KR100479478B1 - Object-based transcoding method with the importance-degree of each object and apparatus thereof - Google Patents

Abstract

The present invention discloses an object-based transcoding method and apparatus in consideration of object-specific importance.

According to the present invention, input data including an image is divided into objects according to the shape, size, contrast, and connectivity of an object in the image, and a predetermined importance index is determined for each object for the divided objects. And encoding the input data including the image by inserting the importance index for each object as a watermark in a corresponding object in the spatial domain or the frequency domain, and inserting a value according to the importance of the object as a watermark and associating it with the watermark. By transcoding by assigning different bit rate for each object, it is possible to insert the importance of each object into the object and to perform compression according to the object for each object, so that it is possible to transmit video and video data effectively. Clearer display than objects and unique watermarks prevent illegal duplication The distribution of copyrighted works through the communication network is effective.

Description

{Object-based transcoding method with the importance-degree of each object and apparatus}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image processing, and more particularly, to a method and apparatus for transcoding by an object unit capable of copy protection and efficient compression.

Video and video data are inevitably compressed due to the huge amount of information in transmission and storage. In many cases, in order to solve the problems caused by limited channel capacity or storage capacity when transmitting and storing compressed video and video data, a technique of recompressing data at a desired bit rate is necessary.

In addition, MPEG-4, a new video coding standard, can express images based on objects, and can encode image objects having arbitrary shapes into independent bit streams, and decode only necessary objects from bit streams. have. Compared with the existing frame-based compression techniques, multimedia information is represented and encoded based on objects, and MPEG-4 can accommodate a variety of newly born multimedia applications. In addition, JPEG2000, which seeks to provide better image quality than the previous standard at a low bit rate of 0.25 bits or less per pixel, allows arbitrary access to important areas within the image or a region of interest (ROI).

As such, standardization is currently progressing or recently completed compression techniques are targeting an image object having an arbitrary shape. Therefore, in consideration of various applications of MPEG-4 and standardization directions of JPEG2000 in the future, the conventional transcoding unit also needs a transcoding method of an object unit deviating from the existing frame basis.

In the case of using the conventional transcoding method, the characteristics of the objects existing in the image are stored in a database, and when the transcoding for each object is needed, the importance of the object is calculated using the information. However, these informations do not accurately reflect the importance of the objects, and it is necessary to manage a separate database to calculate the importance, so that the data necessary for coding or compressing the image is large, and the size of the system for implementing the size is inefficient. .

In this case, there is a problem in that efficient compression is not performed when uniform compression is performed without considering the importance of each object inserted into the image data.

The technical problem to be solved by the present invention is to solve the above problems, by inserting an element that can indicate the importance of the object directly into the object, security or copyright protection by making the importance of the object as a watermark (watermark) A method and apparatus for enabling an apparatus, and a transcoding method and apparatus for applying different bit rates for each object based on the importance can be determined by extracting the inserted information at any time when transcoding on a per-object basis. To provide.

Another technical problem to be solved by the present invention is to provide a computer-readable recording medium which records a program including the above method.

In accordance with an aspect of the present invention, an object-based transcoding method includes: (a) dividing input data including an image into objects according to shapes, sizes, contrasts, and connectivity of objects in the image; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark, and in step (c), The encoded materiality index is repeatedly inserted into an object in blocks of a predetermined size, or when the input data is video / video data, compression is performed in the format of MPEG-4 or JPEG2000 when encoding in the step (c). At this time, the encoded importance index is inserted using coefficients of the DCT block in each object, or if the input data is bitstream data, only the object to which the importance index is to be inserted in step (c) is decoded to count the DCT block. It is characterized by inserting the encoded importance index using this coefficient.

In accordance with another aspect of the present invention, there is provided an object-based transcoding apparatus comprising: an image data receiver configured to receive input data including an image; An object dividing unit dividing the received input data into objects according to a shape, size, contrast, and connectivity of an object included in an image; An index input unit which receives an importance index for each object of the divided objects and encodes the importance index for each corresponding object; And a coder for inserting the received importance index for each object into a corresponding object by inserting the watermark in a spatial domain and a frequency domain into a corresponding object to encode the input data. Is repeatedly inserted into an object in blocks of a predetermined size or compressed in the form of MPEG-4 or JPEG2000 when the input data is video / video data, and the coefficients of the DCT blocks in each object are compressed. Insert the encoded importance index or decode only the object to which the importance index is to be inserted if the input data is bitstream data to obtain a coefficient of the DCT block, and use the coefficient to insert the encoded importance index. Characterized in that.

