fast Fourier Transform) in the embedding processing is performed, and the digital watermark is embedded. After that, a watermark extraction process 24 is performed using the same algorithm as the above-described information 9. The restoration signal 23 here is obtained by removing the embedded information corresponding to the embedding position. The information is obtained by performing an inverse transform 28 on the frequency transform 26.

FIG. 7 shows in detail a processing procedure for separating the audio ID information and the original image information and further reproducing the same by the method shown in FIG. The processing shown in the figure can be executed by an input device such as a display unit, a display, a keyboard, and an information processing device such as a display device having an interface circuit capable of transmitting information to the outside. Processing.

When voice ID information separation processing is instructed, a predetermined watermark removal is performed. Data indicating the position where the ID voice data for watermarking is embedded is obtained from the key data in accordance with the output processing algorithm (31). Based on the position, the restoration data is temporarily separated from the voice ID data (32). At the time of this separation, the order of the digital voice ID information is not considered. In the process of FIG. 7, the voice parameters are determined from the separated restored data (33). It is assumed that the order of the data of the digital voice ID information is determined in advance at the embedding stage. Using the same order as above, refer to the number of quantization bits in the speech configuration parameter and reconstruct the digital speech ID information from the data temporarily separated in step 32 ( 3 4). When the digital voice ID information is reconstructed and completely separated, noise removal is performed on the separated digital voice ID information (35). In FIG. 7, voice reproduction processing is performed on the separated digital noise ID information from which noise has been removed, and the meaning of the digital voice ID information is extracted or recognized (36). If the voice ID information is multiplexed and embedded, the process of steps 33 to 36 is performed for all voice ID information via step 37.

In the case of realizing the method of FIG. 6, although not particularly shown, the frequency of the information 9 is converted before step 32. In addition, the information corresponding to the original image information separated in step 32 is subjected to inverse conversion of frequency conversion to obtain restoration information. Their processing is different from that of FIG.

Fig. 8 shows a basic configuration example of an information processing system assuming that the above method is applied to businesses such as electronic approval, electronic commerce, and information distribution.

In FIG. 8, reference numerals 40 and 41 denote information processing devices, respectively. The first information processing device 40 separates the digitized audio ID information having meaning as ID information into the frequency domain of the original digitized image information as invisible digital watermark information. It is possible to output image information embedded detachably. The second information processing device 41 inputs the image information output by the first information processing device 40 and uses the input image information as electronic approval information or signature information. . Then, the second information processing apparatus 41 separates the digital audio ID information from the image information and reproduces the digital audio ID information, thereby making it possible to check whether or not the image information is genuine.

The process in which the first information processing device 40 embeds digital audio ID information in image information is the same as the process in FIG. In the general electronic business, information 9 in which digital audio ID information 3 is embedded as a digital watermark is encoded (compressed) standardized as MPEG (Moving Picture Experts Group), JPEG (Joint Photographic Experts Group), etc. ) The processing (42) is performed, and the encoded data is encrypted (44) before being transmitted on the network. The transmission path through which the encrypted data is transmitted is a disturbance that has the effect of removing or destroying the embedded watermark information.

The voice ID information embedded as watermark information needs to be strong enough to resist such disturbances. As described above, the strength of the voice ID information is variable according to the voice parameters such as the number of quantization bits and the sampling frequency, so that the required strength can be easily set. Even if voice quality deteriorates due to lack of data, it is sufficient to have enough information to finally recognize the meaning of voice.In the case of voice data, for example, even if 50% Because it can be understood, it is excellent in terms of the strength against disturbance in terms of the sound information. In addition, if the voice time is also 1 to 3 seconds, it is possible to reproduce 20 to 50 words and phrases, which is enough to prove the ID. As the intensity increases, the amount of embedded information increases, and the degree of undesired deterioration with respect to the original image information increases accordingly. Although there is a trade-off between the two, the watermark strength can be easily increased by changing the audio parameters, and it is easy to improve the audio quality. Will be possible.

