JP4584115B2 - Image encoding apparatus, image encoding method, image encoding program, and computer-readable recording medium recording the image encoding program - Google Patents

Description

  The present invention relates to still image and moving image encoding and decoding processing, and relates to an image encoding device, an image encoding method, an image encoding program, and a computer-readable recording medium on which an image encoding program is recorded.

  In recent years, with the development of IT technology, it has become possible to handle digital image data in various devices such as televisions, personal computers, digital color copiers, digital cameras, digital video cameras, etc., and the frequency of use has increased. In addition, in order to increase the efficiency of image data exchange between devices, increase the speed of internal processing, and increase the efficiency of storage, image encoding technology, which is a technology for reducing the volume of digital image data, has become important. Image coding techniques can be divided into a lossy image coding system with image quality degradation and a lossless image coding system without image quality degradation.

In recent years, a technique for improving the image quality by encoding the entire image data under different conditions depending on the region is used instead of encoding the entire image data under a certain condition. This is mainly used to encode a region that is not good in the lossy encoding method using the lossless encoding method to prevent image quality deterioration. As a technique for encoding image data under different conditions depending on regions, there is Patent Document 1.
Japanese Patent Application Laid-Open No. 11-339002 is a technique in which image data is separated into a plurality of layers and each layer is encoded under different conditions. The layer used for each region is switched. In particular, by using a transparent color for the upper layer, the layer to be used is switched, and in addition to the data obtained by encoding the upper layer by the irreversible encoding method, the transparent color data is also included for designating the transparent color region.

  In general, the lossless image encoding method has a lower compression rate than the lossy image encoding method and does not involve image quality deterioration, and thus the capacity after encoding cannot be reduced so much. When the upper layer image in which the specific pixel value is designated as a transparent color is encoded by the lossless encoding method, there is a problem that the capacity of the encoded data of the upper layer image is enlarged. As a technique for avoiding this problem, there is a technique for encoding an upper layer image using an irreversible image encoding method as in Patent Document 1. However, since the encoding of the upper layer image is accompanied by image quality degradation, the specific color of the upper layer cannot simply be a transparent color. This is because the transparent color pixel value loses its function as a transparent color even if it is a very close value. Therefore, in Patent Document 1, it is necessary to separately hold transparent color data that defines the upper transparent color region.

  Eventually, this transparent color data is required in addition to the image data of the layer, and this transparent color data accurately conveys the transparent color area to the decoding side. Since the compression rate is small, the capacity is compressed. Further, when decoding, compared with a method that does not hold transparent color data, a decoding time for the transparent color data is required. This hinders comfortable use particularly in a portable terminal with a slow processing speed such as a cellular phone or a portable information terminal (Personal Digital Assistance).

  The present invention provides an image encoding method that enables an upper layer image to be encoded by an irreversible encoding method in a state in which a transparent color region can be specified, and eliminates the need for transparent color data. An object of the present invention is to provide an image encoding device using an encoding method, an image encoding program, and a computer-readable recording medium on which the image encoding program is recorded.

  Most lossy encoding methods can control the degree of image quality degradation by setting parameters during encoding. If the deterioration is set to be small, the compression rate is naturally reduced, but still a high compression rate can be ensured as compared with the lossless image encoding method. By setting the parameters at the time of encoding, it is possible to ensure a high compression rate while maintaining the required image quality. This realizes a high compression rate and enables high-speed decoding processing.

  According to one aspect of the present invention, input image data is separated into layer image data of N layers (N is a natural number of 2 or more) corresponding to the input image data, and at least one layer image data is small. An image encoding apparatus that performs encoding using an image encoding method of an irreversible encoding method that performs function conversion processing in units of regions, wherein n-th layer image data (n is 1 or more, A natural number less than or equal to N. The smaller the n, the higher the layer when the composite process is performed.) Area information data input processing means for inputting area information data that can specify at least the effective area to be expressed; Based on the above, when layer image data to be encoded by the image encoding method is encoded by the image encoding method and the region using the pixel value of the input image data in units of small regions and decoded When the pixel value pattern before encoding of the flat area where the pixel values in the small area are the same flat value is used as the flattening conversion pattern, the area to be encoded using the flattening conversion pattern is determined as the area. And a layer separation processing unit that generates n-th layer image data, and an image code that encodes the n-th layer image data created by the layer separation processing unit using an image coding method to create n-th layer image coded data. Data generation processing means and output data generation processing means for generating output data including at least the n-th layer image encoded data generated by the image encoded data generation processing means.

  Preferably, the image encoding device further includes resolution conversion processing means for converting the vertical resolution and the horizontal resolution of the image data, and the small area has two straight lines parallel to the vertical direction and two parallel to the horizontal direction. The layer separation processing means determines the area of the n-th layer image data, determines the size of the block in the vertical direction as 1 / Xn (Xn is a real number, and the n-th layer The resolution of the image data in the vertical direction) and the size in the horizontal direction are 1 / Yn (Yn is a real number, the resolution in the horizontal direction of the n-th layer image data) times, and each is determined in units of regions, and the image code The converted data generation processing means converts the area using the input image data by the resolution conversion processing means to convert the vertical resolution of the corresponding area of the input image data to Xn times and the horizontal resolution to Yn times. image data Using, to produce the n-th layer image data.

  Preferably, the small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction, and the layer separation processing means is used for determining the area of the n-th layer image data. An area is represented by a common multiple of the vertical size and a common multiple of the horizontal size of each block of the image data of each layer image data encoded by the image encoding method corresponding to each block of the nth layer image data, respectively. Judged in units of blocks.

  Preferably, the small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction, and the layer separation processing means is used for determining the area of the n-th layer image data. The least common multiple of the vertical size and the least common multiple of the horizontal size of each block of the image data of each layer image data that is encoded by the image encoding method, corresponding to each block of the nth layer image data, respectively. And is determined in units of blocks with horizontal dimensions.

  Preferably, the image encoding device further includes resolution conversion processing means for converting the vertical resolution and the horizontal resolution of the image data, and the small area has two straight lines parallel to the vertical direction and two parallel to the horizontal direction. The layer separation processing means uses an image encoding method corresponding to each block of the n-th layer image data. 1 / Xi (Xi is a real number and the vertical resolution of the i-th layer image data. 1 ≦ i ≦ N and i includes n) Not a natural number.) A common multiple of the multiplied size, and a common multiple of the size obtained by multiplying the horizontal size of the block by 1 / Yi (Yi is a real number and the horizontal resolution of the i-th layer image data). Are vertical and horizontal sizes respectively. The encoded image data creation processing means determines the vertical resolution of the corresponding area of the input image data by Xn (Xn is a real number) by the resolution conversion processing means for the area using the input image data. , The resolution in the vertical direction of the n-th layer image data) times and the resolution in the horizontal direction to Yn (Yn is a real number and the horizontal resolution of the n-th layer image data) times, respectively. The nth layer image data is generated.

  Preferably, the image encoding device further includes resolution conversion processing means for converting the vertical resolution and the horizontal resolution of the image data, and the small area has two straight lines parallel to the vertical direction and two parallel to the horizontal direction. The layer separation processing means uses an image encoding method corresponding to each block of the n-th layer image data. 1 / Xi (Xi is a real number and the vertical resolution of the i-th layer image data. 1 ≦ i ≦ N and i includes n) No natural number.) The least common multiple of the multiplied size and the minimum size of the horizontal size of the block multiplied by 1 / Yi (Yi is a real number and the horizontal resolution of the i-th layer image data) times. Common multiple, the vertical and horizontal sizes, respectively The encoded image data creation processing means determines the vertical resolution of the corresponding area of the input image data by Xn (Xn is a real number) by the resolution conversion processing means for the area using the input image data. , The resolution in the vertical direction of the n-th layer image data) times and the resolution in the horizontal direction to Yn (Yn is a real number and the horizontal resolution of the n-th layer image data) times, respectively. The nth layer image data is generated.

  Preferably, the region of the (n + 1) -th layer image data corresponding to the region using the pixel value of the input image data in the n-th layer image data is decoded in the region of the (n + 1) -th layer image data encoded by the image encoding method. A flattening conversion pattern having a flat value is used.

  Preferably, the layer image data is an irreversible encoding method that performs function conversion processing in units of small regions, and is different from the image encoding method when the small image is encoded and decoded by the second image encoding method. An area using pixel values of input image data in n-th layer image data when a pixel value pattern before encoding of a flat area in which the pixel values in the area are the same flat value is used as the second flattening conversion pattern And the second flattening conversion pattern is used in the area of the (n + 1) -th layer image data that is encoded by the second image encoding method.

  Preferably, the flat value is a value used as a transparent color in decoding of the n-th layer image data.

Preferably, the output data creation processing means further includes a transparent color value in the output data.
According to another aspect of the present invention, a computer having an arithmetic processing unit separates the input image data into layer image data of N layers (N is a natural number of 2 or more) respectively corresponding to the input image data, An image encoding method in which at least one layer image data is encoded by an irreversible encoding image encoding method in which function conversion processing is performed in units of small regions. It is possible to specify at least the effective area to be expressed by the n-th layer image data (n is a natural number of 1 or more and N or less. The smaller n is, the higher the layer is arranged in the upper layer). The step of inputting the region information data, and the layer separation processing means, on the basis of the region information data, for the layer image data to be encoded by the image encoding method, A flattened conversion pattern is a pixel value pattern before encoding of an area using a prime value and a flat area where the pixel value in the small area becomes the same flat value when encoded and decoded by the image encoding method. When determining the region to be encoded using the flattening conversion pattern, the step of generating the n-th layer image data and the image encoded data creation processing means are created by the layer separation processing means A step of encoding the n-th layer image data by an image encoding method to generate the n-th layer image encoded data; and an output data generation processing unit configured by the n-th layer image code generated by the image encoded data generation processing unit. Generating output data including at least the digitized data.

  According to still another aspect of the present invention, the input image data is separated into layer image data of N layers (N is a natural number of 2 or more) corresponding to the input image data, of which at least one layer image data is An image encoding program for causing a computer having an arithmetic processing unit to execute an image encoding method for encoding by an irreversible encoding type image encoding method for performing function conversion processing in units of small regions. However, the effective area to be expressed by the n-th layer image data (n is a natural number of 1 or more and N or less. When n is reduced, the effective region to be expressed in the upper layer) is individual layer image data. At least the step of inputting identifiable region information data, and the processing device, on the basis of the region information data, the layer image data that is encoded by the image encoding method in units of small regions, The pixel value pattern before encoding of the area using the pixel value of the input image data and the flat area where the pixel value in the small area becomes the same flat value when encoded and decoded by the image encoding method. When a flattening conversion pattern is used, a step of determining an area to be encoded using the flattening conversion pattern and generating n-th layer image data; Are encoded by an image encoding method to generate n-th layer image encoded data, and an arithmetic processing unit generates output data including at least the generated n-th layer image encoded data. Let it run.