According to the present invention for solving the above technical problem, a method of transcoding encoded image data in which a importance index is inserted into a watermark for each object, the method comprising: (a) an object according to the type of data; Decoding each other; (b) separating the decrypted objects for each object; (c) extracting an importance index for each separated object; (d) allocating a predetermined bit rate for each object according to the importance index for each object; And (e) transcoding and encoding each object according to the bit rate allocated to each object.

According to an embodiment of the present invention for solving the above technical problem, an apparatus for transcoding image data in which a importance index is inserted as a watermark for each object, wherein each object decodes the image data for each object according to a type of data. A decoder; An object separator that separates the decoded objects by objects; An importance index extractor configured to extract the importance index by considering the importance index as the predetermined noise for each of the separated objects; A bit rate allocation unit for allocating a predetermined bit rate for each object according to the extracted importance index; And a transcoding unit configured to transcode and encode each object according to the bit rate allocated to each object.

According to another aspect of the present invention, a computer-readable recording medium having recorded thereon an object-based transcoding program includes: (a) shape, size, and contrast of input data including an image; Dividing into objects according to connectivity; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark, and in step (c), The encoded materiality index is repeatedly inserted into an object in blocks of a predetermined size, or when the input data is video / video data, compression is performed in the format of MPEG-4 or JPEG2000 when encoding in the step (c). At this time, the encoded importance index is inserted using coefficients of the DCT block in each object, or if the input data is bitstream data, only the object to which the importance index is to be inserted in step (c) is decoded to count the DCT block. It is characterized by inserting the encoded importance index using this coefficient.

According to the present invention for solving the above-mentioned technical problem, a recording medium recording a method of transcoding video data having an importance index inserted as a watermark for each object, comprising: (a) assigning the video data to a type of data; Decoding each object accordingly; (b) separating the decrypted objects for each object; (c) extracting an importance index for each separated object; (d) allocating a predetermined bit rate for each object according to the importance index for each object; And (e) transcoding and encoding each object according to the bit rate allocated to each object.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 illustrates a flow of an object-based transcoding method considering the importance of each object according to the present invention.

The method divides the input data including the image into objects according to the shape, size, contrast and connectivity of the objects in the image (step 100), and a predetermined importance index for each object for the divided objects. In operation 110, the importance index is encoded for each corresponding object (step 110), and the input index including the image is encoded by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain (120). step).

2 is a block diagram illustrating an example of a configuration of an object-based transcoding apparatus in consideration of importance of each object according to the present invention.

The apparatus includes an image data receiver 200 for receiving input data including an image, and an object divider 210 for dividing the received input data into objects according to the shape, size, contrast, and connectivity of an object included in the image. ), An index input unit 220 for receiving an index of importance for each object of the divided objects and encoding the corresponding index for each object, and inserting the received index of importance for each object as a watermark in a spatial domain and a frequency domain to correspond to each other. And a coder 230 that inserts an object and encodes the input data.

The coder 230 inserts an input importance index for each object into the object as a watermark in the spatial domain, and inserts a watermark in the frequency domain into the corresponding object by inserting the importance index for each object as a watermark in the frequency domain. A frequency domain watermark inserting unit 250 and an MPEG-4 processing unit 26 for encoding the object in which the watermark is inserted and outputting the MPEG-4 stream are included.

Image data is received through the image data receiver 200. It may be for encoding the image or video data using the present invention, or may be for the purpose of encoding and storing or transmitting.

The object dividing unit 210 divides the video / image data into which the watermark is to be inserted, by object unit using a semi-automatic or automatic image dividing technique (step 100). The technique of segmenting an image is to segment the image by object using the shape, size, contrast, and connectivity between objects in the image. When the image data input through the image data receiver 200 is data according to the MPEG format, dividing the MPEG format data into object units can be easily performed according to the MPEG standard.

The shape and number of objects to be divided can be selected automatically or arbitrarily by the user. According to a predetermined criterion for the divided object, the user automatically selects the importance of the object and assigns the importance index to each object, and the result is transmitted through the index input unit 220 (step 110).

