CN1148970C - Digit image coding/decoding device and method using water mark - Google Patents

Abstract

Provided are a digital image encoding/decoding device using a watermark and a method thereof. The digital image coding device includes: a discrete wavelet transform part, which outputs M×M discrete wavelet coefficients for the main image to be transmitted through discrete wavelet transform; an effective coefficient extraction part, which extracts an upper limit effective coefficient; a pseudo-random number generator, which generates Pseudo-random numbers; coefficient replacement/combination part, obtain replacement position data representing N×N replacement positions, replace and combine coefficients corresponding to N×N replacement positions with N×N effective wavelet systems. The current image is used to correct the distorted main image, so the information flow is continuous. No errors are generated in correcting image processing even if the image changes suddenly.

Description

Digital image encoding/decoding apparatus and method using watermark

This application is related to provisional patent application US Serial No. 60/091,540, entitled "Method of Assessing Image Quality Using Watermarking," filed July 1, 1998 by the applicant of the present invention.

technical field

The present invention relates to a digital image encoding device, a decoding device, and an encoding and decoding method using a watermark, and more particularly, relates to making information flow by correcting a distorted main image using a signature image when the distortion of the main image is severe. A continuous digital image encoding device, decoding device, and encoding and decoding method.

Background technique

Watermarking is a technique of transmitting an image to be transmitted together with an encrypted image for the purpose of providing security and authentication. The receiver side receives an image (hereinafter referred to as a main image) transmitted together with a secretly transmitted encrypted image (hereinafter referred to as a signature image). The signature image is decoded by a specific decryption device.

FIG. 1 is a conceptual block diagram illustrating a conventional digital image codec apparatus for generating a watermarked image and extracting a signature image from the watermarked image. In a conventional digital image encoding device to be described, during encoding, a main image and a secretly transmitted signature image are subjected to DCT (Discrete Cosine Transform), thereby extracting DCT coefficients of the corresponding images. The DCT coefficients are encoded by an encoder. Here, the DCT coefficient components of the signature image are encoded by a separate encryption encoder for encryption control operations. In this way, the DCT coefficients of the main image and the DCT coefficients of the signature image inserted into the DCT coefficients of the encoded main image can be transmitted. IDCT (Inverse Discrete Cosine Transform) is performed on the DCT coefficients of the main image and the DCT coefficients of the signature image inserted into the DCT coefficients of the main image, thereby obtaining a watermarked image in which only the main image can be seen but not the signature image.

The signature image watermarked on the main image is decoded by a specific decryption device to restore it. During restoration, DCT is performed on the watermarked image to extract its DCT coefficients. Next, the extracted DCT coefficients obtained by performing DCT on the original main image are separated from the DCT coefficients of the watermarked image, thereby extracting the DCT coefficient components of the encoded signature image. Then, the decoder decodes the DCT coefficient components of the coded signature image. Here, the encryption decoder performs the control operation of decryption. IDCT is performed again on the DCT coefficients of the signature image decoded by the decryption control operation, thereby restoring the signature image.

As described above, in a conventional digital image encoding device using a watermark, if it is determined that the current main image is severely distorted, the current image is corrected by the decoding unit using the previous main image. However, if a previous image is used instead of the current image in corrective image processing, the flow of information becomes discontinuous. In addition, when an image changes suddenly, an error may occur in correcting image processing.

Contents of the invention

An object of the present invention is to provide a digital image encoding/decoding apparatus and method thereof capable of correcting a distorted main image transmitted from a heavily noisy environment using a current image.