  According to still another aspect of the present invention, there is provided a computer-readable recording medium on which an image encoding program is recorded. The image encoding program has N pieces (N is a natural number of 2 or more) corresponding to input image data. Input image data is separated into layer image data of layers, and at least one of the layer image data is encoded by an irreversible encoding image encoding method in which function conversion processing is performed in units of small regions An image encoding program for causing a computer having an arithmetic processing unit to execute the method, wherein the arithmetic processing unit is n-th layer image data (n is a natural number of 1 or more and N or less. A step of inputting region information data that can identify at least an effective region to be expressed in the case where the composition processing is performed as the size of the processing is smaller. The layer image data that is encoded by the image encoding method based on the region information data is encoded by the region using the pixel value of the input image data in units of small regions and the decoding process. When the pixel value pattern before encoding of the flat region in which the pixel values in the small region become the same flat value when the flattening conversion pattern is used as the flattening conversion pattern, the region to be encoded using the flattening conversion pattern Determining the region and generating the n-th layer image data, and the arithmetic processing device encoding the generated n-th layer image data by an image encoding method to generate n-th layer image encoded data; And an arithmetic processing unit causing a computer to execute output data including at least the generated n-th layer image encoded data.

  According to the present invention, even in an irreversible encoding method, an irreversible signal having a flattening conversion pattern, which is a pixel value pattern in which all pixel values in a region when encoded and decoded become flat values having the same value. In the case of the encoding method, with respect to the area of the flattening conversion pattern, it is guaranteed that the image after encoding becomes a flat area having a flat value determined by a parameter at the time of encoding. Therefore, if an image of another layer is used for the area having the flat value, an image corresponding to the input image at the time of encoding can be reproduced.

  Also, if the flat value is designated as a transparent color on the decoder side, if the transparent area is filled with a flattening conversion pattern, the transparent color is surely obtained at the time of decoding. That is, it is possible to reliably specify the transparent color determined by the parameters at the time of encoding in units of small areas that perform function conversion of the irreversible area method having the flattening conversion pattern.

  By using an irreversible encoding method for image data of a layer including a transparent color, an encoding process can be performed so as to ensure that a specific area has a transparent color at the time of decoding. According to the image encoding method of the present invention or the image encoding apparatus using this image encoding method, it is possible to encode the upper layer image by the irreversible encoding method in a state where the transparent color region can be specified. Transparent color data is unnecessary. This realizes a high compression rate and enables high-speed decoding processing.

  First, terms used in the specification and claims of the present invention are defined. The transformation of input image data by a function including non-orthogonal transformation such as non-orthogonal wavelet as well as orthogonal transformation represented by DCT (discrete cosine transformation), DFT (discrete Fourier transformation) and orthogonal wavelet transformation is called “function transformation”. Define. Further, an area having the same pixel value is defined as a “flat area”, and a pixel value in the flat area is defined as a “flat value”. To describe in a specific example, a region where the pixel values in the region are all 0 is “a flat region with a flat value of 0”, a region where all the pixel values in the region are 255 “a flat region with a flat value of 255”, A region having all values N is referred to as a “flat region having a flat value N”.

  Further, in the lossy image encoding method in which function conversion processing is performed in units of regions, 1 obtained by encoding the pixel values of the one region under at least one specific parameter condition and decoding the encoded data. When one decoding region becomes a flat region, the pixel value pattern of the region before encoding is defined as a “flattening conversion pattern”. In other words, the “flattening conversion pattern” does not have to be a flat area before encoding, but after encoding under specific parameter conditions, decoding the encoded data results in a flat area before encoding. It is an input area. Furthermore, an irreversible image encoding method in which a flattened conversion pattern exists is defined as “an irreversible image encoding method having a flattened conversion pattern”.

  In addition, in the lossy image encoding method in which function conversion processing is performed in units of regions, one decoding region obtained by encoding the pixel value of the one region under an arbitrary parameter condition and decoding the encoded data Is a flat region, the pixel value pattern of the region before encoding is defined as a “constant flattening conversion pattern”. In other words, the “constant flattening conversion pattern” does not have to be a flat area before encoding, but is independent of the parameter conditions at the time of encoding, that is, after encoding under arbitrary parameter conditions, This is an input area before encoding that becomes a flat area when the converted data is decoded. Furthermore, an irreversible image encoding method in which a constant flattening conversion pattern exists is defined as “an irreversible image encoding method having a constant flattening conversion pattern”.

  In addition, the flattening conversion pattern itself is a flat region before encoding, and the flattening conversion pattern is such that the flat value when encoded and decoded under the condition of specific parameters matches the flat value before the encoding. Is defined as “immobilization flattening conversion pattern”. In other words, the “non-moving flattening conversion pattern” is a flat region before encoding, and after encoding under specific parameter conditions, decoding the encoded data results in a flat value that is the same as the flat value before encoding. This is an input area before encoding, which is a flat area. That is, when encoding is performed under specific parameter conditions, the region does not change due to encoding. Furthermore, an irreversible image encoding method in which a non-moving flattening conversion pattern exists is defined as “an irreversible image encoding method having a non-moving flattening conversion pattern”.

  In addition, the flattening conversion pattern itself is a flat region before encoding, and the flattening conversion pattern is such that the flat value when encoded and decoded under an arbitrary parameter condition matches the flat value before encoding. Is defined as “constant immobilization flattening conversion pattern”. In other words, the “constant immovable flattening conversion pattern” is a flat area before encoding, and is encoded data after encoding under any parameter condition, regardless of the parameter conditions at the time of encoding. Is an input area before encoding that becomes a flat area having the same flat value as the flat value before encoding. That is, when encoding is performed under arbitrary parameter conditions, the region does not change due to encoding. Furthermore, an irreversible image encoding method in which a constant immobilization flattening conversion pattern exists is defined as “an irreversible image encoding method having a constant immobilization flattening conversion pattern”.

  Specific examples of each flattening conversion pattern based on the above definition are shown in FIGS. Each of them uses JPEG, and specifically, an experiment was performed in which encoding was performed using cjpeg attached to IJG libjpeg and decoding was performed using djpeg. A block of 8 pixels × 8 pixels (hereinafter referred to as 8 × 8 pixels), which is a JPEG processing unit, is used as a region unit. FIG. 1 illustrates a flattening conversion pattern. FIG. 1A shows an area before encoding, which is an area composed of pixel values of 0 and 1, and is not a flat area. When the quantization parameter Q is JPEG-encoded with a specific 75 and decoded, a flat area (b) composed of pixel values 0 is obtained. However, when the specified quantization parameter Q = 75 is JPEG-encoded and decoded with a quantization parameter of Q = 100, for example, a region (c) composed of pixel values of 0 and 1 is obtained and does not become a flat region. In this case, since the area (a) before encoding becomes a flat area when encoded and decoded with the specified quantization parameter Q = 75, it can be seen that it is a flattening conversion pattern.

  FIG. 2 illustrates a constant flattening conversion pattern. FIG. 2A shows a region before encoding. FIG. 2A is a region having a configuration in which a pixel value 0 occupies most but a part of the pixel value 1 is also included. Not an area. When JPEG encoding and decoding is performed with an arbitrary quantization parameter (Q = 0 to 100), a flat area (b) composed of pixel values 0 is obtained. In this case, since the area before encoding becomes a flat area when encoded and decoded with an arbitrary quantization parameter, it can be seen that it is a constant flattening conversion pattern.

  FIG. 3 is a diagram for explaining an immovable flattening conversion pattern. FIG. 3A shows a region before encoding, and is a flat region composed of flat values of pixel values 129. When the quantization parameter Q is JPEG-encoded with a specific 75 and decoded, a flat area (b) composed of flat values of pixel values 129 is obtained. This is because the flat value matches the flat value before encoding, and the area does not change due to encoding. However, when the specified quantization parameter Q = 75 is JPEG-encoded and decoded with a quantization parameter of Q = 0, for example, a flat region (c) composed of flat values of pixel values 128 is obtained. This is a flat region both before and after encoding, but the flat value 128 after encoding and decoding changes from the flat value 129 before encoding. In this case, the flat value when decoded with the specified quantization parameter Q = 75 is the same as the flat value before encoding, that is, the region is not changed by encoding. Thus, it can be seen that this is a stationary flattening conversion pattern.

  FIG. 4 is a diagram for explaining a constant immobilization flattening conversion pattern. (A) showing a region before encoding is a flat region in which pixel values are all 0 (flat value 0), and an arbitrary quantization parameter (Q = Even if JPEG encoding and decoding is performed at 0 to 100), the area remains the same as the flat area (b) where the pixel values are all 0 (flat value 0). Therefore, in this case, it can be seen that this is a constant immobilization flattening conversion pattern.

  As is clear from the above definition, there is an inclusion relationship shown in FIG. 5 between the area data and each flattening conversion pattern, and there is an inclusion relationship shown in FIG. 6 between each lossy encoding method. For example, the JPEG image encoding method and the MPEG image encoding method are irreversible image encoding methods in which encoding is performed using discrete cosine transform that is function conversion in units of 8 × 8 pixel regions. The flattening conversion patterns of the JPEG image encoding method and the MPEG image encoding method are all those in which all the 8 × 8 pixel regions have the same value (that is, the flat region). This is because there is no change in the pixel value in the 8 × 8 pixel region because the AC component after the discrete cosine transform is 0, and the flat region is obtained after encoding. In the constant flattening conversion pattern of the JPEG image encoding method and the MPEG image encoding method, all pixel values in the 8 × 8 pixel region are 0 (that is, a flat region having a flat value of 0). This is because when the DC component is 0 in the flattening conversion pattern, it is not affected by quantization. (In other flattening conversion patterns, if the setting of the quantization width (parameter condition at the time of encoding) is different, the flat value after encoding may change, so it cannot be said to be a constant flattening conversion pattern.) In addition, the constant flattening conversion pattern in which all pixel values in the 8 × 8 pixel area are 0 is converted into a flat area having a flat value of 0 after encoding and decoding. It can be said that there is. From the above, the JPEG image encoding method and the MPEG image encoding method are “irreversible image encoding methods having a constant stationary flattening conversion pattern”.

  The technical points of the present invention will be described. By defining the specific pixel value as a “transparent color”, which is a pixel value that makes the image transparent, the image data can be expressed by being separated into a plurality of layers. At the time of decoding, the pixel value of the lower layer is used for display by setting the specific pixel value of the upper layer to a transparent color. Thus, the layer displayed effectively is decided by using a transparent color. When the input image data is separated into multiple layers of image data, and some or all layers are encoded, and the layer image data to be encoded contains a transparent color, the specified pixel value Is assigned a special effect of “transparency”, which basically has a completely different meaning from other pixel values. Even if the value of the specific pixel value is changed by encoding, even if it is very close, the targeted effect cannot be obtained at all. Therefore, conventionally, when encoding image data of a layer including a transparent color, a lossless encoding method has been used.

  The present invention is a flattening that is a pixel value pattern before encoding a flat region in which all pixel values of the region when encoded and decoded are flat values even when the lossy encoding method is used. If it is an irreversible encoding method having a conversion pattern, with respect to the flattened conversion pattern area, it is guaranteed that the image after encoding will be a flat area having a flat value determined by parameters at the time of encoding. ing. Therefore, if the flat value is designated as a transparent color on the decoder side (of course, a transparent color pixel value may be included in the header or auxiliary information of the encoded data and passed to the decoder side), Filling the transparent area with a flattening conversion pattern ensures a transparent color during decoding. That is, it is possible to reliably specify the transparent color determined by the parameters at the time of encoding in units of small areas that perform function conversion of the irreversible area method having the flattening conversion pattern.

  In addition, in the case of an irreversible image encoding method having a constant flattening conversion pattern that becomes a flattening conversion pattern when encoded and decoded regardless of the parameter conditions at the time of encoding, in a small area unit for performing function conversion Even if the parameters at the time of encoding are set arbitrarily, it is possible to reliably specify the transparent color.