The encrypted data transmitted via the transmission path is subjected to encryption / decryption processing (47), and further, the second information processing device 41 performs decryption (decompression) processing for the MPEG code (48). ). The composited information corresponds to the image information 9 in which the digital watermark is embedded. When the image information is used for electronic approval or digital signature, whether or not the image information is genuine is determined at the level of the visible information and further at the level of the invisible digital watermark. For example, when it is necessary to determine at the level of the invisible digital watermark, as described with reference to FIG. 6, the voice ID information is separated through the frequency conversion process 26 and the watermark extraction process 24, and separated. The noise may be removed from the obtained information by filtering (50), voice recognition may be performed (51), and the meaning of the voice of the ID may be recognized. If the audio ID information before embedding is encoded, the information of 22 is decoded to obtain the audio ID information.

For reference, FIG. 9 shows the flow of processing by the information processing apparatus in FIG. 8 in detail.

FIG. 10 highlights the embedding location when audio ID information is multiplexed and embedded as a digital watermark in image information.

If the rights relationship of the original image information 60 is diversified, it is assumed that each right holder embeds its own ID information. Area 61 is a voice ID information area of the copyright holder, 62 is a voice ID information area of the owner of the original image information, and 63 is a voice ID information area of the authorized duplication right holder of the original image information. In each area, unique digital voice ID information is embedded one after another as digital watermark information. 6 1D, 6 2D, 6 3D Means the segment data of the voice ID information.

In the example of FIG. 10, since the previous embedding center is known, the next digital watermark information is embedded in a new embedding center shifted from that. That is, in each of the regions 61, 62, and 63, the position of the center of gravity of each region can be determined from the position data of the embedded audio ID information. When newly embedding digital audio ID information, a new embedding position may be determined so as to deviate from the existing center of gravity position. As for the center of gravity position of each area, a place for storing the history or the latest position may be secured in the information 5 or 9.

Fig. 11 shows an example of embedding a digital watermark with audio ID information in electronic approval documents and electronic identification documents.

In the electronic approval document 70, the shadow image information of the digital approval marks 71 and 72, which are liable to be illegally copied, is used as the original image information, in which audio ID information is embedded as digital watermark data.

Also, in the electronic identification card 73, the digitized face photo 84, which is liable to be illegally copied, is used as the original image information, in which audio ID information is embedded as digitized watermark data.

Figure 12 illustrates an overview of an electronic commerce or electronic approval system using voice ID information.

First, in the copyright production department 81, the copyright owner 82 creates a sound ID information bulletin 83 that only the user can know. The voice ID information 83 is added to the copyright information A 85 produced by the copyright holder 82 by the embedding method shown in FIG. 1 or FIG. 2 in the product sales department (or electronic approval creation department) 84. Using a program 86 for embedding the voice ID information 83, new copyright information B 87 with the voice ID information embedded is obtained. This copyright information B877 is used as an approved product or information as a distribution product. This work When it is necessary to check whether the right information B 87 is a genuine or illegally used item, the checking organization 90 performs an illegal judgment. In this case, the copyrighted work B 88 to be checked is passed through the detection program 91 for extracting the digital voice ID information by the voice ID information detection and separation method shown in FIG. 5 or FIG. 2 is detected and compared with the voice ID information registered in advance in the database 96 of the management unit 93, and the pass / fail is determined.

Here, the copyright holder 82 creates the voice ID information 83, but it is not always necessary for the copyright owner 82 to create it, and the copyright manager 95 creates it. Is also good. In addition, although the management department 93, the production department 81, and the check department 90 are configured separately, they need not necessarily be configured separately. Alternatively, any two departments may be integrated.

In general, a digital watermark is used only by the copyright holder in the event that encrypted information that would otherwise have no meaning in an unauthorized copy is decrypted in an unauthorized manner and that is illegally copied as restoration information. It is for extracting digital watermark information by comparing it with original information possessed, claiming copyright, or identifying an unauthorized copyist. Therefore, the audio ID watermark information extraction processing shown in FIGS. 8 and 9 is often used only when investigating fraud. The extraction process is performed by, for example, a watermark management company.

In FIG. 12, the product sales department (or electronic approval creation department) 84 can be regarded as a first information processing device, and the check organization 90 can be regarded as a second information processing device. The first information processing device 84 executes the audio ID embedding program 86 to generate and output the copyright information B 87 that incorporates the digital audio ID information into the copyright information A 85 . Second information processing device 90 inputs the copyright information B 87, extracts the digital voice ID information 92 from the information 87 as necessary, and can compare this with the ID information of the genuine copyright holder registered in the database in advance To

FIG. 13 shows an example of an information processing device (also referred to as a combination device) that uses the audio ID information for digital watermarking, and an example of an information processing network including the device.