In order to realize the above-mentioned purpose of the present invention, a kind of digital image coding device is provided, and this device comprises: a discrete wavelet (wavelet) transformation part, is used for carrying out discrete wavelet transformation to the main image to be transmitted on it, thereby output M × M discrete wavelet coefficients, wherein M is a predetermined positive integer; an effective coefficient extraction part is used to extract a predetermined percentage of upper effective coefficients from M×M wavelet coefficients; a pseudo-random number generator is used for encryption according to A predetermined rule corresponding to the code generates a pseudo-random number; a coefficient replacement/combination part is used to obtain replacement position data representing N×N replacement positions, where N is a predetermined positive integer smaller than M, and N×N effective wavelets are used Coefficient replacement and combination of coefficients corresponding to N×N replacement positions selected among M×M wavelet coefficients; an inverse discrete wavelet transform section for inverse discrete wavelet transform on the replaced M×M wavelet coefficients a transform to produce a combined main image; and a compression section to compress the combined main image and the encryption code.

In addition, the predetermined positive integer N is preferably less than or equal to 0.5M.

According to another aspect of the present invention, there is provided a digital image decoding apparatus for decoding a bitstream encoded using a watermark, the apparatus comprising: a decompression section for decompressing the compressed bitstream to restore the main image and predetermined encryption code; an image quality evaluation section for evaluating the quality of the restored main image, and if the restored main image quality is lower than a predetermined level, a control signal that becomes a first logic state is generated, and if the restored main image quality generating a control signal that changes to a second logic state at no lower than a predetermined level; a discrete wavelet transform section for performing discrete wavelet transform on the restored main image in response to the first logic state control signal to obtain M×M wavelets Coefficient, M is a predetermined positive integer; a pseudo-random number generator, used to generate pseudo-random numbers according to the rules corresponding to the recovery encryption code; a coefficient separation part, used to use pseudo-random numbers to obtain N×N replacement positions Separation position data, N is another predetermined positive integer smaller than M, and N × N wavelet coefficients separating N × N separation positions from M × M wavelet coefficients; an inverse discrete wavelet transform part for N ×N wavelet numbers for inverse discrete wavelet transform to produce the restored signature image; an image scale conversion section for adding the scale of the restored signature image to the scale of the main image, and an image selection section, For selecting a scale-increasing signature image in response to a first logic state control signal, and selecting a restored main image in response to a second logic state control signal, to output the selected image as the main image.

According to another aspect of the present invention, there is provided a digital image codec device using watermark, the device includes: a first discrete wavelet transform part, used for performing discrete wavelet transform on the main image to be transmitted, thereby outputting M×M discrete wavelet coefficients, wherein M is a predetermined positive integer; an effective coefficient extracting part for extracting a predetermined percentage of upper effective coefficients from the M×M wavelet coefficients; a first pseudo-random number generator, Used to generate pseudo-random numbers according to predetermined rules corresponding to encrypted codes; a coefficient replacement/combination part is used to obtain replacement position data representing N×N replacement positions, where N is a predetermined positive integer smaller than M, and N×N N effective wavelet coefficients replace and combine coefficients corresponding to N×N replacement positions selected in M×M wavelet coefficients; a first inverse discrete wavelet transform part is used for replacing M×M wavelets The coefficients undergo an inverse discrete wavelet transform to produce the combined main image; a compression section compresses the combined main image and encryption code; a decompression section decompresses the compressed bitstream to recover the main image and encryption Code; an image quality evaluation part for evaluating the quality of the restored main image, and if the restored main image quality is lower than a predetermined level then generating a control signal that becomes a first logic state, and if the restored main image quality is not low At a predetermined level, a control signal that becomes a second logic state is generated; a second discrete wavelet transformation part is used to perform discrete wavelet transformation on the recovered main image in response to the first logic state control signal, so as to obtain M×M sub-wavelets The wave coefficient, M is a predetermined positive integer; a second pseudo-random number generator, used to generate pseudo-random numbers according to the rules corresponding to the restored encrypted code; a coefficient separation part, used to use the pseudo-random numbers to obtain representation N×N Separation position data of replacement positions, N is another predetermined positive integer less than M, and N × N wavelet coefficients corresponding to N × N separation positions are separated from M × M wavelet coefficients; a second inverse discrete sub A wavelet transform part is used to inverse discrete wavelet transform on N×N wavelet coefficients to generate the restored signature image; an image scale conversion part is used to increase the scale of the restored signature image to the scale of the main image and an image selection section for selecting a scale-increased signature image in response to a first logic state control signal, and selecting a restored main image in response to a second logic state control signal to output the main image.

