JP2002083316A - Image deforming device and image deforming method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To employ image deforming methods of different types for desired image deformations. SOLUTION: A modeling section 120 makes a vertex calculation of a polygon unit for an object subject to an image deformation and calculates a pre-filter coefficient according to a reduction ratio. A DDA(digital differential analyzer) 130 sets texture addresses for mapping a raw image for the object and stores them in an address buffer 140 together with the specified information on the selection of the deformation process method. A pre-filter processing section 210 conducts the pre-filtering of image data according to the calculated pre-filter coefficient. The image data applied with the pre-filtering process are stored in a texture memory 230 by a texture memory control 220. Externally generated image data are also stored in the texture memory 230 as required. An interpolation 240 reads the image data of the texture memory 230 based on the texture addresses and employs the deformation process method according to the specified information for an interpolation process.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image deforming apparatus and an image deforming method thereof, and more particularly to an image deforming apparatus and an image deforming method for performing a special effect by performing a deformation process on a predetermined image.

[0002]

2. Description of the Related Art Conventionally, as a method of giving a special effect to an image, a method of performing deformation, such as moving, enlarging, reducing, or rotating an image, on an appropriate original image is known. An image transformation apparatus for processing images has been used.

In an image transformation apparatus, an image pre-filtered in accordance with a reduction ratio of an image is written in an image memory (hereinafter, referred to as a texture memory), and then transformed according to a transformation address of the image (hereinafter, referred to as a texture address). By reading the image data of the neighboring points from the texture memory and interpolating between the neighboring point image data (hereinafter, referred to as interpolation), an image deformation effect (hereinafter, referred to as texture mapping) is realized. Hereinafter, the texture mapping method according to this method is referred to as a pre-filter method.

On the other hand, in the field of computer graphics in recent years, a technique based on a mipmap method has been mainly used for image deformation. In the mipmap method, instead of performing pre-filtering, a reduced image of an image to be deformed in advance, for example, x1, x1 / 2, x1 / 4, x
After writing 1/8 in the texture memory, the image data of the neighboring points are read out from the texture memory and interpolated in accordance with the texture address and the designated data from which scaled image the image data is read. Is performing texture mapping.

[0005] Comparing the interpolation by the prefilter method and the interpolation by the mipmap method, each has the following features. The pre-filter method performs optimal pre-filtering according to the reduction of the image, and thus has a feature that the reduced image has higher image quality than the mipmap method prepared in advance in stages.

The mipmap method has a feature that images having different reduction ratios can be simultaneously displayed in an image of one field since images having different reduction ratios are written in the texture memory in advance. Displaying images at the same time in this manner is called a multi-object. On the other hand, in the pre-filter method, since only pre-filtering can be performed at one reduction rate in an image of one field, multi-object is impossible.

[0007] In recent years, image transformation apparatuses have performed texture mapping by hardware to increase the speed of image processing.

[0008]

However, the conventional image transformation apparatus has a problem that it is difficult to perform a desired image transformation because both the pre-filter system and the mipmap system cannot be mixed. .

As described above, the pre-filter method has a feature that high image quality can be provided, and the mipmap method has a feature that multi-objects are possible.
In order to enhance the special effects of the image, it is desirable that the most appropriate method can be selected for the desired image deformation processing by utilizing these characteristics. For example, in the case of performing a deformation for adding a special effect to a moving image, it is desirable to perform processing using a pre-filter method that provides high image quality. When it is desired to add an image such as a frame to a predetermined image, processing is performed by a mipmap method.

However, in order to realize both the prefilter method and the mipmap method, it is necessary to mount hardware for performing each processing. For this reason, there has been no case where both types are conventionally mixed in one image deforming apparatus.

The present invention has been made in view of the above points, and an object of the present invention is to provide an image deformation apparatus and an image deformation method capable of performing desired image deformation.