The information processing network shown in Fig. 13 is a LAN (local area network), a WAN (wide area-network) such as an Internet network, and a wireless communication network. A system, and the one indicated by 104 indicates a transmission medium such as an optical fiber, an ISDN line, or a wireless line in the system. The transmission medium 104 includes, but is not limited to, a host computer device 103, and terminal computer devices 100, 101, and 102, typically shown via communication adapters 105, 106, and 107 such as routers and terminal adapters. The connected c-terminal computer device 100 includes, but is not limited to, a semiconductor integrated circuit processor (MPU) 109, and a display controller (DI SPC) 113 on an external bus 108, and a network. Network controller (NE TC) 114 and DRAM 115, and a disk controller connected to a peripheral circuit (not shown) built in the data processor 109. (FDC) 110, keyboard controller (KEY C) 111, and integrated device 'electronics' controller (IDEC) 112 are provided. The DISPC 113 controls drawing on the video RAM (VRAM) 121, and displays the drawn display data on the display (DISP) 120. The NE TC 114 is connected to the communication adapter 105, and performs transmission / reception information buffering and communication protocol control. DRAM 1 15 It is used for the program area and work area of the mouth processor 109. The FDC 110 is connected to a floppy disk drive device 116 to read information from and write information to a floppy disk 130 which is an example of a recording medium. A keyboard 117 is connected to the KEYC 111. A hard disk drive (HDD) 118 and a CD-ROM drive (CDRD) 119 are connected to IDE C 112. The HDD 118 has a magnetic disk which is another example of the recording medium. The CD RD 119 has a CD-ROM 131 which is another example of the recording medium. The other terminal computer devices 101 and 102 have the same configuration as above.c For example, when the digital audio ID information embedding process of FIG. 1 or FIG. 2 is performed using the terminal computer device 100, or When the digital voice ID information extraction processing shown in FIG. 5 or FIG. 6 is performed, a program for that is installed from the floppy disk 130 or the CD-OM 131 into the hard disk drive 118 by a user, for example. At this time, the program is recorded in the floppy disk 130 and the CD-ROM 131 in advance. In some cases, the set-up of the terminal computer 100 may provide the program preinstalled on the hard disk drive.

When executing the installed program, the data processor 109 loads the program into the DRAM 115, and sequentially fetches and executes instructions from the DRAM 115. A part of the program stored in the CD-ROM 131 can be directly taken out from the CD-ROM and executed. Thus, the terminal combination device 100 can install the program via the floppy disk 130 or the like, or can execute the program directly from the hard disk drive device 118 or the like. Therefore, according to the recording media 130 and 131 storing the program, the original Embedding of ID information by digital watermarking and embedded ID information, which has a small amount of embedded information relative to image information and is strong against disturbance to embedded original information and against missing information This can be easily realized by the terminal computer 100 as an example of the information processing device.

Further, the terminal-viewing device 100 can download the program from the host computer device 103. In other words, the host convenience device 103 holds, for example, the compressed program in a hard disk device or the like. After the terminal computer 100 establishes communication with the host computer 103, the terminal computer 100 designates the program and instructs the download, whereby the program is transmitted to the transmission medium 94. And transmitted to the hard disk drive 1118 of the terminal combination device 100. The downloaded program is then decompressed and installed in a predetermined program storage area. In this manner, the terminal computer 100 can easily acquire the program for embedding or separating the digital audio ID information via the transmission medium 94 on a network. Therefore, the transmission medium 94 has a small amount of embedded information relative to the original image information as described above, and is strong against disturbance to the embedded original information and also against information loss. It is useful for the terminal computer 100 to easily realize embedding of the ID information by the digital watermark and separation of the embedded ID information. The recording medium such as the floppy disk 130 and the CD-ROM 131 is a medium for recording and distributing the image information in which the digital audio ID information is embedded in the original image information as digital watermark information. is there. In addition, the transmission medium 104 can store the image information obtained by embedding the digital audio ID information in the original image information as digital watermark information in an information processing apparatus such as an information processing apparatus 100. It is also a medium for transmission to devices.

The information processing apparatus 100 or the like that inputs the image information recorded on the recording mediums 130 and 131 or the image information transmitted via the transmission medium 104 is provided with a digital watermark as necessary. It is possible to relatively easily realize the process of separating the digital voice ID information as information and confirming the content even in a low-cost information processing device.