According to another aspect of the present invention, there is provided a digital image coding method using watermark, the method includes the steps of: (a) performing discrete wavelet transform on the main image to be transmitted, thereby outputting M×M discrete wavelets Coefficients, wherein M is a predetermined positive integer; (b) extracting a predetermined percentage of upper effective coefficients from M×M wavelet coefficients; (c) generating a pseudo-random number according to a rule corresponding to a predetermined encryption code; and (d) obtaining Replacement position data representing replacement positions of N×N pixels, where N is a predetermined positive integer smaller than M; (e) replacing and combining N selected among M×M wavelet coefficients with N×N effective wavelet coefficients ×N coefficients at replacement positions; (f) performing an inverse discrete wavelet transform on the replaced M×M wavelet coefficients to produce a combined main image; and (g) compressing the combined main image and N×N pixels location data.

According to yet another aspect of the present invention, there is provided a digital image decoding method for decoding a bitstream coded using a watermark, the method comprising the steps of: (a) evaluating the quality of the combined main image and determining whether the quality of the combined main image is greater than or equal to a predetermined level; (b) if it is determined in step (a) that the quality of the combined main image is greater than or equal to a predetermined level, then set the combined main image as the main image; (c) if it is determined in step (a) that the combined main image is lower than a predetermined level, then set the signature image embedded in the combined main image as the main image; and (d) decompress the compressed bitstream to restore the combined main image and the predetermined encryption code; wherein step (c) further The method includes the following steps: performing discrete wavelet transform on the recovered main image to obtain M×M wavelet coefficients, M is a predetermined positive integer; generating a pseudo-random number according to a predetermined rule corresponding to the restored encryption code; using the pseudo-random number to obtain Represents the separation position data of N×N replacement positions, N is another positive integer less than M; separate N×N wavelet coefficients corresponding to N×N separation positions from M×M wavelet coefficients; for N×N performing an inverse discrete wavelet transform on wavelet coefficients to produce a restored signature image; converting the scale of the restored signature image to the scale of the main image; and setting the scale-transformed signature image as the main image.

Description of drawings

The above objects and advantages of the present invention will become more apparent by describing in detail preferred embodiments of the present invention with reference to the accompanying drawings, in which:

1 is a conceptual block diagram illustrating a conventional digital image codec apparatus for generating a watermarked image and extracting a signature image from the watermarked image;

Fig. 2 is a block diagram illustrating a digital image codec device according to the present invention;

Fig. 3 is a flow chart showing the steps of the digital image encoding method according to the present invention; and

Fig. 4 is a flowchart showing the steps of the digital image decoding method according to the present invention.

Detailed ways

Referring to FIG. 2 , the digital image codec device according to the present invention includes an encoding unit 20 and a decoding unit 22 . The encoding unit 20 includes a discrete wavelet transform section 202 , a significant coefficient extracting section 204 , a pseudorandom number generator 206 , a coefficient replacing/combining section 208 , an inverse discrete wavelet transform section 210 and a compressing section 212 . In addition, the decoding unit 22 includes a decompression section 222, an image quality evaluation section 224, a discrete wavelet transform section 226, a pseudorandom number generator 228, a coefficient separation section 230, an inverse discrete wavelet transform section 232, An image scale conversion section 234 and an image selection section 236 .

First, the operation of the encoding unit 20 is described with reference to FIGS. 2 and 3 .

The discrete wavelet transformation part 202 receives the main image to be transmitted, and performs discrete wavelet transformation (step 302 ) to obtain discrete wavelet coefficients. As known in the art, discrete wavelet coefficients consist of a square matrix. In this embodiment, it is assumed that the discrete wavelet coefficients are M×M, and M is a predetermined positive integer.

The significant coefficient extracting section 204 extracts 25% upper limit significant coefficients from the M*M wavelet coefficients (step 304).