Determining the importance is, for example, when a given image is divided into two objects, foreground and background, the importance index of A is usually used for the foreground because the foreground is more important than the background. Allocate an importance index named B. Of course, the A index is more important than the B index.

Or you might want to have the highest priority for an object that is a specific object (for example, a particular person or object), lower priority for another object, and lower priority for another object.

In this way, the sign 230 determines the importance index determined for each object in the object. At this time, the importance index is inserted as a watermark for preventing or tracking illegal copying. In other words, watermarking techniques are used.

In the present invention, the importance index for an object is inserted into the object by using a commonly used watermarking technique. As an example of the method, in order to be robust to the geometric attack on the object, the importance of the object in step 110 is referred to. An index is encoded for each corresponding object, and in step 120, the encoded importance index is repeatedly inserted into the object in units of blocks having a predetermined size in a spatial domain.

 A watermark sequence inserted into the object extracted by the user definition is generated in N * N block units. For example, when generating a watermark sequence of 32 * 32 blocks, one block of 1024 (32 * 32) lengths exists in the block. The block is divided into two 512 length slots by user keys, and then encoded and allocated to the first slot and a message related to the object importance index in the second slot.

The synchronization signal is used to predict the geometric transformation applied to the image, and uses a 512-length random variable consisting of -1 and 1. The message coding method uses an M-ary modulation method which is widely used in communication theory. For example, if the index of object importance to be inserted is A, the ASCII code for this is 65. Using this ASCII code as a seed, we generate a 512-length sequence and assign this value to that slot.

After the watermark encoding process, all binary codes (-1, 1) are allocated to 32 * 32 blocks. In order to consider the geometric manipulation of the object, the N * N watermark block generated in encoding the message is repeatedly inserted. If FIG. 3A is an extracted object, FIG. 3B shows a state in which N * N watermark blocks are repeatedly inserted.

When I (x, y) is a brightness value in the extracted object, an example of the insertion method is shown in Equation 1 below.

At this time Is a pixel unit weight at a position obtained by using a human visual system (HVS) for an image to strongly insert a watermark. Denotes a value at (x, y) in the watermark sequence generated during message encoding.

Extracting the inserted watermark as described above will be described again below.

In the present invention, another method of inserting an importance index of an object into the object as a watermark may include a watermark in consideration of general attacks such as real-time watermark insertion and detection or noise addition, without considering geometric attacks on the object. You can use the insert method.

To this end, in step 110, the importance index of the object is encoded for each corresponding object, and when the input data is video / video data, the compression is performed in the format of MPEG-4 or JPEG2000 when encoding in step 120. In this case, it is preferable to insert the encoded importance index using coefficients of the DCT block in each object.

Alternatively, when the input data is bitstream data, in step 120, it is preferable to obtain only coefficients of the DCT block by decoding only the object to which the importance index is to be inserted, and insert the encoded importance index using the coefficients.

The spatial domain watermark inserter 240 and the frequency domain watermark inserter 250 insert the importance index as a watermark in the spatial domain and the frequency domain, and insert the index into each corresponding object. The MPEG-4 processor 260 encodes the objects including the importance index in a format defined by the MPEG-4 standard (120). At this time, encoding in the MPEG-4 format is one example, and the gist of the present invention is retained even if other object compression techniques such as JPEG2000 are used.

As described above, when compressed or encoded video data or bitstream video data is transmitted in a network environment different from the existing channel capacity, or moved to a storage medium that is limited than the existing storage capacity, a desired bit rate for different environments It may be necessary to recompress at a compression rate suitable for furnaces or other environments. In this case, it is inefficient to apply the same bit rate to each object because the importance of each object is different. Therefore, bit rate adjustment according to the importance of each object is a necessary process. A method for satisfying these requirements is described.

4 illustrates a flow of an object-based transcoding method in consideration of importance of each object according to the present invention.

In this method of transcoding encoded image data having an importance index inserted as a watermark for each object, the image data is decoded for each object according to the type of data (step 400), and the decoded objects are separated for each object (step 400). Step 410), extracting the importance index for each separated object (step 420), assigning a predetermined bit rate for each object according to the importance index for each object (step 430), and transcoding for each object according to the bit rate allocated for each object Encoding is performed (step 440).

5 is a block diagram showing the configuration of an object-based transcoding device considering the importance of each object according to the present invention.