The effects of the various embodiments for embedding the digital audio ID information described above as a digital watermark and separating and embedding the digital watermark for reproduction are summarized as follows.

1) Digital voice ID information is adopted as digital watermark information having meaning as ID information, so even if there is information loss of, for example, 50%, it can be recognized as meaning and watermark strength strong against disturbance can be easily realized. . In addition, it is easy to select a desired level of disturbance according to the demand by easily changing the audio parameters such as the number of quantization bits and the sampling frequency.

[2] In the case of digital audio ID information, the quality of the audio reproduced and extracted may be the lowest quality that can be understood.For example, when the number of quantization bits is 4 bits and the sampling frequency is 4 kHz, For example, a watermark data amount of 16 kbit / s (uncompressed) is sufficient. A reproduction time of 1 to 3 seconds is enough for use as ID information.

[3] Digital audio ID information is used as digital watermark information that has meaning as ID information, so even if noise is included in the extracted audio ID watermark information, it is only necessary to be able to recognize the meaning, and it is somewhat difficult to recognize the meaning. Even in a simple state, the meaning can be easily recognized and improved by removing noise by filtering. Therefore, the process of extracting the watermark information is easy, and the information processing device for that can be handled at a simple cost. [4] By adding automatic voice recognition, automatic recognition of ID information is also possible. <[5] Digit audio ID information does not have to be embedded in a position pattern like embedding image ID information. It is also easy to embed multiple in the original image information. Even for complicated rights related to digital image information, the ID information of the copyright holder or the rights holder can be easily and separably embedded.

[6] An embedding method using a watermark of a video device disclosed in the above-mentioned Japanese Patent Application Laid-Open No. H10-290404, or image information described in Japanese Patent Application Laid-Open No. H11-16686 The method of embedding the visible sound in the processing method is an invention that is captured from the viewpoint of using multimedia, which does not protect copyright, and the invention relating to the method of embedding the voice ID information of the present invention is the point of interest. And the means of realization. Therefore, the watermark embedding and separating methods and the like described above realize effects that cannot be achieved and cannot be achieved by the known techniques.

The invention made by the present inventor has been specifically described based on the embodiments, but the present invention is not limited thereto, and can be variously modified without departing from the gist thereof.

For example, the frequency transform is not limited to the above, and may be a wavelet transform. The basic embedding algorithm of the watermark information is not limited to the above, and the watermark information may be embedded by dividing the original image information into a number of blocks. Industrial applicability

INDUSTRIAL APPLICABILITY The present invention can be widely applied to various electronic approval, electronic commerce, and electronic distribution systems, or various information processing devices such as personal computers and mobile communication terminals used for the systems.

Claims

Scope of the request. An information processing method characterized by embedding digital audio ID information having meaning as ID information as invisible digital watermark information in original digital image information in a separable manner.

A conversion process of frequency-converting the digital original image information, an embedding process of embedding digital audio ID information having meaning as ID information in predetermined frequency components obtained in the conversion process as invisible digital watermark information, An inverse transformation process for performing an inverse transformation of the frequency transformation on the information that has undergone the embedding process. (A plurality of methods for embedding digital watermark information in digitized original image information.) Embedding position determination processing for determining the embedding position of the embedded image with reference to the digital original image information, order determination processing for determining the embedding order for the determined embedding position, and the embedding position according to the embedding order. And embedding digital audio ID information having meaning as ID information as invisible digital watermark information. Information processing method to be considered.

A conversion process for frequency-converting the digital original image information, and an embedding position determination process for determining, as embedding positions, a plurality of embedding frequency components for embedding digital watermark information from the frequency components obtained in the conversion process. An order determining process for determining the order of embedding for the determined plurality of embedding positions; and in accordance with the order of embedding, digitized voice ID information having meaning as ID information is embedded in the embedding position at an invisible electronic transmission. However, the information processing method includes: a process of embedding the information as information; and an inverse conversion process of performing an inverse conversion on the information after the embedding process with respect to the frequency conversion.

5. Parameter / digital calculation for calculating the digital / audio data configuration parameters defining the number of quantization bits, sampling frequency, and voice time of the digital / audio data ID information based on the digital original image information. 5. The information processing method according to claim 3, further comprising a process.