  In addition, the flattening conversion pattern itself is a flat region before encoding, and is encoded under specific parameter conditions. The flat value when the decoding is the same as the flat value before the encoding is the same. If the irreversible encoding method has a non-reversible encoding method, there is an advantage that the non-moving flattening conversion pattern itself can be a transparent color determined by parameters at the time of encoding.

  In addition, the flattening conversion pattern itself is a flat region before encoding, and is encoded under an arbitrary parameter condition, and the flat value when decoded is the same as the flat value before encoding. An irreversible encoding method having a pattern has an advantage that even if the parameters for encoding the constant immobilization flattening conversion pattern itself are arbitrarily set, a transparent color can be obtained.

  The basic technology common to all the embodiments is to use an irreversible encoding method for image data of a layer including a transparent color, and to control and encode an area that is desired to be transparent at the time of decoding to ensure a transparent color at the time of decoding. This is a technology for performing the conversion process. The image encoding apparatus of the present invention can encode the upper layer image by the irreversible encoding method in a state where the transparent color region can be specified, and does not require the transparent color data. This realizes a high compression rate and enables high-speed decoding processing.

[Embodiment 1]
The image encoding apparatus according to the present embodiment is an apparatus that separates input image data into first layer image data that is an upper layer and second layer image data that is a lower layer, and encodes each. At the time of decoding, it is assumed that a layer effective for display is determined by setting the specific pixel value of the first layer to a transparent color. That is, the display of the second layer image is effective for the portion of the first layer image projected by the transparent region. The display of the first layer image is effective for the region portion that is not transparent in the first layer image. The specified specific pixel value is assigned a special effect of “transparency” that is basically unrelated to other pixel values. Even if the value of the specific pixel value is changed by encoding, even if it is very close, the targeted effect cannot be obtained. For this reason, conventionally, a lossless encoding method has been used when encoding a first layer of image data including a transparent color after setting a pixel to be transparent to a transparent color.

  In the present embodiment, even in the irreversible encoding method, a technique for controlling the layer separation and performing the encoding process so that the transparent portion of the image data of the first layer is surely rendered transparent after decoding. It is. An image encoding method used for encoding the first layer image data is referred to as a first layer image encoding method. The first layer image encoding method may be “an irreversible image encoding method having a flattening conversion pattern”. An image encoding method used for encoding the second layer image data is referred to as a second layer image encoding method. Any encoding method can be applied to the second layer image encoding method.

  In the present embodiment, both the first layer image data and the second layer image data are encoded by the JPEG image encoding method. In addition, encoding is performed under different conditions in each layer, such as by setting a parameter for encoding so that the image quality degradation due to the encoding of the first layer image data is smaller than the image quality deterioration due to the encoding of the second layer image data. It is desirable. This is because it is easy to obtain the advantage of encoding in multiple layers.

  First, the conceptual processing flow of the present invention will be described with reference to FIGS. FIG. 7 shows an example in which encoding is performed using a constant immobilization flattening conversion pattern for a portion where it is desired to be transparent. (A) is an input image, and will be described until the first layer image encoded data (g) and the second layer image encoded data (h) are created from this input image data. (B) shows an area to be displayed in the first layer image. Here, an area including black-displayed characters “ABCDEFGH” indicates an area to be displayed in the first layer image.

  The area to be displayed in the first layer image of the area information data (b) may be created manually by the user in accordance with the input image data. For example, an area in which “character data” exists is created. In the case of the area to be displayed in the first layer image, the “character” may be created by automatically discriminating from the input image using an image pattern discrimination program.

Next, layer separation information data shown in (d) is created.
As shown in (c), the layer separation information data (d) divides the region information data into function conversion processing units of the first layer image encoding method, and is displayed on the first layer within the function conversion processing unit region. If there is a region to be processed, the processing unit region for the function conversion is set as a first layer region. (The area filled with black in the layer separation information data (d) is the first layer area.) Further, the other area is the second layer area (the area painted with black in (e) corresponds). .

  In order to effectively display the first layer region of the layer separation information data (d) from the input image data (a) and make the other second layer region transparent, for example, a “constant immobilization flattening conversion pattern” The second layer region of the layer separation information data (d) is effectively displayed from the first layer image data (e) and the input image data (a), and is indicated by hatching in the first layer region ((f)). The second layer image data (f) that can be filled with a single color and increase the compression ratio is created.

  Specifically, the constant immobilization flattening conversion pattern for filling the second layer region of the first layer image data (e) is, as shown in FIG. 4A, all pixel values are 0 (flat value is 0). ) Flat region. The portion f1 indicated by diagonal lines in the second layer image data (f) may be filled with a single color because this portion is the first layer region to be displayed in the first layer image. This is because the second layer image data is not used for display. A more effective method for determining the pixel value for the f1 portion will be described later. The first layer image data (e) and the second layer image data (f) are encoded to generate first layer image encoded data (g) and second layer image encoded data (h).

  Note that the first layer image encoded data (g) and the second layer image encoded data (h) are output after being combined with the output image encoded data by the image encoded data combining process. The output image encoded data includes a header portion and a data portion, and the header portion stores the first layer image encoded data and the second layer image encoded data offset values from the beginning of the file. Information necessary for separating the layer image encoded data and the second layer image encoded data is provided. The “image encoded data joining process” will be described later with reference to FIGS. 9 and 10, and the configuration of the “output image encoded data” will be described later with reference to FIG.

  Therefore, strictly speaking, in the next FIG. 8, before the decoding process, the output image encoded data is separated into the first layer image encoded data (g) and the second layer image encoded data (h). However, in the description of FIGS. 7 and 8, this description is omitted for the sake of clarity.

  8 shows the first layer image encoded data (g) encoded in FIG. 7, the first layer image decoded data (j) obtained by decoding the second layer image encoded data (h), and the second layer image. It shows a process until a decoded image (m) obtained by synthesizing the decoded data (k) is created. In the decoded image (m) that has been correctly combined, the pixels of the region portion (second layer region) that are shown in black in the first layer image decoded data (j) are transparent, so this portion is It can be seen that the pixels of the second layer image decoded data (k) are applied to the region portion of the second layer image decoded data (k) to be projected, and the input image is effectively displayed. The pixel value (transparent color) of the pixel to be made transparent may be embedded in the header of the output image encoded data by the encoder and read at the decoding side on the decoding side.

  In addition, if a pixel value to be transparent is stored in the area to be transparent in the first layer image data at the time of decoding in the header portion of the output image encoded data, the decoding side can dynamically change the transparent color according to the data. It is also possible to change the pixel value. Here, “dynamically” means “not fixed”. In the first layer image data (e) in FIG. 7, the constant immobilization flattening conversion pattern is used for the region to be transparent, but instead using a flattening conversion pattern that becomes a flat region when decoded after encoding. In other words, the flat value after conversion may be designated as a transparent color at the time of decoding.

  Next, the system configuration of the image coding apparatus according to the present embodiment will be described with reference to FIG. The image encoding device 0104 receives region information data from the region information input device 0101 and input image data from the image input device 0102, and outputs the image encoded data to the image encoded data output device 0103.

  The image encoding device 0104 includes a region information data buffer 0105 for storing region information data from the region information input device 0101 and a plurality of layers of images generated by separating input images based on the region information data, for example, Layer separation information creation processing unit 0106 for creating layer separation information data for specifying images of the first layer and the second layer, a layer separation information data buffer 0108 for storing the created layer separation information data, and image input Based on the input image data buffer 0107 for storing the input image data from the device 0102 and the layer separation information data, the first layer image and the second layer image are separated from the input image data, and at least the first layer image The second layer region in the data includes a layer separation processing unit 0109 that performs processing for filling with the above-described planarization conversion pattern.

  The image encoding device 0104 further includes a first layer image data buffer 0110 and a second layer image data buffer 0111 for storing the first layer image data and the second layer image data output from the layer separation processing unit 0109, respectively. And the first layer image data in the first layer image data buffer 0110 is encoded by the first image image encoding method in which the region filled with the flattening conversion pattern at the time of decoding becomes a flat region. A first image encoding processing unit 0112; a second image encoding processing unit 0113 that encodes the second layer image data of the second layer image data buffer 0113 by a second image image encoding method; First and second image encoded data buffers 0114 and 0115 for storing data encoded by the second image encoding processing units 0112 and 0113, respectively. The image encoded data concatenation processing unit 0116 for generating output image encoded data by combining the encoded data in the first and second image encoded data buffers 0114 and 0115, and the generated output image encoded data And an output image encoded data buffer 0117 to be stored.

  In the following description, as an example of image encoding, both the first layer image data and the second layer image data are image encoded by the JPEG method.

FIG. 10 is a flowchart showing the flow of processing of the image encoding device 0104.
Here, the function of each part of the image encoding device 0104 and the processing performed by the image encoding device 0104 will be described with reference to FIGS. 9 and 10. The region information data buffer 0105 stores the region information data input from the region information input device 0101 (S0502). The format of the region information data is not limited as long as it is a data format that can specify the region to be expressed by the first layer image data of the image region by masking, coordinate designation, or the like. In the following, although not particularly limited, as shown in FIG. 7B, the area to be displayed on the first layer is specified by the information of the color black (character color in FIG. 7B). It will be explained as a thing. The layer separation information creation processing unit 0106 reads the region information data from the region information data buffer 0105 and outputs the layer separation information data (S0503).

  This layer separation information creation processing unit 0106 is an area unit processed by function conversion when encoding the first layer image data (8 × 8 pixel unit which is a processing unit of DCT conversion in the JPEG image encoding method). Then, it is determined whether or not it is the first layer area for each area unit of the input image. For example, it should be displayed on the first layer in the region unit processed by function conversion when encoding the first layer image data (8 × 8 pixel unit which is the DCT conversion processing unit in the JPEG image encoding method) When the region is included, the region is set as the first layer region. In addition, the ratio of pixels that specify the input image in the area unit processed by function conversion at the time of encoding the first layer image data (in the JPEG image encoding method, the unit of 8 × 8 pixels that is the processing unit of DCT conversion) May be set as the first layer region, for example.

  The layer separation information data buffer 0108 stores the layer separation information data output by the layer separation information creation processing unit 0106 (S0504). On the other hand, the input image data buffer 0107 receives and stores input image data from the image input device 0102 (S0505). The layer separation processing unit 0109 reads the layer separation information data from the layer separation information data buffer 0108 and the input image data from the input image data buffer 0107, respectively, and outputs the first layer image data and the second layer image data ( S0506). Here, the layer separation processing unit 0109 fills the first layer area of the first layer image data with the pixel value of the input image and fills the other areas with the flattening conversion pattern based on the layer separation information data. The area filled with the flattening conversion pattern becomes a flat area in the subsequent encoding process, and if the flat value is designated as a transparent color, the image information of the second layer is included in the area filled with this flattening conversion pattern. Will be applied.

  Here, for example, a constant immobilization flattening conversion pattern of the JPEG image encoding method is used as the flattening conversion pattern. That is, all the pixel values in the region of the first layer image data that is not the first layer region in the layer separation information data are set to 0. Of the regions of the second layer image data based on the layer separation information data, the second layer region may be filled with pixel values of the input image data, and the first layer region may be filled with arbitrary pixel values. In the present embodiment, the value of the input image data is filled in the second layer image data regardless of the layer separation information data. The first layer image data output by the layer separation processing unit 0109 is stored in the first layer image data buffer 0110, and the second layer image data is stored in the second layer image data buffer 0111, respectively (S0507).