[0012]

According to the present invention, in order to solve the above-mentioned problems, in an image deforming apparatus for performing a deforming process on a predetermined image to generate a special effect, an object to be deformed in units of polygons Modeling means for performing a vertex operation of the above, calculating a pre-filter coefficient according to the reduction ratio of the polygon, Texture address setting means for setting a texture address for mapping a predetermined image predetermined to the object, Texture address storage means for storing designation information relating to the selection of the method of the deformation processing together with the texture address in association with the object, and inputting image data of the predetermined image as required, and according to the pre-filter coefficient Pre-filter processing means for performing pre-filter processing; Image data storage means for storing image data processed by the filter processing means and image data generated externally as necessary based on the texture address; and image data processed by the pre-filter processing means. And image data control means for controlling writing of the image data generated externally to the image data storage means and reading of the image data stored in the image data storage means; and Texture transformation means for performing mapping by performing interpolation processing on image data of the image data storage means read out via data control means using a transformation processing method according to the designation information. Device is provided.

In the image deforming apparatus having such a configuration, the modeling means calculates a vertex operation in units of polygons and calculates a pre-filter coefficient corresponding to the polygon reduction rate for the object to be image-deformed. The texture address setting means sets a texture address for mapping a predetermined image, which will be described later, to the object using the polygon generated by the modeling means. To memorize. The designation information and the texture address are stored in association with the object. The pre-filter processing means performs pre-filtering of predetermined image data to be pasted according to the pre-filter coefficient calculated by the modeling means. The pre-filtered image data is stored in the image data storage by the image data controller. Further, the image data control means stores image data generated externally for use in texture mapping in the image data storage means based on the texture address, if necessary. The texture mapping means reads out the image data of the image data storage means via the image data control means based on the texture address stored in the texture address storage means, and performs an interpolation process using a deformation processing method according to the designated information. And perform mapping.

According to another aspect of the present invention, there is provided an image deformation method for performing a deformation process on a predetermined image to generate a special effect. Calculate a pre-filter coefficient according to a reduction ratio, set a texture address for mapping a predetermined image to the object, and store designation information relating to selection of a method of the deformation processing in association with the object together with the texture address. The image data of the predetermined image is input as necessary, and a pre-filter process is performed according to the pre-filter coefficient, and the pre-filtered image data and an image generated externally as necessary. Data based on the texture address and store the data in the texture address. Hazuki, the read pre-filtered image data, image transformation method characterized by comprising the step of performing the mapping and interpolation processing using techniques deformation processing in accordance with the designation information, is provided.

In the image deforming method of such a procedure, a vertex calculation is performed on an object to be deformed in units of polygons and a pre-filter coefficient is calculated according to a polygon reduction ratio. Using the generated polygons, a texture address for mapping a predetermined image to the object as described later is set, and stored together with the designation information on the selection of the deformation processing method. The designation information and the texture address are stored in association with the object.
The image data to be texture-mapped is pre-filtered according to the calculated pre-filter coefficients.
The image data that has been prefiltered in this way is stored in a predetermined storage area. If necessary, image data generated externally for use in texture mapping can be stored. The texture mapping process reads out pre-filtered image data or externally generated image data based on a texture address, performs an interpolation process using a transformation process method according to the designated information, and performs mapping.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram of an image transformation device according to an embodiment of the present invention.

The image transformation apparatus according to the present invention comprises an address processing block 100 for setting a texture address for an object to be transformed, and a video processing block 200 for processing image data.

An address processing block 100 includes a host 110 for controlling the entire apparatus, a modeling section 120 for performing polygon processing and prefilter coefficient calculation, and a DDA (Digital Differential) as texture address setting means for setting a texture address.
An analyzer (hereinafter, referred to as DDA) 130 and an address buffer 140 as a texture address storage unit for storing a texture address.

The host 110 controls the entire apparatus, and sends information such as image deformation (wire frame) data and selection of an image to be texture-mapped for each object to the modeling unit 120.

The modeling unit 120 calculates a vertex of three-dimensional coordinates in units of polygons and calculates a pre-filter coefficient according to a reduction ratio of the polygons for an object to be image-deformed.

The DDA 130 operates on the object
A texture address for mapping a predetermined image is assigned using the polygon calculated by the modeling unit 120, and the texture address is written to the address buffer 140. The predetermined image is an image to be texture-mapped for each object specified via the host 110. In particular,
The polygon data calculated by the modeling unit 120 is converted into pixel data and written into the address buffer 140. In the address buffer 140, a read address for pasting a video image to a polygon is written. This read address is a texture address, and the operation of pasting a video image to a polygon is texture mapping.