6. It further includes a parameter selection process for selecting a digit parameter configuration parameter that specifies the number of quantization bits, sampling frequency, and audio time of the digit parameter ID information from the parameter parameter table. The information processing method according to claim 3 or 4, wherein the information processing method is characterized in that:

7 · The digital audio ID information is an audio parameter of a sampling frequency selected from the range of 3 to 40 kHz, the number of quantization bits selected from 3 to 16 bits. 5. The information processing method according to claim 3, wherein the information is audio information specified by evening.

8. The digital voice ID information is characterized by being formed by quantizing with a 1-bit oversampling method and specified by voice parameters having a sampling frequency of 3 to 40 kHz. The information processing method according to claim 3 or 4.

9. The embedding position determination processing is performed such that when the digital audio ID information is already embedded as invisible digital watermark information, the current audio is taken to have a center of gravity different from the centroids of a plurality of already embedded embedding positions. 5. The information processing method according to claim 3, wherein an embedding position of the ID information is determined.

10 Separation processing for separating the digital audio ID information from the information in which the digital audio ID information having meaning as the ID information is embedded in the original digital image information in a separable manner as invisible digital watermark information. And minutes An audio reproduction process for reproducing the separated digital audio ID information.

11. Frequency conversion processing for performing frequency conversion on information in which digital audio ID information having meaning as ID information is embedded as invisible digital watermark information in a predetermined frequency component of the original digital image information And a sound reproducing process for separating the digitized audio ID information from the information obtained by the frequency conversion process, and a sound reproducing process for reproducing the separated digitized voice ID information. Method.

1 2. A process for embedding the digital audio ID information having meaning as ID information in the original digital image information in a separable manner as invisible digital watermark information, and processing the digital audio ID information from the embedded information. A recording medium characterized by recording a program for causing an information processing device to execute a process for separating the information and a program for causing the information processing device to read the program.

1 3. First conversion processing for frequency-converting the original digital image information, and digital audio ID information having meaning as ID information at a predetermined frequency component obtained in the first conversion processing as invisible digital watermark information. An embedding process for embedding, an inverse conversion process for performing an inverse conversion on the frequency conversion on the information that has passed through the embedding process, a second conversion process for converting the inversely converted information into frequency components, and a second conversion process The recording process is characterized in that a separation process for separating the above-mentioned digitized voice ID information from the information obtained in step (1) and a program for causing the information processing apparatus to execute (3) are recorded in the information processing apparatus in a readable manner. Medium.

4. A recording medium in which image information is recorded in an information processing device in a readable manner, wherein the image information is a digital original image information, and a digital audio ID having meaning as ID information is invisible digital watermark. As information A recording medium comprising information embedded in a separable manner.

5. A recording medium in which image information is recorded in an information processing apparatus in a readable manner, wherein the image information is digitized audio ID information having a meaning as ID information at a predetermined frequency component of the digitized original image information. Is embedded as invisible electronic watermark information, and the digital audio

A recording medium characterized in that the ID information is separable information; c6. A recording medium on which image information is recorded in an information processing apparatus in a readable manner, wherein the image information has a meaning as an ID. The digital audio ID information is embedded as invisible digital watermark information in the sample value area of the original digital image information or in the frequency-converted frequency domain of the original digital image information, and the embedded digital information is embedded. A recording medium, wherein the evening voice ID information is information that can be separated from the sample value area or the frequency area by an information processing device.

7. Digit audio having meaning as ID is used as invisible electronic watermark information in the sample value area of the original digital image information or in the frequency-converted frequency domain of the original digital image information. A first information processing device that outputs image information embedded so as to be separable; and a second information processing device that inputs the image information and uses the input image information as electronic approval information. An information processing system, comprising:

8. Digital voice ID information with meaning as ID can be separated as invisible electronic watermark information into the sample value area of the original digital image information or into the frequency-converted frequency domain of the original digital image information. A first information processing device that outputs image information embedded in a document, and a second information processing device that inputs the image information and uses the input image information as electronic signature information Information processing characterized by being system.

1 9. Digit audio having meaning as ID The ID information is used as invisible electronic watermarking information in the sample value area of the original digital image information or in the frequency domain obtained by frequency-converting the original digital image information. An information processing system, comprising: an information processing device that stores image information embedded separably in a document and outputs the stored image information in response to a distribution request.