The apparatus for transcoding video data having an importance index inserted as a watermark for each object includes: an object-based decoder 400 for decoding the video data for each object according to the type of data, and separating the decoded objects for each object. The object-separation unit 510, an importance index extractor 520 for extracting an importance index by considering the importance index for each separated object as a predetermined noise, and a predetermined bit rate according to the extracted importance index for each object. A bit rate allocator 530 is allocated and a transcoding unit 540 is transcoded and encoded for each object according to the bit rate allocated to each object.

The object decoder 500 decodes the bitstream or the image data which is a result of encoding in the same place as the MPEG-4 processor 260 of FIG. 2. At this time, decoding is performed according to the type of the bitstream or the image data, that is, the encoded method. This may be determined by looking at the extension specified in the image data or the bitstream file, or may decode the bitstream encoded according to a predetermined format.

If the input data may not be coded at all, in this case, the object-based decoding unit 500 transfers the data to the next step without performing separate decoding.

The data processed by the object-based decoder 500 separates the objects from one object to another by using a semi-automatic or automatic image segmentation technique by the user in the object-specific separator 510 (step 410).

The importance index extractor 520 extracts the importance index of the object inserted into the object from the separated objects (step 420). The data input to the object-based decoder 500 extracts the inserted watermark since the importance index is inserted as the watermark.

In this case, the reference numeral 420 may include a block having a predetermined size in which the importance index is encoded to insert the watermark for strengthening the geometric attack on the object described in the process of inserting the watermark at 120. In order to extract the watermark repeatedly inserted into the object in units, predicting the inserted importance index as a kind of noise and extracting the importance index based on the prediction, the image data using the predicted importance index Finding a geometric transformation applied to the image; reconstructing the original image data by inversely transforming the image data based on the geometric transformation applied to the image; and decoding the extracted importance index.

When extracting a watermark directly from an image with a watermark embedded or attacked, it is difficult to extract an accurate watermark because the power of the image is much larger than that of the watermark. Therefore, the watermark is regarded as a kind of noise, for example, Gaussian noise, and the watermark is predicted using a commonly used noise canceling filter, such as a Wiener filter. In other words, the watermark is regarded as a predetermined noise and the watermark is predicted using a filter that removes the noise.

The following equation is an example of one of the filter coefficients that can be used in the present invention.

The geometric transformation applied to the image may be predicted using the auto-correlation value in the predicted watermark image. The original image may be reconstructed by inversely transforming the image based on the predicted geometric transformation. Modeling of geometric transformations may include affine transformations using six parameters, or pulse-perspective transformations using eight or twelve parameters.

The object is extracted from the reconstructed image, and the inserted message is extracted using the M-ary modulation method from the extracted object.

The bit rate assignment unit 530 allocates a bit rate suitable for the importance of the object for each object by using the extracted information. This bit rate may be determined based on the criteria as described in FIG. 1 or FIG. 2. Alternatively, when the data inputted to the object-decoding unit 500 is ultimately directed, for example, whether the maximum compression should be performed due to the lack of storage space or if the bandwidth of the channel to be transmitted is determined, what compression rate is determined according to the bandwidth. What should be applied will be determined. In this case, the bit rate may be determined to uniformly compress all objects uniformly, and of course, the bit rate may be separately determined for each object, and the bit rate may be determined in conjunction with the importance index.

The transcoding unit 540 transcodes and encodes each object according to the bit rate determined as described above (step 440). This coded data may be transmitted over a network or stored on another storage medium, depending on the aspect in which it is used.

When the image data input to the object-based decoder 500 is multimedia data, the object-specific separator 510 separates the image data into video data and audio data in step 410, and separates the separated video data into objects. In addition, the transcoding unit 540 configures the transcoded objects into a video frame after step 440, and synthesizes the separated audio component in the video frame.

This process is explained in detail.

First, when the input is multimedia data, a video signal is divided into a video bitstream and an audio bitstream using a demultiplexer (step 410). The importance index for each object is extracted from the divided video bitstreams using a watermarking technique (step 420). The bit rate corresponding to each object is allocated using the extracted importance index for each object and the size and temporal importance of the objects in the image (step 430). As a result of bit rate allocation, more important bits are allocated for objects that are not important. Transcoding for each object is performed based on the allocated bit rate as described above.

Examples of the transcoding techniques used at this time are as follows.