The address buffer 140 is a DDA 130
The texture address generated by the above and the specification information on the selection of the deformation processing method are stored in association with the object. Here, the designation information on the selection of the method of the deformation processing is setting information on whether to select the pre-filter method or the mipmap method.

The video processing block 200 is a pre-filter processing section 210 for inputting original image data for performing texture mapping, performing pre-filter processing, and image data control means for controlling reading and writing of image data for performing texture mapping. It comprises a texture memory control 220, a texture memory 230 as image data storage means for storing image data to be subjected to texture mapping, and an interpolation 240 as texture mapping means for performing texture mapping processing.

The pre-filter processing unit 210 receives a video image as an original image on which texture mapping is to be performed, and performs pre-filter processing according to the pre-filter coefficients calculated by the modeling unit 120. The pre-filter processing unit 210 includes an H-direction pre-filter 211 that performs pre-filtering (anti-alias filter) processing in a pixel direction (hereinafter, referred to as an H direction) of an image according to a pre-filter coefficient, and an image scan line direction (hereinafter, V). HV scan conversion 212 for performing HV scan conversion for performing pre-filtering
It comprises a V-direction pre-filter 213 for performing V-direction filtering in accordance with pre-filter coefficients after HV scan conversion, and a field / frame conversion 214 for converting a field image to a frame image for improving image quality.

The texture memory control 220
According to the instruction of the host 110, the pre-filter processing unit 21
The image data prefiltered by 0 is written to the texture memory 230. These are mainly used as pre-filter image data. Further, if necessary, image data such as a title or a pattern created by the host 110 or the like is written in the texture memory 230. These are used as mipmap image data. Further, separately from the image data created by the host or the like, a reduced image created by the pre-filter processing unit 210 can be registered as a mipmap image. Further, the texture memory control 220 reads out the image data stored in the texture memory 230 according to the instruction of the host 110, and executes the interpolation 2
Send to 40.

The texture memory 230 stores prefilter image data and mipmap image data according to the texture memory control 220. The interpolation 240 performs an interpolation process on the image data of the texture memory 230 read out via the texture memory control 220 based on the texture address, and completes the texture mapping according to the deformation of the image. At this time, one of the interpolation processing by the pre-filter method and the interpolation processing by the mip map method is selected according to the specification information stored in association with the object.

The operation of the image deforming apparatus having such a configuration and an image deforming method will be described. First, the operation of the address processing block 100 will be described. In the address processing block 100, information such as image deformation data and selection of an image to be texture-mapped for each object is input from the host 110 to the modeling unit 120. The modeling unit 120 calculates the vertices of the three-dimensional coordinates of the object in units of polygons and calculates the pre-filter coefficient according to the reduction ratio of the polygon. The pre-filter coefficients are used in the pre-filter processing unit 210. DDA1
Reference numeral 30 sets a texture address for performing texture mapping on the object based on the polygon data calculated by the modeling unit 120.
At this time, the interpolation processing method is stored in association with the object together with the texture address.

A specific example will be described. FIG. 2 is a configuration diagram of a texture address space of the image transformation device according to one embodiment of the present invention. SDTV (Standard Def)
Initiation Television) and HDTV
(High Definition Televisi
on)).

The texture address space of the image transformation device according to the present invention is a two-dimensional space of texture addresses U and V. Signal format is SD
In the case of TV, 1024 × 102 per object
It has 4 spaces and can be divided into 16 objects. Since the interpolation by the pre-filter method or the mipmap method can be set for each arbitrary area of the texture address space, the interpolation method of up to 16 objects can be selected. For HDTV, 2 per object
It is possible to select an interpolation method with a space of 048 × 2048 and up to four objects.

The selection of the prefilter method or the mipmap method will be described in an embodiment. FIG. 3 is an example of the image special effect. This indicates the “brick effect”.