The pseudo-random number generator 206 generates a pseudo-random number according to a rule corresponding to a predetermined encryption code input by the user's option, and the coefficient replacement/combination section 208 selects N*N replacement positions (step 306). In addition, the coefficient replacing/combining section 208 replaces and combines the coefficients at N×N replacement positions selected among the M×M wavelet systems with N×N effective wavelet coefficients (step 308). Here, since 25% upper effective coefficients are selected among the M×M wavelet coefficients in step 304, N is set to be half of M, ie 0.5M. In view of the characteristics of the watermark and considering the impact on the original main image, N is preferably less than or equal to 0.5M. The effective wavelet coefficients of the main image to be transmitted are randomly distributed in the combined main image along with the wavelet coefficients of the signature image.

The inverse discrete wavelet transform section 210 performs inverse discrete wavelet transform on the replaced M×M wavelet coefficients to generate a combined main image (step 310). Since the combined main image has the signature image included therein with the watermark, it is not much different from the original main image, and the signature image cannot be seen clearly with the naked eye.

The compression section 212 compresses the combined main image and the encryption code to output a compressed bit stream (step 312). Here, the encryption code is also compressed for the purpose of selecting the separation position with the same interval as the replacement position used during encoding.

As described above, the encoding unit 20 pseudo-randomly rearranges the upper-limit effective wavelet coefficients of the main image to be transmitted among the own wavelet coefficients of the main image. A valid portion of the main image is then watermarked in the main image itself.

The main image watermarked by the encoding unit 20 is decoded by the decoding unit 22 .

The operation of the decoding unit 22 will now be described with reference to FIGS. 2 and 4 .

The decompression section 222 decompresses the compressed bit stream to restore the main image and encrypted code (step 402). The encryption code is used to determine the generation rules of the pseudo-random numbers used for coefficient replacement and combination in the coding unit 20 . In the decoding unit 22, a pseudo-random number generation rule for separating coefficients described later is determined from an encryption code.

The image quality evaluation section 224 evaluates the quality of the restored main image and determines whether the quality of the restored main image is greater than or equal to a predetermined level (step 42). Therefore, if the restored main image quality is lower than the predetermined level, the image quality evaluating part 224 generates a control signal that becomes logic "high", and if the restored main image quality is not lower than the predetermined level, the image quality evaluating part 224 generates a logic "high" control signal. "LOW" control signal. For example, image quality evaluation can be performed in the same manner as distortion evaluation.

The DWT part 226 performs DWT on the restored main image in response to the logic "high" control signal to obtain M×M wavelet coefficients, where M is a predetermined positive integer (step 460).

The pseudo-random number generator 228 generates pseudo-random numbers according to the rules corresponding to the recovered encrypted code (step 462). The pseudo-random number generator 228 performs the same operation as the pseudo-random number generator 206 of the encoding unit 20 . Therefore, if a restored encryption code is input, that is, the same encryption code as used in the encoding unit 20, a pseudo-random number is generated according to the same rule as used in the encoding unit 20.

The coefficient separation section 230 selects N*N separation positions using a pseudo-random number, N being another predetermined integer smaller than M (step 464). In addition, the coefficient separation section 230 separates N×N wavelet coefficients at N×N separation positions from the M×M wavelet coefficients (step 466 ). Here, N is determined in the encoding unit 20 during coefficient replacement and combination, and N is preferably less than or equal to 0.5M.

The inverse discrete wavelet transform section 232 performs inverse discrete wavelet transform on the N×N wavelet coefficients to generate a restored signature image.

The image scale conversion section 234 converts the scale of the restored signature image into the scale of the main image (step 470). Set the scale-converted signature image as the main image.

Finally, the image selection section 236 selects the scale-converted signature image in response to the logic "high" control signal, and selects the restored main image in response to the logic "low" control signal to output the selected image as the main image.