The first is to delete the AC coefficient below the specified value among the AC components in the DCT block in the object. In this way, it is possible to convert the bit rate with a simple operation.

In the second case, a bit rate conversion method through requantization is performed through VLD, inverse quantization of a DCT block, and then requantized and VLC to achieve an accurate bit rate.

The transcoded video bitstream and audio bitstream will be synthesized using a multiplexer, transmitted over the network as needed, or stored in a suitable storage medium.

6 is a block diagram illustrating an example of a configuration of an apparatus for transcoding by inserting a importance index and assigning a bit rate to each object of image data according to the present invention. As described above, the function of the present invention receives image data, divides it by object, inserts and decodes the importance index, splits the result by object, inserts the importance index, and assigns a bit rate after transcoding. It is. 6 shows an example of combining these functions and implementing them in one device.

The image data receiver 600 receives image data, and the object splitter 610 divides the image data for each object. When importance is input for each object through the importance index unit 620, the importance index inserter 630 inserts the importance index into each object as a watermark. When the bit rate assignment unit 640 determines an appropriate bit rate for each object, the transcoding unit 650 transcodes each object into which each watermark is inserted according to the bit rate.

The function of each component of the apparatus of FIG. 6 is substantially the same as each component shown in FIGS. 2 and 5, and thus detailed description of each component is omitted.

Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. Examples included in the above description are introduced for the understanding of the present invention, and these examples do not limit the spirit and scope of the present invention. It will be apparent to those skilled in the art that various embodiments in accordance with the present invention in addition to the above examples are possible. The scope of the present invention is shown not in the above description but in the claims, and all differences within the scope will be construed as being included in the present invention.

In addition, it can be easily understood by those skilled in the art that each of the above steps according to the present invention can be variously implemented in software or hardware using a general programming technique.

And some steps of the invention may also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, CD-RW, magnetic tape, floppy disks, HDDs, optical disks, magneto-optical storage devices, and carrier wave (eg, Internet It also includes the implementation in the form of). The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

According to the present invention, input data including an image is divided into objects according to the shape, size, contrast, and connectivity of an object in the image, and a predetermined importance index is determined for each object for the divided objects. And encoding the input data including the image by inserting the importance index for each object as a watermark in a corresponding object in the spatial domain or the frequency domain, and inserting a value according to the importance of the object as a watermark and associating it with the watermark. By transcoding by assigning different bit rate for each object, it is possible to insert the importance of each object into the object and to perform compression according to the object for each object, so that it is possible to transmit video and video data effectively. Clearer display than objects and unique watermarks prevent illegal duplication The distribution of copyrighted works through the communication network is effective.

The embedded watermark can be used for copyright protection and authentication of video and video, and can also determine the integrity of data transmitted over a network. In addition, when transcoding is required, watermarks may be extracted to determine importance of each object, and according to the determination, video and image data may be transcoded and transmitted by applying different bit rates for each object. In addition, the method according to the present invention solves the problem of allocating the same bit rate to all objects, which is a conventional method. It can be allocated to transmit video and video data more effectively.

1 illustrates a flow of an object-based transcoding method considering the importance of each object according to the present invention.

2 is a block diagram illustrating an example of a configuration of an object-based transcoding apparatus in consideration of importance of each object according to the present invention.

3A is an example of an object extracted from image data.

FIG. 3B illustrates a state in which N * N watermark blocks are repeatedly inserted into the object of FIG. 3A.

4 illustrates a flow of an object-based transcoding method in consideration of importance of each object according to the present invention.

5 is a block diagram showing the configuration of an object-based transcoding device considering the importance of each object according to the present invention.

6 is a block diagram illustrating an example of a configuration of an apparatus for transcoding by inserting a importance index and assigning a bit rate to each object of image data according to the present invention.