20. Input means and arithmetic control means, wherein the input means inputs digital original image information, and the arithmetic control means has a meaning as ID information in the inputted digital original image information. An information processing apparatus characterized in that digitized voice ID information can be separated and embedded as invisible digital watermark information.

21. A storage unit and an operation control unit, wherein the storage unit stores digital original image information, and the operation control unit performs frequency conversion on the digital original image information stored in the storage unit, Digitized voice ID information having meaning as ID information is embedded as invisible digital watermark information in the predetermined frequency component obtained by the frequency conversion, and the information after the embedding is inversely converted to the frequency conversion. An information processing apparatus characterized in that the information processing apparatus can perform the processing.

22. A storage unit and an operation control unit, wherein the storage unit stores digital original image information, and the operation control unit refers to the original original image information stored in the storage unit. A plurality of embedding positions for embedding digital watermark information in the original digital image information are determined, an embedding order for the determined embedding position is determined, and the embedding position is given meaning as ID information in accordance with the embedding order. Digital voice ID information can be embedded as invisible digital watermark information An information processing apparatus characterized in that the information processing apparatus is capable of functioning.

3. Includes storage means and arithmetic control means, wherein the storage means stores digital original image information, wherein the arithmetic control means frequency-converts the digital original image information stored in the storage means, A plurality of embedding frequency components for embedding digital watermark information are determined as embedding positions from among the frequency components obtained by the frequency conversion, and an embedding order for the determined plurality of embedding positions is determined. It is possible to embed digital audio ID information having meaning as ID information in the embedding position as invisible electronic watermark information in the embedding position in accordance with the order of the above, and to perform the inverse conversion for the frequency conversion on the information after the embedding process 4. The arithmetic control means, when determining the embedding position, has digital voice ID information. When embedded as invisible digital watermark information, the embedding position of the current voice ID information is determined so that a center of gravity different from the centroid of a plurality of already embedded embedding positions is adopted. The information processing device according to claim 23.

5. An input unit, a storage unit, and an arithmetic control unit, wherein the storage unit stores the invisible digital watermark information in a sample value area of the original digital image information or a frequency-converted frequency area of the original digital image information. The embedding position for embedding and the order of the positions are determined in advance by embedding control information, and the audio parameter information of the digital audio ID information, which has meaning as digital watermark information, and the audio parameter information that is predetermined. The input means inputs digital original image information, and the arithmetic and control means selects required voice parameters and embedding control information from the storage means, and stores a sample value area or data of digital original image information. The digital audio ID information according to the selected audio parameter is stored in the frequency domain of the original image information after the frequency conversion. An information processing device characterized by being embedded in a position and an order specified by control information.

6. Input means, storage means, and arithmetic control means, wherein the storage means stores the invisible digital watermark information in the sample value area of the original digital image information or the frequency-converted frequency area of the original digital image information. The embedding position for embedding and the embedding control information in which the order of the positions are pre-encrypted, and the voice configuration parameters of the digital voice ID information that have meaning as digital watermark information The input means inputs digital original image information, and the arithmetic control means selects required encrypted voice parameters and encrypted embedding control information from the storage means, and selects The decrypted encrypted voice parameter and the encrypted embedded control information are decrypted, and the sampled value area of the digital original image information or the frequency-converted frequency of the digital original image information is decoded. Information processing for embedding digit voice ID information in accordance with the decoded voice parameter information in a wave number domain in a position and an order specified by the decoded embedding control information. apparatus.

7. An input means and an arithmetic control means, the input means being capable of separating digital original image information into which digital audio ID information having meaning as ID information is separably embedded as invisible digital watermark information. An information processing apparatus, wherein the arithmetic control means is capable of separating the digital voice ID information from the input information.

8. Includes input means and arithmetic control means, wherein the input means has digital audio ID information having meaning as ID information embedded in predetermined frequency components of the digital original image information as invisible digital watermark information. Information, and the arithmetic and control unit performs frequency conversion on the input information, and is capable of separating the digital voice ID information from information obtained by the frequency conversion. apparatus.

9. The claim range, wherein the arithmetic control means is further capable of performing a noise removal filtering process on the separated digitized voice ID information. An information processing apparatus according to claim 1.

20. The method according to claim 29, wherein the arithmetic control means is further capable of performing a speech recognition process on the digitized speech ID information subjected to the noise removal filtering process. Item.