  The first layer image encoding processing unit 0112 reads the first layer image data from the first layer image data buffer 0110, outputs the JPEG encoded first layer image encoded data, and the second layer image encoding processing unit In step 0113, the second layer image data is read from the second layer image data buffer 0111, and the JPEG-encoded second layer image encoded data is output (S0508). The first layer image encoded data output from the first layer image encoding processing unit 0112 is stored in the first layer image encoded data buffer 0114 and the second layer image encoded data output from the second layer image encoding processing unit 0113. The encoded data is respectively stored in the second layer image encoded data buffer 0115 (S0509). The image encoded data joint processing unit 0116 reads the first layer image encoded data from the first layer image encoded data buffer 0114 and the second layer image encoded data from the second layer image encoded data buffer 0115, respectively. The output image encoded data is output (S0510).

  FIG. 11 is a conceptual diagram showing the structure of output image encoded data. This structure is composed of a header part and a data part, in which the first layer image encoded data and the offset value from the file head of the second layer image encoded data are stored in the header part, and the first layer image encoded data And the information necessary for separating the second layer image encoded data. As described above, any structure may be used as long as the first layer image encoded data and the second layer image encoded data can be separated. Further, if a pixel value to be transparent in the first layer image data is stored in the header portion at the time of decoding, it becomes possible to dynamically change the pixel value to be transparent in accordance with the data on the decoding side. Here, “dynamically” means “not fixed”.

  Referring to FIG. 10 again, the output image encoded data output from image encoded data joining processing unit 0116 is stored in output image encoded data buffer 0117 (S0511).

  Area information data buffer 0105, input image data buffer 0107, layer separation information data buffer 0108, first layer image data buffer 0110, second layer image data buffer 0111, first layer image encoded data buffer 0114, second layer image code The encoded data buffer 0115 and the output image encoded data buffer 0117 are realized by a RAM (Random Access Memory) such as a flash memory or a hard disk.

  The layer separation information creation processing unit 0106, the layer separation processing unit 0109, the first layer image encoding processing unit 0112, the second layer image encoding processing unit 0113, and the image encoded data joining processing unit 0116 are, for example, dedicated circuits respectively. Realized by hardware. Alternatively, for example, the function may be realized by an arithmetic processing circuit such as a computer based on software.

  The modified example of the above description will be supplemented below. The first layer image encoding method may be “an irreversible image encoding method having a flattening conversion pattern”. When the first layer image data is generated by the layer separation processing unit 0109, the second layer region may be filled with a flattening conversion pattern of the first layer image encoding method. The flattening conversion pattern of the first layer image encoding method becomes a flat region by encoding by the first layer image encoding method. By setting the flat value of the flat area as the transparent color designation pixel value at the time of decoding of the first layer, the transparent color can be designated, and the transparent color area is used for display by the decoding result of the second layer which is the lower layer. It is done.

  In the creation of the second layer image data by the layer separation processing unit 0109, the first layer region may be basically filled with arbitrary pixel values because the second layer image data is not used for display during decoding. For example, it may be filled with pixel values that are advantageous for encoding by the second layer image encoding method. Thereby, the capacity | capacitance of 2nd layer image coding data can be reduced, and the capacity | capacitance of output image coding data can also be reduced.

  However, since the lossy encoding method is used for encoding the first layer image data, pixel values other than the region that is desired to be designated as a transparent color after encoding and decoding are encoded and decoded, and the pixel value that is specified as a transparent color after the decoding. It can be. In this case, the pixel value of the second layer image encoded data is used for display for the pixel that happens to have the pixel value designated as transparent color. Considering this, it is desirable to fill the first layer area of the second layer image data with the pixel value designated as the transparent color in the first layer in the creation of the second layer image data in the layer separation processing unit 0109.

The transparent color is an agreement in the first layer until it gets tired, and there is no concept of a transparent color in the second layer.
Therefore, in the second layer image data, the pixel value designated as the transparent color in the first layer is a simple pixel value that does not have the function of transparency. In the first layer image, the pixel other than the pixel to be designated as the transparent color becomes transparent when the pixel value designated as the transparent color is obtained after encoding and decoding, and the second layer image is displayed. In the creation of the second layer image data in the part 0109, the first layer region of the second layer image data is filled with the pixel value designated as the transparent color in the first layer, and the displayed pixels are transparent in the first layer. Since the pixel value specified for the color is encoded, an appropriate pixel value is used for display.

  Furthermore, a better way of filling the pixel values of the first layer area of the second layer image data is to use a flattening conversion pattern of the second layer image encoding method in which the flat value is a pixel value that is a transparent color in the first layer. Is to use. Then, when the pixel value other than the pixel that is designated as the transparent color in the first layer image becomes the pixel value designated as the transparent color after encoding and decoding, the decoded data of the second layer is the transparent color in the first layer. It becomes the pixel value itself.

  That is, when the “irreversible image coding method having a flattening conversion pattern” is used as the second layer image coding method, the flat value after encoding and decoding of the flattening conversion pattern is the first layer image code. If the flat value that matches the flat value after encoding and decoding by the first layer image encoding method is set as the transparent color of the first layer, the first layer of the second layer image data. The pixel value of the area can be represented by the pixel value completely designated as the transparent color of the first layer for each processing unit to which the function conversion of encoding is performed.

  That is, if the flattening conversion patterns that are encoded by the first layer image encoding method and the second layer image encoding method and have the same flat value after decoding are defined as the flattening conversion pattern 1 and the flattening conversion pattern 2, respectively. When the first layer image data is created by the layer separation processing unit 0109, the second layer region is filled with the flattening conversion pattern 1, and the second layer image data is created by encoding the second layer image data. When the unit area on which the function conversion process is performed is the first layer area, this unit area is filled with the flattening conversion pattern 2, and the other unit areas are filled with the pixel values of the input image data.

  Further, when both the function conversion processing unit of the image encoding method for encoding the first layer image data and the function conversion processing unit of the image encoding method for encoding the second layer image data are both blocks, both function conversion processing is performed. The first layer region or the second layer region performed by the layer separation information creation processing unit 0106 in block units in which the common multiple of the vertical size of the unit block and the common multiple of the horizontal size are the vertical size and horizontal size, respectively. If this determination is made, the first layer region can be completely filled with the flattening conversion pattern 2 by creating the second layer image data in the layer separation processing unit 0109.

  More preferably, when both the function conversion processing unit of the first layer image encoding method and the function conversion processing unit of the second layer image encoding method are blocks, the first layer region performed by the layer separation information creation processing unit 0106 This determination can be made in units of blocks in which the common multiple of the vertical size and the common multiple of the horizontal size of the block in both function conversion processing units are set to the vertical size and horizontal size, respectively. By setting the size of the function transform block of each irreversible encoding method of each layer that is irreversibly encoded for layer separation, each region can be set efficiently.

  Further, the layer separation information creation processing unit 0106 performs the block separation in which the least common multiple of the vertical size and the least common multiple of the horizontal size of the block in the unit of both function conversion processing units are the vertical size and horizontal size, respectively. If the presence / absence of the one-layer region is determined, the first-layer region can be completely filled with the flattening conversion pattern 2 in the creation of the second-layer image data in the layer separation processing unit 0109, and both function conversion can be performed. If the presence / absence of the first layer area is determined in block units of the least common multiple of processing units, the transparent designation area can be designated most finely.

[Embodiment 2]
The image coding apparatus according to the present embodiment, like the image coding apparatus according to the first embodiment, separates input image data into first layer image data that is an upper layer and second layer image data that is a lower layer. Is a device for encoding. The difference from the first embodiment is that a process for performing resolution conversion for changing the resolution of each layer image data is added. By changing the resolution for each layer, the resolution can be lowered depending on the layer, so that the compression rate is increased. . In addition, since it is possible to assign the capacity reduced in the layer having a lower resolution to a layer that requires more bits in compression, the image quality can be improved. The system configuration of the image coding apparatus according to the present embodiment is such that the layer separation processing unit 0109 in the system configuration of the image coding apparatus according to the first embodiment shown in FIG. 9 is similar to the layer separation processing unit 0205 of FIG. It is the same except for the point. Further, the processing flow is the same as that of the first embodiment except for the layer separation processing S0506 in FIG. Therefore, in the image coding apparatus according to the present embodiment, the same portions are denoted by the same reference numerals, and description of the same configuration and processing as those of the first embodiment described above is omitted.

  However, the layer separation information creation process (S0503 in FIG. 10) to the layer separation process (S0506 in FIG. 10) will be supplementarily described below.

  That is, in the second embodiment, the process flow of the layer separation processing unit 0205 shown in FIG. 12 is replaced with another process corresponding to the layer separation process (S0506) in FIG. FIG. 13 shows a process S0602 that replaces such S0506.

  Hereinafter, referring to FIG. 12 and FIG. 13, the first layer resolution conversion processing unit 0206 reads the input image data from the input image data buffer 0107 and converts the resolution to T1 (for example, 0.5 times). The first layer resolution conversion image data is output, and the second layer resolution conversion processing unit 0207 reads the input image data from the input image data buffer 0107 and converts the resolution to T2 times (for example, 0.25 times). Two-layer resolution converted image data is output (S0604). T1 and T2 may be positive real numbers. Such a resolution conversion method can use bilinear, for example. However, it is not necessary to limit to bilinear, and any method can be used as long as it is a resolution conversion method such as bicubic or simple thinning.

  In the present embodiment, for the sake of simplicity, the vertical resolution and the horizontal resolution are the same, and the resolutions of the first layer and the second layer are simply described as T1 times and T2 times. Yes. However, it goes without saying that the vertical resolution and horizontal resolution of each layer can be set to different values.

  The first layer resolution conversion image data buffer 0208 stores the first layer resolution conversion image data output from the first layer resolution conversion processing unit 0206, and the second layer resolution conversion image data buffer 0209 stores the second layer resolution conversion image data. The second layer resolution conversion image data output by the processing unit 0207 is stored (S0605). The pixel value change processing unit 0210 receives the layer separation information data from the layer separation information data buffer 0108, the first layer resolution conversion image data from the first layer resolution conversion image data buffer 0208, and the second layer resolution conversion image data buffer 0209. The second layer resolution conversion image data is read, and the first layer image data and the second layer image data are output (S0606). The first layer image data output from the pixel value change processing unit 0210 is stored in the first layer image data buffer 0110, and the second layer image data is stored in the second layer image data buffer 0111, respectively (S0507).

  In the first layer image data, the region designated as the first layer region by the layer separation information data is filled with the pixel value of the input image, and the other regions are filled with the flattening conversion pattern. If the area filled with the flattening conversion pattern is decoded in a later encoding process, the pixel value becomes a constant value.If this value is designated as a transparent color, the area filled with the flattening conversion pattern becomes transparent. Thus, the image information of the second layer is applied. Here, it is assumed that a JPEG image encoding method constant immobilization flattening conversion pattern is used as the flattening conversion pattern. That is, all the pixel values in the area of the first layer image data not designated as the first layer area by the layer separation information data are set to 0. In the second layer image data, the region other than the first layer region may be filled with pixel values of the input image data, and the first layer region may be filled with arbitrary pixel values. In the present embodiment, the pixel values of the input image data are filled in. That is, the input image data itself or a copy is used.