Claims (12)

delete (a) dividing the input data including the image into objects according to the shape, size, contrast and connectivity of the objects in the image; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark; In the step (c), in the spatial domain, the object-based transcoding method considering the importance of each object, characterized in that repeatedly inserting the encoded importance index in the unit of a block of a predetermined size. (a) dividing the input data including the image into objects according to the shape, size, contrast and connectivity of the objects in the image; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark; When the input data is video / video data, the encoding is performed in the format of MPEG-4 or JPEG2000 when encoding in the step (c), wherein the encoded importance index is obtained by using coefficients of the DCT block in each object. Object-based transcoding method considering the importance of each object to insert. (a) dividing the input data including the image into objects according to the shape, size, contrast and connectivity of the objects in the image; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark; When the input data is bitstream data, in step (c), decode only the object to which the importance index is to be inserted to obtain a coefficient of the DCT block, and insert the encoded importance index using the coefficient. Object-based transcoding method considering star importance. A method of transcoding encoded image data having an importance index inserted as a watermark for each object, (a) decoding the image data for each object according to the type of data; (b) separating the decrypted objects for each object; (c) extracting an importance index for each separated object; (d) allocating a predetermined bit rate for each object according to the importance index for each object; And (e) transcoding by encoding the object according to the object bit rate according to the bit rate allocated for each object. The method of claim 5, In step (c), (c1) predicting the inserted importance index as a kind of noise and extracting the importance index based on the prediction; (c2) finding a geometric transformation applied to the image data using the predicted importance index; (c3) restoring original image data by inversely transforming image data based on a geometric transformation applied to the image; And (c4) object-based transcoding method considering importance of each object, comprising: decoding the extracted importance index. The method of claim 5, When the input image data is multimedia data, In step (b), the image data is divided into video and audio data, and the separated video data is classified by object. (f) composing the transcoded objects into video frames, and synthesizing the separated audio components into the video frames. An image data receiver configured to receive input data including an image; An object dividing unit dividing the received input data into objects according to a shape, size, contrast, and connectivity of an object included in an image; An index input unit which receives an importance index for each object of the divided objects and encodes the importance index for each corresponding object; And And a coder for inserting the received importance index for each object into a corresponding object by inserting a watermark in a spatial domain and a frequency domain to encode the input data. The coder repeatedly inserts the encoded importance index into an object in units of blocks of a predetermined size in a spatial domain, or compresses it in the format of MPEG-4 or JPEG2000 when encoding the input data when the input data is video / video data. In this case, the coefficient of the DCT block is obtained by inserting the encoded importance index using coefficients of the DCT block in each object, or decoding only the object to which the importance index is to be inserted when the input data is bitstream data. An object-based transcoding apparatus considering the importance of each object, characterized by inserting the encoded importance index by using a coefficient. An apparatus for transcoding video data having an importance index inserted as a watermark for each object, An object-based decoder which decodes the image data for each object according to a type of data; An object separator that separates the decoded objects by objects; An importance index extractor configured to extract the importance index by considering the importance index as the predetermined noise for each of the separated objects; A bit rate allocation unit for allocating a predetermined bit rate for each object according to the extracted importance index; And An object-based transcoding apparatus in consideration of the importance of each object, characterized in that it comprises a transcoding unit for transcoding by encoding according to the object according to the bit rate allocated to each object. The method of claim 9, wherein the importance index extractor predicts the inserted importance index as a kind of noise, extracts an importance index based on the prediction, and uses the predicted importance index to perform a geometric transformation applied to the image data. And an object-based transcoding apparatus considering object-specific importance, reconstructing the original image data by inversely transforming the image data based on a geometric transformation applied to the image, and decoding the extracted importance index. (a) dividing the input data including the image into objects according to the shape, size, contrast and connectivity of the objects in the image; (b) determining a predetermined importance index for each of the divided objects and encoding the importance index for each corresponding object; And (c) encoding the input data including the image by inserting the importance index for each object into a corresponding object in a spatial domain or a frequency domain as a watermark; In step (c), the encoded importance index is repeatedly inserted into an object in blocks of a predetermined size in a spatial domain. When the input data is video / video data, the encoding is performed in the format of MPEG-4 or JPEG2000 when encoding in the step (c), wherein the encoded importance index is obtained by using coefficients of the DCT block in each object. Insert or When the input data is bitstream data, in step (c), decode only the object to which the importance index is to be inserted to obtain a coefficient of the DCT block, and insert the encoded importance index using the coefficient. A computer-readable recording medium that records a program of an object-based transcoding method that takes into account the importance of each. In a recording medium that records a method of transcoding video data having an importance index inserted as a watermark for each object, (a) decoding the image data for each object according to the type of data; (b) separating the decrypted objects for each object; (c) extracting an importance index for each separated object; (d) allocating a predetermined bit rate for each object according to the importance index for each object; And and (e) transcoding and encoding each object according to the bit rate allocated to each object, wherein the object-based transcoding program is recorded.