Therefore, if the quality of the restored main image is greater than or equal to a predetermined level, the decoding unit 22 selects the combined main image as the main image (step 44), and then outputs it. On the other hand, if the quality of the restored main image is lower than the predetermined level, the decoding unit 22 restores the signature image embedded in the main image, and sets the restored signature image as the main image (step 46), and then outputs.

As described above, according to the present invention, since a distorted main image transmitted from a heavily noisy environment is corrected using the current image, the flow of information is continuous. In addition, even when the image changes suddenly, no error occurs in the process of correcting the image.

In addition, the digital image coding and decoding unit of the digital image codec device according to the present invention can be realized separately by one digital image coding device and one digital image decoding device.

In addition, digital image encoding and decoding methods can be formulated in a computer program. In addition, the digital image encoding and decoding methods can be realized with a general-purpose digital computer for operating a program employing a medium used in the computer. The media include magnetic recording media such as floppy disks or hard disks, and optical recording media such as CD-ROMs or DVDs. In addition, programmers in the field related to the present invention can easily derive the function programs, codes, and code segments.

Claims (8)

1. A digital image encoding device using a watermark, comprising: A discrete wavelet transform part, which is used to perform discrete wavelet transform on the main image to be transmitted, thereby outputting M×M discrete wavelet coefficients, where M is a predetermined positive integer; an effective coefficient extracting part for extracting a predetermined percentage of upper effective coefficients from the M×M wavelet coefficients; A pseudo-random number generator for generating N×N pseudo-random numbers according to a predetermined rule corresponding to the encryption code, wherein N is a predetermined positive integer smaller than M, and the N×N pseudo-random numbers represent the corresponding M×M A replacement position selected from M×M positions of wavelet coefficients; a coefficient replacement/combination section for obtaining replacement position data representing N×N replacement positions, replacing and combining N×N effective wavelet coefficients with N×N replacements selected among M×M wavelet coefficients The coefficient corresponding to the position; an inverse discrete wavelet transform section for performing an inverse discrete wavelet transform on the substituted M×M wavelet coefficients to produce a combined main image; and A compression section that compresses the combined main image and encrypted code. 2. The digital image encoding device according to claim 1, wherein the predetermined positive integer N is less than or equal to 0.5M. 3. A digital image decoding apparatus for decoding a bitstream encoded using a watermark, comprising: a decompression section for decompressing the compressed bitstream to recover the main image and predetermined encryption code; an image quality evaluation section for evaluating the quality of the restored main image, and generating a control signal that changes to a first logic state if the restored main image quality is lower than a predetermined level, and if the restored main image quality is not lower than a predetermined level generating a control signal to change to a second logic state; a discrete wavelet transform section, configured to perform discrete wavelet transform on the recovered main image in response to the first logic state control signal, to obtain M×M wavelet coefficients, where M is a predetermined positive integer; A pseudo-random number generator for generating pseudo-random numbers according to the rules corresponding to the recovery encryption code; a coefficient separation section for obtaining separation position data representing N×N replacement positions using a pseudorandom number, N being another predetermined positive integer smaller than M, and separating the N×N separation positions from the M×M wavelet coefficients The N×N wavelet coefficients of ; an inverse discrete wavelet transform section for performing an inverse discrete wavelet transform on the N×N wavelet coefficients to produce a recovered signature image; an image scale conversion section for increasing the scale of the recovered signature image to the scale of the main image; and An image selection section for selecting the scale-increased signature image in response to the first logic state control signal, and selecting the restored main image in response to the second logic state control signal to output the selected image as the main image. 