  Here, a supplementary explanation will be given of the layer separation information creation processing (S0503 in FIG. 10). In the layer separation information creation process (S0503), as described in the first embodiment, whether or not to set the first layer area in units of areas to be subjected to function conversion when encoding the first layer image data. It is a process which determines. For example, when the first layer image encoding method is the JPEG image encoding method, in the first embodiment, the region for determining whether or not to be the first layer region is 8 × 8 pixel units in terms of the resolution of the input image. However, in this embodiment, since the resolution of the first layer image data to be encoded is T1 times, the area unit processed by function conversion when encoding the first layer image data is In terms of the resolution of the input image, it should be noted that the unit is (8 / T1) × (8 / T1) pixel units. For example, if the region information data includes a region to be displayed on the first layer in units of regions processed by function conversion when encoding the first layer image data, the region is referred to as a first layer region. To do. In addition, when the ratio of pixels that specify the input image is a certain value or more in a region unit processed by function conversion at the time of encoding the first layer image data, the region is set as the first layer region. May be.

  As described above, when the vertical resolution of the first layer is set to T1 vertical magnification and the horizontal resolution of the first layer is set to a value different from T1 horizontal magnification, the first layer image data The area unit processed by function conversion when encoding is 8 / T1 vertical × 8 / T1 horizontal) pixel unit. The second layer image data also needs to be processed under the same considerations as the first layer image data.

  The first layer resolution conversion image data buffer 0208 and the second layer resolution conversion image data buffer 0209 are realized by a RAM (Random Access Memory) such as a flash memory or a hard disk.

  The first layer resolution conversion processing unit 0206, the second layer resolution conversion processing unit 0207, and the pixel value change processing unit 0210 are each realized by hardware such as a dedicated circuit, for example. Further, the function may be realized by an arithmetic processing circuit such as a computer based on software, for example.

  The modified example of the above description will be supplemented below. The first layer image encoding method may be “an irreversible image encoding method having a flattening conversion pattern”. In creating the first layer image data by the pixel value change processing unit 0210, the second layer region may be filled with a flattening conversion pattern of the first layer image encoding method. The flattening conversion pattern of the first layer image encoding method becomes a flat region by the encoding by the first layer image encoding method, but the flat value is made transparent as a transparent color designation pixel value at the time of decoding. Color can be specified.

  In the creation of the second layer image data by the pixel value change processing unit 0210, the second layer image data in the first layer region is not used for display at the time of decoding. What is necessary is just to fill with the pixel value advantageous for the encoding by a 2nd layer image encoding system. Thereby, the capacity | capacitance of 2nd layer image coding data can be reduced, and the capacity | capacitance of output image coding data can also be reduced. However, since the lossy encoding method is used for encoding the first layer image data, pixel values other than those that are desired to be designated as transparent colors after encoding may become pixel values that are designated as transparent colors after encoding. It is possible. In this case, the pixel value of the second layer image encoded data is used for display for the pixel that happens to have the pixel value designated as transparent color. In consideration of this, it is desirable that the second layer image data in the first layer region is filled with the pixel value designated as the transparent color in the first layer in the creation of the second layer image data in the pixel value change processing unit 0210. .

  Only the pixel value of the first layer image data defines the transparent color, and it does not become transparent after decoding. Therefore, the pixels of the second layer image data are decoded into the pixel values originally possessed by the pixels. This is because the second layer image data does not have a transparent color, and the first layer image data and the second layer image data have different encoding (compression) conditions such as encoding parameters. . When the pixel value other than the pixel desired to be designated as the transparent color in the first layer image becomes the pixel value designated as the transparent color after encoding and decoding, the pixel becomes transparent and the second layer image is displayed.

  Further, when the “lossy image coding method having a flattening conversion pattern” is used as the second layer image coding method, the flat value after encoding and decoding is included in the first layer image code. If the flat value that matches the flat value after encoding and decoding by the first layer image encoding method is set as the transparent color of the first layer, the first layer of the second layer image data. The pixel value of the area can be represented by the pixel value completely designated as the transparent color of the first layer for each processing unit to which the function conversion of encoding is performed.

  That is, if the flattening conversion patterns that are encoded by the first layer image encoding method and the second layer image encoding method and have the same flat value after decoding are defined as the flattening conversion pattern 1 and the flattening conversion pattern 2, respectively. In the creation of the first layer image data by the pixel value change processing unit 0210, the second layer area is filled with the flattening conversion pattern 1, and the second layer image data is created by encoding the second layer image data. When the unit area on which the function conversion process is performed is the first layer area, this unit area is filled with the flattening conversion pattern 2, and the other unit areas are filled with the pixel values of the input image data.

  Furthermore, both the function conversion processing unit of the image encoding method for encoding the first layer image data and the function conversion processing unit of the image encoding method for encoding the second layer image data are both the first layer resolution conversion processing unit 0206. In the case of a block obtained by multiplying the resolution conversion magnification in the input image resolution by the resolution of the input image and a block obtained by multiplying the resolution conversion magnification in the second layer resolution conversion processing unit 0207 by the resolution of the input image, these blocks are converted to the first layer resolution. The block size divided by the resolution conversion magnification in the conversion processing unit 0206 and the block size divided by the resolution conversion magnification in the second layer resolution conversion processing unit 0207 are common multiples of the respective vertical sizes, Whether the first layer region or the second layer region is performed by the layer separation information creation processing unit 0106 in block units in which the common multiple of the horizontal size is the vertical size and the horizontal size. By performing the determination, the creation of the second layer image data in the pixel value change processing unit 0210, the first layer region are all it is possible to fill a planarization conversion pattern 2. By setting the size of the function transform block of each irreversible encoding method of each layer that is irreversibly encoded for layer separation, each region can be set efficiently.

  Note that, as described above, the function conversion processing unit of the image encoding method for encoding the first layer image data by setting the resolutions of the first layer and the second layer to values different from the vertical direction and the horizontal direction, respectively. And the function conversion processing unit of the image encoding method for encoding the second layer image data, the vertical resolution conversion magnification (T1 length) in the first layer resolution conversion processing unit 0206 and the first layer resolution conversion processing unit The horizontal resolution conversion magnification (T1 horizontal) in 0206 is multiplied by the resolution of the input image in the vertical direction and horizontal direction, respectively, and the resolution is converted in the vertical direction in the second layer resolution conversion processing unit 0207. The resolution is converted by multiplying the resolution conversion magnification (T2 vertical) and the horizontal resolution conversion magnification (T2 horizontal) in the second layer resolution conversion processing unit 0207 by the resolution of the input image in the vertical and horizontal directions, respectively. If with a block For be described.

  In this case, the first layer block obtained by dividing the resolution-converted first layer block by T1 length and T1 width, and the second layer block obtained by converting the resolution by T2 length and T2 width. Layer separation information creation processing unit in units of blocks, each of which is a common multiple of the vertical size of the divided second layer block, and a common multiple of each horizontal size. In step 0106, it is determined whether the region is the first layer region or the second layer region. By doing so, it is possible to fill the entire first layer region with the flattening conversion pattern 2 by creating the second layer image data in the pixel value change processing unit 0210.

  More preferably, both the function conversion processing unit of the image encoding method for encoding the first layer image data and the function conversion processing unit of the image encoding method for encoding the second layer image data both have the first layer resolution. In the case of the block multiplied by the resolution conversion magnification in the conversion processing unit 0206 and the block multiplied by the resolution conversion magnification in the second layer resolution conversion processing unit 0207, the resolution conversion magnification in the first layer resolution conversion processing unit 0206 The least common multiple of the vertical size and the least common multiple of the horizontal size of the divided block size and the block size divided by the resolution conversion magnification in the second layer resolution conversion processing unit 0207 are respectively If the layer separation information creation processing unit 0106 determines whether the region is the first layer region or the second layer region, the pixel value change processing unit 0210 may Creating the second layer image data, the first layer region are all it is possible to fill a planarization conversion pattern 2. For this reason, if it is determined whether the first layer region or the second layer region in block units of the least common multiple of both function conversion processing units, the transparent designation region can be designated most finely.

  Note that, as described above, the function conversion processing unit of the image encoding method for encoding the first layer image data by setting the resolutions of the first layer and the second layer to values different from the vertical direction and the horizontal direction, respectively. And the function conversion processing unit of the image encoding method for encoding the second layer image data, the vertical resolution conversion magnification (T1 length) in the first layer resolution conversion processing unit 0206 and the first layer resolution conversion processing unit The horizontal resolution conversion magnification (T1 horizontal) in 0206 is multiplied by the resolution of the input image in the vertical direction and horizontal direction, respectively, and the resolution is converted in the vertical direction in the second layer resolution conversion processing unit 0207. The resolution is converted by multiplying the resolution conversion magnification (T2 vertical) and the horizontal resolution conversion magnification (T2 horizontal) in the second layer resolution conversion processing unit 0207 by the resolution of the input image in the vertical and horizontal directions, respectively. If with a block For be described.

  In this case, the first layer block obtained by dividing the resolution-converted first layer block by T1 length and T1 width, and the second layer block obtained by converting the resolution by T2 length and T2 width. Layer separation information is created in block units with the least common multiple of the vertical size of the divided second layer block and the least common multiple of each horizontal size as the vertical size and horizontal size. The processing unit 0106 determines whether it is the first layer region or the second layer region. By doing this, the transparent designated area can be designated most finely.

  In the above, the block obtained by multiplying the resolution conversion magnification in the first layer resolution conversion processing unit 0206 by the resolution of the input image, and the block obtained by multiplying the resolution conversion magnification in the second layer resolution conversion processing unit 0207 by the resolution of the input image. The case where the determination of whether the region is the first layer region or the second layer region has been described. However, as a modification, first, the layer separation information creation processing unit 0106 determines whether the first layer region or the second layer region in units of blocks of input image data. Thereafter, the first layer region of the first layer image data using the input image data and the second layer region of the second layer image data are processed by the first layer resolution conversion processing unit 0206 and the second layer resolution conversion processing unit 0207. The first layer image data and the second layer image data may be generated using the respective image data obtained by multiplying the input image data by the resolution conversion magnification.

  This means that the first layer region or the second layer performed by the layer separation information creation processing unit 0106 is a block unit having a common multiple of the vertical size, a common multiple of the horizontal size, or a least common multiple of each block. The same applies to the determination of the area. Also, the vertical resolution of the first layer is T1 vertical magnification, the horizontal resolution of the first layer is T1 horizontal magnification, the vertical resolution of the second layer is T2 vertical magnification, and the horizontal resolution of the second layer is The same applies to the case where the T2 horizontal magnification is set to a different value.

[Embodiment 3]
The image encoding apparatus according to the present embodiment is an apparatus that separates input image data into a plurality of layers and encodes them. Even if there are three or more layers to be divided, the decoding side can be combined into one image by specifying a transparent color as in the first embodiment. The image coding apparatus according to the present embodiment performs coding by separating into N layers. The N layer is from the first layer to the Nth layer as shown in FIG. 14, and is assumed to be a lower layer as the layer number increases. Further, a transparent color that is a pixel value that is transparent at the time of decoding is defined for each layer other than the Nth layer.

  The input image data is separated into a plurality of layers, each is encoded, and output image encoded data obtained by joining the encoded layer image encoded data of each layer is output. The output image encoded data is decoded in the order of the Nth layer, the N-1th layer, the N-2th layer,..., The first layer, and the pixel values other than the transparent color are overwritten. . Thereby, the pixels of the entire image are decoded. Note that an image encoding method in the case of performing image encoding processing on n-th layer image data is referred to as an n-th layer image encoding method.

  Here, at least one of the N layers of image data is set to the above-described “lossy image encoding method having a flattening conversion pattern”. When using the `` lossy image coding method having a flattening conversion pattern '', for a flat region that becomes a flat value after decoding, an image can be reproduced by using a decoded image lower than the layer, In particular, by designating the flat value as the transparent color in the layer, it is possible to reproduce the image only by superimposing the decoded images of the first to Nth layers. For example, a reversible image coding method may be used for some layers having a transparent region. At this time, with respect to the lossless image encoding method, it is not always necessary to perform encoding in units of regions. Further, uncompressed images may be used for some of the first to Nth layers. For the Nth layer, since there is no provision for a transparent color, any coding method that does not process in units of regions may be used.

  In the following description, for simplicity of explanation, it is assumed that the first to Nth layer image data is image-encoded by the JPEG method as an example of image encoding.

  FIG. 15 is a diagram showing a system configuration of the image coding apparatus according to the present embodiment. The image encoding device 0304 receives region information data from the region information input device 0301 and input image data from the image input device 0302, and outputs the image encoded data to the image encoded data output device 0303.

  The operation of the image coding apparatus 0304 of the present embodiment is also basically the same as that of the first embodiment described with reference to FIG. However, in response to the separation of the input image into N layers, the configuration and function of each unit of the image encoding device 0304 need to be changed as described below.

  Here, the function of each unit of the image encoding device 0304 and the processing of the image encoding device 0304 will be described with reference to FIGS. 15 and 10.

  The area information data buffer 0305 stores area information data input from the area information input device 0301 (S0502). Here, the region information data only needs to be able to specify the region to be displayed by the n-th layer image data of the image region by mask or coordinate designation. The layer separation information creation processing unit 0306 reads the region information data from the region information data buffer 0305 and outputs the layer separation information data (S0503).

  Here, a supplementary description will be given of the layer separation information creation processing S0503. The layer separation information creation processing S0503 is performed in units of areas where function conversion processing is performed when encoding n-th layer image data (8 × 8 pixel units, which are DCT conversion processing units in the JPEG image encoding method). This is a process for determining whether or not it is the nth layer region to be displayed in the nth layer. Similarly to the first embodiment, for example, an area including an area to be displayed in the nth layer is included in the nth layer in a unit of area where function conversion is performed when encoding the nth layer image data. The layer area. In addition, when the ratio of pixels that specify an input image is a certain value or more in the unit of area where function conversion processing is performed when encoding the n-th layer image data, the area is set as the n-th layer area. May be.

  The layer separation information data output by the layer separation information creation processing unit 0306 is stored in the layer separation information data buffer 0308 (S0504). The input image data is received from the image input device 0302 and stored in the input image data buffer 0307 (S0505). The layer separation processing unit 0309 reads the layer separation information data from the layer separation information data buffer 0308 and the input image data from the input image data buffer 0307, and outputs each layer image data (S0506). According to the layer separation information data of the nth layer image data, the nth layer region is the pixel value of the input image, and the first layer region to the (n−1) th layer in the first layer image data to the (n−1) th layer image data. The layer region may be filled with an arbitrary pixel value, and may be filled with a pixel value advantageous for encoding by each layer image encoding method.

  A region other than the first layer region to the n-th layer region in the n-th layer image data, that is, a region to be made transparent is filled with a flattening conversion pattern. The area filled with the flattening conversion pattern becomes a flat area in the subsequent encoding process. If the flat value is designated as a transparent color, the area filled with the flattening conversion pattern is the (n + 1) -th layer image. Information will be applied. Here, it is assumed that a JPEG image encoding method constant immobilization flattening conversion pattern is used as the flattening conversion pattern. That is, all the pixel values in the n-th layer image data region other than the first layer region to the n-th layer region are set to 0 by the layer separation information data.

  Each layer image data output by the layer separation processing unit 0309 is stored in the layer image data buffer group 0310 (S0507). The layer image encoding processing unit group 0311 reads each layer image data from the layer image data buffer group 0310, and outputs each layer encoded data obtained by encoding each layer image data (S0508). Each layer image encoded data output from the layer image encoding processing unit group 0311 is stored in the layer image encoded data buffer group 0312 (S0509). The image encoded data joining processing unit 0313 reads each layer image encoded data from the layer image encoded data buffer group 0312 and outputs the output image encoded data (S0510). The output image encoded data output from the image encoded data joint processing unit 0313 is stored in the output image encoded data buffer 0314 (S0511).

  FIG. 16 is a conceptual diagram showing a structure of output image encoded data according to the present embodiment. This structure also includes a header portion and a data portion, and each header value stores each offset value from the beginning of the file of each layer image encoded data, so that information necessary for separating each layer image encoded data is provided. ing. In this way, any structure may be used as long as each layer image encoded data can be separated. In addition, if a pixel value to be transparent in each layer image encoded data is stored in the header portion at the time of decoding, it is possible to dynamically change the pixel value to be transparent according to the data on the decoding side. Here, “dynamically” means “not fixed”.

  The area information data buffer 0305, the input image data buffer 0307, the layer separation information data buffer 0308, the layer image data buffer group 0310, the layer image encoded data buffer group 0312, and the output image encoded data buffer 0314 are a flash memory, a hard disk, etc. This is realized by a RAM (Random Access Memory).

  The layer separation information creation processing unit 0306, the layer separation processing unit 0309, the layer image encoding processing unit group 0311, and the image encoded data joining processing unit 0313 are each realized by hardware such as a dedicated circuit. For example, the function may be realized by an arithmetic processing circuit such as a computer based on software.

  The modified example of the above description will be supplemented below. The image encoding method of the layer that performs encoding having a transparent color may be “an irreversible image encoding method having a flattening conversion pattern”. When the n-th layer image data is generated by the layer separation processing unit 0309, the regions other than the first layer region to the n-th layer region in the n-th layer image may be filled with a flattening conversion pattern of the n-th layer image encoding method. . In the transparent color designation, the flattening conversion pattern of the n-th layer image encoding method becomes a flat region after decoding by encoding by the n-th layer image encoding method, but the flat value is converted into the n-th layer at the time of decoding. The transparent color designation pixel value may be used.

  In the generation of the nth layer image data by the layer separation processing unit 0309, the first layer region to the (n−1) th layer region are not used for display at the time of decoding, and thus are basically filled with arbitrary pixel values. What is necessary is just to fill with the pixel value advantageous for the encoding by each layer image encoding system. Thereby, the capacity | capacitance of each layer image coding data can be reduced, and the capacity | capacitance of output image coding data can also be reduced.

  However, since the lossy encoding method is used for encoding the n-th layer image data, pixel values other than those that are desired to be designated as a transparent color after encoding may become pixel values that are designated as transparent colors after encoding. It is possible. In this case, for the pixel that happens to have a pixel value designated as transparent, the pixel value of the layer image encoded data below the layer is used for display. In consideration of this, in the generation of the nth layer image data in the layer separation processing unit 0309, the image data of the first layer region to the (n−1) th layer region is the first layer to the (n−1) th layer. It is desirable to fill with a pixel value designated as a transparent color in any one of the layers.

  Furthermore, the n-th layer image encoding method is “an irreversible image code having a flat value conversion pattern” in order to better fill pixel values in the first layer region to the (n−1) -th layer region in the n-th layer image data. In the case of “the conversion method”, the flattening conversion pattern of the (n−1) th layer image encoding method whose flat value is a pixel value which is a transparent color in the (n−1) th layer is used. Then, in the (n−1) -th layer image, when the pixel value other than the pixel that is designated as the transparent color and the pixel value that is designated as the transparent color after encoding is decoded, the decoded data of the n-th layer is the (( n-1) The pixel value itself is a transparent color in the layer.

  In addition, when the image encoding method is “lossy image encoding method having a flat value conversion pattern” and the function conversion processing unit of the image encoding method is a block, the common multiple of the vertical size of each block, the horizontal If the layer separation information creation processing unit 0306 determines the presence or absence of the first layer region to the n-th layer region in units of blocks each having a common multiple of the vertical size and horizontal size, the layer separation processing unit 0309 In the creation of the n-th layer image data, the first layer region to the (n−1) th layer region can all be filled with the flattening conversion pattern (n−1). Here, a flattening conversion pattern for obtaining a flat region having the flat value of the (n−1) -th layer image encoding method is defined as a flattening conversion pattern (n−1). By setting the size of the function transform block of each irreversible encoding method of each layer that is irreversibly encoded for layer separation, each region can be set efficiently.

  More preferably, when the function conversion processing unit of each layer image encoding method is a block, the least common multiple of the vertical size and the least common multiple of the horizontal size of each block are If the layer separation information creation processing unit 0306 determines the presence or absence of the first layer region to the (n−1) th layer region for each block size, the layer separation processing unit 0309 can create the n-th layer image data. The first layer region to the (n−1) th layer region can all be filled with the planarization conversion pattern (n−1). Therefore, if the presence / absence of the first layer region to the (n−1) th layer region is determined in block units of the least common multiple of the function conversion processing unit of each layer image encoding method, the transparent designation region can be designated most finely.

[Embodiment 4]
Similar to the image coding apparatus according to the third embodiment, the image coding apparatus according to the present embodiment is an apparatus that separates input image data into N pieces of layer image data and codes them.

  The difference between the image coding apparatus of the fourth embodiment and the third embodiment is that a process of applying resolution conversion for changing the resolution of each layer image data is added, so that the compression rate and image quality can be changed by changing the resolution for each layer. It is possible to increase.

  The system configuration of the image coding apparatus according to the present embodiment is such that the layer separation processing unit 0309 in the system configuration of the image coding apparatus according to the third embodiment shown in FIG. 15 is similar to the layer separation processing unit 0404 in FIG. Except for the configuration, it is the same as the image coding apparatus of the third embodiment. Further, the processing flow is the same as that of the third embodiment except for the layer separation processing S0506 of FIG. Therefore, the description of the same configuration and processing as those of the third embodiment described above is omitted for the image encoding device of the present embodiment.

However, the layer separation information creation process (S0503 in FIG. 10) will be supplemented below.
The processing flow of the image coding apparatus according to the present embodiment having the layer separation processing unit 0404 as shown in FIG. 17 basically corresponds to the layer separation processing S0506 in FIG. 10 according to the flowchart shown in FIG. The processing to be performed corresponds to the processing replaced with S0602 in FIG.

  That is, the layer resolution conversion processing unit group 0405 reads the input image data from the input image data buffer 0307, and outputs each layer resolution converted image data converted to each layer predetermined resolution (S0604). The n-th layer resolution converted image data (n is an arbitrary natural number satisfying 1 ≦ n ≦ N) is obtained by converting the resolution of input image data to Tn times (Tn is a real number larger than 0, for example, 0.5 times). Bilinear can be used for such a resolution conversion method. However, it is not necessary to limit to bilinear, and any method can be used as long as it is a resolution conversion method such as bicubic or simple thinning.

  In the present embodiment, for the sake of simplicity, the vertical resolution and the horizontal resolution are the same, and the resolution of the nth layer is simply described as Tn times. However, it goes without saying that the vertical resolution and horizontal resolution of each layer can be set to different values.

  The layer resolution conversion image data buffer group 0406 stores each layer resolution conversion image data output by the layer resolution conversion processing unit group 0405 (S0605). The pixel value change processing unit 0407 reads the layer separation information data from the layer separation information data buffer 0308 and each layer resolution conversion image data from the layer resolution conversion image data buffer group 0406, and outputs each layer image data (S0606). Each layer image data output by the pixel value change processing unit 0407 is stored in the layer image data buffer group 0310 (S0507).