4. The digital image decoding device according to claim 3, wherein the predetermined positive integer N is less than or equal to 0.5M. 5. A digital image codec device using a watermark, comprising: A first discrete wavelet transform part, which is used to perform discrete wavelet transform on the main image to be transmitted, thereby outputting M×M discrete wavelet coefficients, wherein M is a predetermined positive integer; an effective coefficient extracting part for extracting a predetermined percentage of upper effective coefficients from the M×M wavelet coefficients; A first pseudo-random number generator, used to generate N×N pseudo-random numbers according to predetermined rules corresponding to the encryption code, wherein N is a predetermined positive integer less than M, and the N×N pseudo-random numbers represent A replacement position selected from the M×M positions of the ×M wavelet coefficients; a coefficient replacement/combination section for obtaining replacement position data representing N×N replacement positions, replacing and combining N×N effective wavelet coefficients with N×N replacements selected among M×M wavelet coefficients The coefficient corresponding to the position; a first inverse discrete wavelet transform section for performing inverse discrete wavelet transform on the replaced M×M wavelet coefficients to produce a combined main image; a compression section for compressing the combined main image and encrypted code; a decompression section for decompressing the compressed bitstream to recover the main image and encrypted code; an image quality evaluation section for evaluating the quality of the restored main image, and generating a control signal that changes to a first logic state if the restored main image quality is lower than a predetermined level, and if the restored main image quality is not lower than a predetermined level The level generates a control signal that changes to a second logic state; a second discrete wavelet transform part, configured to perform discrete wavelet transform on the recovered main image in response to the first logic state control signal, to obtain M×M wavelet coefficients, where M is a predetermined positive integer; A second pseudo-random number generator for generating pseudo-random numbers according to the rules corresponding to the restored encrypted code; a coefficient separation section for obtaining separation position data representing N×N replacement positions using a pseudo-random number, N being another predetermined positive integer smaller than M, and separating from M×M wavelet coefficients and N×N separation N×N wavelet coefficients corresponding to the position; a second inverse discrete wavelet transform section for performing inverse discrete wavelet transform on the N×N wavelet coefficients to generate a restored signature image; an image scale conversion section for increasing the scale of the recovered signature image to the scale of the main image; and An image selection section for selecting the scale-increased signature image in response to the first logic state control signal, and selecting the restored main image in response to the second logic state control signal to output the main image. 6. A digital image encoding method using a watermark, comprising the steps of: (a) performing discrete wavelet transform on the main image to be transmitted, thereby outputting M×M discrete wavelet coefficients, wherein M is a predetermined positive integer; (b) Extracting a predetermined percentage of upper effective coefficients from the M×M wavelet coefficients; (c) Generate N×N pseudo-random numbers according to the rules corresponding to predetermined encryption codes, wherein N is a predetermined positive integer smaller than M, and N×N pseudo-random numbers represent M The replacement position selected from the ×M positions; (d) obtaining replacement position data representing replacement positions of N×N pixels; (e) replacing and combining coefficients at N×N replacement positions selected among the M×M wavelet coefficients with N×N effective wavelet coefficients; (f) performing an inverse discrete wavelet transform on the replaced M×M wavelet coefficients to produce a combined main image; and (g) Compressing the combined main image and NxN pixel position data. 7. The digital image coding method according to claim 6, wherein the predetermined positive integer N is less than or equal to 0.5M. 8. A digital image decoding method for decoding a bitstream encoded using a watermark, comprising the steps of: (a) decompressing the compressed bitstream to recover the combined main image and predetermined encryption code; (b) assessing the quality of the combined main image and determining whether the quality of the combined main image is greater than or equal to a predetermined level; (c) if it is determined in step (a) that the quality of the combined main image is greater than or equal to a predetermined level, setting the combined main image as the main image; (d) if in step (a) it is determined that the quality of the combined main image is below a predetermined level, setting the signature image embedded in the combined main image as the main image; and Wherein step (d) further comprises the following steps: performing discrete wavelet transform on the recovered main image to obtain M×M wavelet coefficients, where M is a predetermined positive integer; generating a pseudo-random number according to a predetermined rule corresponding to the recovered encrypted code; Obtaining separated position data representing N×N replacement positions using a pseudo-random number, N being another positive integer less than M; separating N×N wavelet coefficients corresponding to N×N separation positions from the M×M wavelet coefficients; performing an inverse discrete wavelet transform on the N×N wavelet coefficients to produce a restored signature image; convert the scale of the recovered signature image to the scale of the master image; and Set the scale-transformed signature image as the main image.

Images