  In the nth layer image data, the region designated as the nth layer region by the layer separation information data is filled with the pixel value of the input image, and the first layer region to the (n−1) th layer region are filled with arbitrary pixels. What is necessary is just to fill with the pixel value advantageous for the encoding by each layer image encoding system.

  The regions other than the first layer region to the nth layer region are filled with a planarization conversion pattern. The area filled with the flattening conversion pattern becomes a constant value if it is decoded later in the encoding process. If this value is designated as a transparent color, the area filled with the flattening conversion pattern is lower than the nth layer. The image information is applied. Here, it is assumed that a JPEG image encoding method constant immobilization flattening conversion pattern is used as the flattening conversion pattern. That is, all the pixel values in the regions other than the first layer region to the nth layer region are set to 0 in the nth layer image data by the layer separation information data.

  In the Nth layer image data, the Nth layer region, which is a region to be displayed on the Nth layer, is filled with the pixel value of the input image data, and all other regions are from the first layer region to the (N−1) th layer region. ) Since it is a layer region, it may be filled with an arbitrary pixel value, and may be filled with a pixel value advantageous for encoding in each layer image encoding method.

Here, a supplementary explanation will be given of the layer separation information creation processing S0503 (FIG. 10).
As described in the third embodiment, the layer separation information creation processing S0503 is an nth layer that should be displayed on the nth layer in units of areas to be subjected to function conversion when encoding nth layer image data. This is a process for determining whether or not it is an area. For example, when the n-th layer image encoding method is a JPEG image encoding method, in the third embodiment, the region for determining whether or not the n-th layer region in the n-th layer image data is a DCT with the resolution of the input image. In this embodiment, since the resolution of the n-th layer image data to be encoded is multiplied by Tn, the n-th layer image data is encoded. It should be noted that the region unit processed by the function conversion at the time is (8 / Tn) × (8 / Tn) pixel unit in terms of the resolution of the input image. For example, it should be displayed on the nth layer in the area unit (8 × 8 pixel unit which is the DCT conversion processing unit in the JPEG image encoding method) in which the function conversion process is performed when encoding the nth layer image data. The region is defined as an nth layer region. In addition, the ratio of pixels that specify the input image in the unit of area where function conversion is performed when encoding the n-th layer image data (in the JPEG image encoding method, the unit of 8 × 8 pixels, which is the processing unit of DCT conversion) May be the n-th layer region, or the like.

  As described above, when the vertical resolution of the n-th layer is set to a value different from Tn vertical magnification and the horizontal resolution of the n-th layer is set to a value different from Tn horizontal magnification, the n-th layer image data The unit of area processed by function conversion when encoding is (8 / Tn vertical) × (8 / Tn horizontal) pixel units.

  In this case, in FIG. 16, information specifying the resolution of each layer may be stored as an encoding parameter stored in the header portion, and the parameter may be referred to for each layer in decoding.

  The layer resolution converted image data buffer group 0406 is realized by a random access memory (RAM) such as a flash memory or a hard disk.

  The layer resolution conversion processing unit group 0405 and the pixel value change processing unit 0407 are each realized by hardware such as a dedicated circuit, for example. Further, the function may be realized by an arithmetic processing circuit such as a computer based on software, for example.

  The modified example of the above description will be supplemented below. The image encoding method of the layer that performs encoding having a transparent color may be “an irreversible image encoding method having a flattening conversion pattern”. In the generation of the n-th layer image data in the pixel value change processing unit 0407, regions other than the first layer region to the n-th layer region may be filled with a flattening conversion pattern of the n-th layer image encoding method. In the transparent color designation, the flattening conversion pattern of the n-th layer image encoding method becomes a flat region after decoding by encoding by the n-th layer image encoding method, but the flat value is converted into the n-th layer at the time of decoding. The transparent color designation pixel value may be used.

  In the creation of the nth layer image data in the pixel value change processing unit 0407, the first layer region to the (n−1) th layer region are not used for display at the time of decoding. And may be filled with pixel values that are advantageous for encoding in each layer image encoding method. Thereby, the capacity | capacitance of each layer image coding data can be reduced, and the capacity | capacitance of output image coding data can also be reduced.

  However, since the lossy encoding method is used for encoding the n-th layer image data, pixel values other than those that are desired to be designated as a transparent color after encoding may become pixel values that are designated as transparent colors after encoding. It is possible. In this case, for the pixel that happens to have a pixel value designated as transparent, the pixel value of the layer image encoded data below the layer is used for display. In consideration of this, when the pixel value change processing unit 0407 generates the n-th layer image data, the image data of the first layer region to the (n−1) th layer region is the first layer to the (n−1) th layer. It is desirable to fill with a pixel value designated as a transparent color in any one of the layers.

  Furthermore, the n-th layer image encoding method is “an irreversible image code having a flat value conversion pattern” in order to better fill pixel values in the first layer region to the (n−1) -th layer region in the n-th layer image data. In the case of “the conversion method”, the flattening conversion pattern of the (n−1) th layer image encoding method whose flat value is a pixel value which is a transparent color in the (n−1) th layer is used. Then, in the (n−1) -th layer image, when the pixel value other than the pixel that is designated as the transparent color and the pixel value that is designated as the transparent color after encoding is decoded, the decoded data of the n-th layer is the (( n-1) The pixel value itself is a transparent color in the layer.

  Also, the image encoding method is “lossy image encoding method having flat value conversion pattern”, and the function conversion processing unit of the image encoding method is the resolution conversion magnification of each layer in the layer resolution conversion processing unit group 0405 of each layer. , The block size of each layer is divided by the conversion magnification of each layer in the layer resolution conversion processing unit 0405, and the common multiple of the vertical size of the block size and the common multiple of the horizontal size are respectively obtained. If the layer separation information creation processing unit 0306 determines the presence or absence of the first layer region to the n-th layer region in units of blocks having a vertical size and a horizontal size, the pixel value update processing unit 0407 performs the n-th processing. In the generation of the layer image data, the first layer region to the (n−1) th layer region can all be filled with the flattening conversion pattern (n−1). Here, a flattening conversion pattern for obtaining a flat region having the flat value of the (n−1) -th layer image encoding method is defined as a flattening conversion pattern (n−1). By setting the size of the function transform block of each irreversible encoding method of each layer that is irreversibly encoded for layer separation, each region can be set efficiently.

  As described above, the n-th layer image data is encoded by setting the vertical resolution of the n-th layer to Tn vertical double and the horizontal resolution of the n-th layer to different values of Tn horizontal double. Both the function conversion processing unit of the image encoding method and the function conversion processing unit of the image encoding method for encoding the i-th layer image data (1 ≦ i ≦ N and i is a natural number not including n) are layer resolution. The vertical resolution conversion magnification (Tn vertical) of the nth layer in the conversion processing unit group 0405 and the horizontal resolution conversion magnification (Tn horizontal) of the nth layer in the layer resolution conversion processing unit group 0405 are respectively A block whose resolution is converted by multiplying the resolution of the input image in the horizontal and horizontal directions, the vertical resolution conversion magnification (Ti vertical) of the i-th layer in the layer resolution conversion processing unit group 0405, and the layer resolution conversion processing unit group The resolution conversion magnification in the horizontal and vertical direction of the i-th layer at 0405 (Ti horizontal As each longitudinal, when multiplied with the resolution of the horizontal direction of the input image resolution is, the block in the i-layer which is converted will be described.

  In this case, the size of the n-th layer block obtained by dividing the n-th layer block whose resolution is converted by Tn vertical and Tn horizontal, and the i-th layer block whose resolution is converted are Ti vertical and Ti horizontal. Layer separation information creation processing in block units in which the common multiple of each vertical size and the common multiple of each horizontal size is the vertical size and horizontal size of the i-th layer block size divided by In part 0306, it is determined whether the region is an n-th layer region or an i-th layer region. Then, in the creation of the n-th layer image data in the pixel value change processing unit 0407, all of the first layer region to the (n−1) th layer region can be filled with the flattening conversion pattern (n−1). It becomes. Here, a flattening conversion pattern for obtaining a flat region having the flat value of the (n−1) -th layer image encoding method is defined as a flattening conversion pattern (n−1).

  More preferably, when the function conversion processing unit of the image encoding method for encoding each layer image data is a block obtained by multiplying the resolution conversion magnification of each layer in the layer resolution conversion processing unit group 0405 of each layer, The block size of each layer divided by the conversion magnification of each layer in the layer resolution conversion processing unit 0405 is the least common multiple of the vertical size and the least common multiple of the horizontal size, respectively. If the layer separation information creation processing unit 0306 determines the presence or absence of the first layer region to the (n−1) th layer region, the n-th layer image is displayed in the pixel value change processing unit 0407. In the creation of data, the first layer region to the (n−1) th layer region can all be filled with the planarization conversion pattern (n−1). Therefore, if the presence / absence of the first layer region to the (n−1) th layer region is determined in block units of the least common multiple of the function conversion processing unit of each layer image encoding method, the transparent designation region can be designated most finely.

  As described above, the n-th layer image data is encoded by setting the vertical resolution of the n-th layer to Tn vertical double and the horizontal resolution of the n-th layer to different values of Tn horizontal double. Both the function conversion processing unit of the image encoding method and the function conversion processing unit of the image encoding method for encoding the i-th layer image data (1 ≦ i ≦ N and i is a natural number not including n) are layer resolution. The vertical resolution conversion magnification (Tn vertical) of the nth layer in the conversion processing unit group 0405 and the horizontal resolution conversion magnification (Tn horizontal) of the nth layer in the layer resolution conversion processing unit group 0405 are respectively A block whose resolution is converted by multiplying the resolution of the input image in the horizontal and horizontal directions, the vertical resolution conversion magnification (Ti vertical) of the i-th layer in the layer resolution conversion processing unit group 0405, and the layer resolution conversion processing unit group The resolution conversion magnification in the horizontal and vertical direction of the i-th layer at 0405 (Ti horizontal As each longitudinal, when multiplied with the resolution of the horizontal direction of the input image resolution is, the block in the i-layer which is converted will be described.

  In this case, the size of the n-th layer block obtained by dividing the n-th layer block whose resolution is converted by Tn vertical and Tn horizontal, and the i-th block whose resolution is converted are Ti vertical and Ti horizontal. Layer separation information in block units with the least common multiple of the vertical size and the least common multiple of each horizontal size divided by the block size of the i-th layer divided by The creation processing unit 0306 determines whether the region is the n-th layer region or the i-th layer region. By doing this, the transparent designated area can be designated most finely.

  In the above, the resolution converted by multiplying the resolution conversion magnification of the nth layer in the layer resolution conversion processing unit group 0405 by the resolution of the input image, and the vertical of the i-th layer in the layer resolution conversion processing unit group 0405. A case has been described in which the resolution conversion magnification in the direction is multiplied by the resolution of the input image to convert the resolution to the i-th layer block. However, as a modification, first, in the block unit of the input image data, it is determined whether the layer separation information creation processing unit 0306 is the n-th layer region or the i-th layer region. Thereafter, for the nth layer region of the nth layer image data using the input image data, the image data obtained by multiplying the input image data by the resolution conversion magnification of the nth layer in the layer resolution conversion processing unit group 0405 is used. N-layer image data may be generated.

  This is because the n-th layer region or the i-th layer performed by the layer separation information creation processing unit 0306 is a block unit having a common multiple of the vertical size, a common multiple of the horizontal size, or a least common multiple of each block. The same applies to the determination of the area. The vertical resolution of the nth layer is Tn vertical, the horizontal resolution of the nth layer is Tn horizontal, the vertical resolution of the ith layer is Ti vertical, and the horizontal resolution of the ith layer is The same applies to the case where the value is set to a value different from Ti horizontal magnification.

  The image encoding method described in the present embodiment is not only put into practical use as an image encoding apparatus, but also an image encoding program using the image encoding method as a program and a computer-readable recording of the image encoding program The present invention can also be applied as a simple recording medium.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is explanatory drawing of the planarization conversion pattern. It is explanatory drawing of a constant planarization conversion pattern. It is explanatory drawing of a fixed flattening conversion pattern. It is explanatory drawing of a permanent immovable flattening conversion pattern. It is inclusion explanatory drawing regarding area | region data. It is inclusion explanatory drawing regarding an irreversible image coding system. It is explanatory drawing of the example which encodes using the fixed stationary flattening conversion pattern in the location which wants to be transparent. It is explanatory drawing until it decodes and synthesize | combines the 1st layer and 2nd layer image coding data encoded in FIG. It is a block diagram of the image coding apparatus of 1st Embodiment. It is a flowchart of a process of the image coding apparatus of 1st Embodiment. It is a structural illustration figure of the output image coding data of 1st, 2nd embodiment. It is a block diagram of the layer separation process part in the image coding apparatus of 2nd Embodiment. It is a flowchart of the layer separation process of 2nd, 4th embodiment. It is structure explanatory drawing when input image data is isolate | separated into multiple layers. It is a block diagram of the image coding apparatus of 3rd Embodiment. It is a structural example figure of the output image coding data of 3rd, 4th embodiment. It is a block diagram of the layer separation process part in the image decoding apparatus of 4th Embodiment.

Explanation of symbols

  0101 region information input device, 0102 image input device, 0103 image encoded data output device, 0104 image encoding device, 0105 region information data buffer, 0106 layer separation information creation processing unit, 0107 input image data buffer, 0108 layer separation information data Buffer, 0109 layer separation processing unit, 0110 first layer image data buffer, 0111 second layer image data buffer, 0112 first layer image encoding processing unit, 0113 second layer image encoding processing unit, 0114 first layer image code Data buffer, 0115 second layer image encoded data buffer, 0116 image encoded data joint processing unit, 0117 output image encoded data buffer, 0201 layer separation information data buffer, 0202 input image data buffer, 0203 first layer image data buffer , 0204 second layer image data buffer, 0205 layer separation processing unit, 0206 first layer resolution conversion processing unit, 0207 second layer resolution conversion processing unit, 0208 first layer resolution conversion image data buffer, 0209 second layer resolution conversion image Data buffer, 0210 pixel value change processing unit, 0301 region information input device, 0302 image input device, 0303 image encoded data output device, 0304 image encoding device, 0305 region information data buffer, 0306 layer separation information creation processing unit, 0307 Input image data buffer, 0308 layer separation information data buffer, 0309 layer separation processing unit, 0310 layer image data buffer group, 0311 layer image encoding processing unit group, 0312 layer image encoded data buffer group, 0313 image encoded data joining processing Part, 0314 out Image encoded data buffer, 0401 layer separation information data buffer, 0402 input image data buffer, 0403 layer image data buffer group, 0404 layer separation processing unit, 0405 layer resolution conversion processing unit group, 0406 layer resolution conversion image data buffer group, 0407 Pixel value change processing unit.

Claims (13)

The input image data is separated into layer image data of N layers (N is a natural number of 2 or more) corresponding to the input image data, and at least one of the layer image data is subjected to function conversion processing in units of small areas. An image encoding apparatus that performs encoding by an image encoding method of a lossless encoding method,
The effective area to be expressed by the n-th layer image data (n is 1 or more and a natural number less than or equal to N. When n is smaller, the effective area is to be expressed in the upper layer). Area information data input processing means for inputting at least identifiable area information data;
Based on the region information data, the layer image data to be encoded by the image encoding method is encoded by the region using the pixel value of the input image data in the small region unit and the image encoding method. When the pixel value pattern before encoding of the flat area in which the pixel value in the small area becomes the same flat value when the decoding process is performed is used as the flattening conversion pattern, encoding is performed using the flattening conversion pattern. A layer separation processing means for determining a region to be generated and generating the n-th layer image data;
Image encoded data creation processing means for encoding the n-th layer image data created by the layer separation processing means by the image coding method and creating n-th layer image coded data;
An image coding apparatus comprising: output data creation processing means for creating output data including at least the n-th layer image coded data created by the image coded data creation processing means. The image encoding device further includes resolution conversion processing means for converting the vertical resolution of the image data and the horizontal resolution,
The small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction.
The layer separation processing unit performs the region determination of the n-th layer image data by multiplying the vertical size of the block by 1 / Xn (Xn is a real number and the vertical resolution of the n-th layer image data). , The size in the horizontal direction is 1 / Yn (Yn is a real number, the resolution in the horizontal direction of the nth layer image data) times, and each is determined in units of regions,
The encoded image data creation processing means, for the area using the input image data, uses the resolution conversion processing means to set the vertical resolution of the corresponding area of the input image data to Xn times and the horizontal resolution to Yn. 2. The image encoding device according to claim 1, wherein the n-th layer image data is generated by using image data converted twice. The small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction.
The layer separation processing means encodes each layer image for encoding the region used for the region determination of the n-th layer image data by the image encoding method corresponding to each block of the n-th layer image data. 2. The image encoding apparatus according to claim 1, wherein determination is made in units of blocks each having a common multiple of a vertical size and a common multiple of a horizontal size of the block of data, each having a vertical size and a horizontal size. The small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction.
The layer separation processing means encodes each layer image for encoding the region used for the region determination of the n-th layer image data by the image encoding method corresponding to each block of the n-th layer image data. 2. The image encoding device according to claim 1, wherein the least common multiple of the vertical size of the block of data and the least common multiple of the horizontal size are determined in units of blocks each having a vertical size and a horizontal size. . The image encoding device further includes resolution conversion processing means for converting the vertical resolution of the image data and the horizontal resolution,
The small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction.
The layer separation processing unit encodes an area used for the area determination of the n-th layer image data by the image encoding method corresponding to each block of the n-th layer image data. Multiplying the vertical size of the block of layer image data by 1 / Xi (Xi is a real number and the resolution in the vertical direction of the i-th layer image data. 1 ≦ i ≦ N and i is a natural number not including n). And the common multiple of the size obtained by multiplying the horizontal size of the block by 1 / Yi (Yi is a real number and the horizontal resolution of the i-th layer image data) times, respectively. Judgment by block size
The encoded image data creation processing means sets the vertical resolution of the corresponding area of the input image data to Xn (Xn is a real number, nth) for the area using the input image data by the resolution conversion processing means. The resolution in the vertical direction of the layer image data) times and the resolution in the horizontal direction to Yn (Yn is a real number, the resolution in the horizontal direction of the nth layer image data) times, The image encoding apparatus according to claim 1, wherein n-layer image data is generated. The image encoding device further includes resolution conversion processing means for converting the vertical resolution of the image data and the horizontal resolution,
The small area is a block composed of a quadrangle surrounded by two straight lines parallel to the vertical direction and two straight lines parallel to the horizontal direction.
The layer separation processing unit encodes an area used for the area determination of the n-th layer image data by the image encoding method corresponding to each block of the n-th layer image data. Multiplying the vertical size of the block of layer image data by 1 / Xi (Xi is a real number and the resolution in the vertical direction of the i-th layer image data. 1 ≦ i ≦ N and i is a natural number not including n). And the least common multiple of the size obtained by multiplying the horizontal size of the block by 1 / Yi (Yi is a real number and the horizontal resolution of the i-th layer image data) times Judgment is made in units of blocks with horizontal size,
The encoded image data creation processing means sets the vertical resolution of the corresponding area of the input image data to Xn (Xn is a real number, nth) for the area using the input image data by the resolution conversion processing means. The resolution in the vertical direction of the layer image data) times and the resolution in the horizontal direction to Yn (Yn is a real number, the resolution in the horizontal direction of the nth layer image data) times, The image encoding apparatus according to claim 1, wherein n-layer image data is generated.   In the nth layer image data, the region of the (n + 1) th layer image data corresponding to the region using the pixel value of the input image data and encoded by the image encoding method is decoded. The image coding apparatus according to claim 1, wherein the flattening conversion pattern that becomes the flat value is used.   This is an irreversible encoding method that performs a function conversion process on the layer image data in units of the small area, and is different from the image encoding method when the small image data is encoded and decoded by a second image encoding method. When the pixel value pattern before encoding of the flat region where the pixel values in the region are the same flat value is used as the second flattening conversion pattern, the pixel value of the input image data is used in the n-th layer image data. 2. The second flattening conversion pattern is used for a region of (n + 1) -th layer image data that corresponds to a region and is encoded by the second image encoding method. The image coding apparatus in any one of -6.   The image coding apparatus according to claim 1, wherein the flat value is a value used as a transparent color in decoding the n-th layer image data.   The image coding apparatus according to claim 9, wherein the output data creation processing means further includes the transparent color value in output data. A computer having an arithmetic processing unit separates the input image data into layer image data of N layers (N is a natural number of 2 or more) corresponding to each of the input image data, of which at least one layer image data is small An image encoding method for performing encoding by an image encoding method of an irreversible encoding method that performs function conversion processing in units of regions,
The area information data input processing means is arranged in the upper layer when the n-th layer image data (n is 1 or more and a natural number less than or equal to N, which is the individual layer image data. ) Inputting region information data in which at least the effective region to be expressed can be specified;
An area using pixel values of the input image data in units of the small area for layer image data encoded by the image encoding method based on the area information data; When the pixel value pattern before encoding of the flat region in which the pixel value in the small region becomes the same flat value when encoded and decoded by the encoding method is used as the flattening conversion pattern, the flattening conversion Determining a region to be encoded using a pattern, and generating the n-th layer image data;
Image encoded data creation processing means encodes the n-th layer image data created by the layer separation processing means by the image coding method, and creates n-th layer image coded data;
An image encoding method comprising: output data generation processing means generating output data including at least the n-th layer image encoded data generated by the image encoded data generation processing means. The input image data is separated into layer image data of N layers (N is a natural number of 2 or more) corresponding to the input image data, and at least one of the layer image data is subjected to function conversion processing in units of small areas. An image encoding program for causing a computer having an arithmetic processing unit to execute an image encoding method for encoding by an image encoding method of a lossless encoding method,
The arithmetic processing unit is an n-th layer image data (n is 1 or more and a natural number less than or equal to N. The n-th layer image data is arranged in the upper layer when n is smaller. Inputting region information data that can identify at least an effective region to be expressed;
An area using pixel values of the input image data in units of the small area for layer image data encoded by the image encoding method based on the area information data, and the image When the pixel value pattern before encoding of the flat region in which the pixel value in the small region becomes the same flat value when encoded and decoded by the encoding method is used as the flattening conversion pattern, the flattening conversion Determining a region to be encoded using a pattern, and generating the n-th layer image data;
The arithmetic processing unit encoding the generated n-th layer image data by the image encoding method to generate n-th layer image encoded data;
An image encoding program that causes a computer to execute a step of generating output data including at least the generated n-th layer image encoded data.   A computer-readable recording medium on which the image encoding program according to claim 12 is recorded.

Images