As an example of a general method of generating a brick effect, an image of a pre-filter method is texture-mapped on a main surface to emphasize image quality.
Still images such as titles are mapped to the sides (sides) A and B by a mipmap method. In such a case, the objects (here, Main, Side,
A and Side B). The arrangement of texture addresses in such a case will be described. FIG.
It is an example of arrangement of a texture address. (A) is an example of a texture address when the signal format is SDTV, and (b) is a data format of the texture address (U or V). As shown in (a), the texture address space is defined as Main, Side
A region is divided by each object of A and Side B. The U or V data indicating the texture address of each object does not use the upper 2 bits as shown in (b). In the unused upper 2 bits, information for specifying whether the interpolation method is the pre-filter method or the mipmap method is set. Thus, the top two texture addresses
The bits enable identification of the interpolation method, so that the same texture memory block
Prefilter image data and mipmap image data can be arranged. The arrangement of the texture address space can be arbitrarily set.

As described above, the texture address and the designation information are set in the address buffer 140. FIG.
Return to the description. In the video processing block 200, the pre-filter processing unit 210 performs pre-filter processing on the input video image according to the pre-filter coefficient calculated by the modeling unit 120. The prefiltered image data is written to the texture memory 230 via the texture memory control 220. In this way, the pre-filter processing is performed, and the texture memory 23
The image data written to 0 is mainly used for the interpolation process of the pre-filter method.

Next, the writing of the mipmap image data will be described. When texture mapping is mainly performed on a still image such as a title or a pattern or when a multi-object is performed, a mipmap method is used. As the mipmap image, several steps of reduced images such as x1, x1 / 2, x1... Are written in the texture memory 230 in advance. In actual writing, several levels of still images created by the host 110 are stored in the texture memory control 2.
20 to the texture memory 230. In addition to the method of writing still image data from the host 110, image data generated by the above-described pre-filter processing may be written to the texture memory 230 as mipmap image data. In this case, several stages of reduced images are sequentially created while pre-filtering and stored. Therefore, host 1
10, it is possible to create mipmap image data with higher image quality than creating a reduced image.

Subsequently, in the interpolation 240, the image data corresponding to the object is read from the texture memory 230 and texture mapped. The interpolation 240 will be described in more detail. FIG. 5 is a configuration diagram of an interpolation portion of the image transformation device according to one embodiment of the present invention. The same components as those in FIG. 1 are denoted by the same reference numerals, and description thereof is omitted.

As described in the section of "Prior Art", interpolation processing is generally performed by hardware in order to perform high-speed processing. Therefore, an interpolation circuit using a pre-filter method and an interpolation circuit using a mipmap method are prepared.

The interpolation 240 according to the present invention
Are a U- and V-data upper 2-bit decoding circuit 241 for identifying the interpolation method, a pre-filter type interpolation circuit 242 for performing a pre-filter type interpolation process, and a mip-map type interpolation circuit for performing a mip-map type interpolation process. It comprises a poration circuit 243 and a switch 244 as selection means for selecting either the pre-filter type interpolation circuit 242 or the mipmap type interpolation 243 based on the identification result.

U, V data upper 2 bits decode circuit 2
41 is a texture address (U, V) for each object.
An identification flag of the interpolation method added to the unused upper two bits of the (data) area is extracted and output to the switch 244 as flag information of the prefilter method / mipmap method.

The pre-filter type interpolation circuit 242 generates interpolation coefficients for the pre-filter method according to the assigned texture address, and performs interpolation processing.

The mipmap type interpolation circuit 243 generates an interpolation coefficient for the mipmap type according to the assigned texture address, and performs interpolation processing.

The switch 244 selects the output of the pre-filter type interpolation circuit 242 or the output of the mipmap type interpolation circuit 243 according to the identification information of the U, V data upper two-bit decode circuit 241.

The above processing functions can be realized by a computer. In this case, the processing contents of the functions that the image transformation apparatus should have are described in a program recorded on a computer-readable recording medium. Then, by executing this program on a computer, the above processing is realized on the computer. Examples of the computer-readable recording medium include a magnetic recording device and a semiconductor memory. For distribution to the market, CD-ROM (Compact Disc Read Only Memory)
The program may be stored and distributed on a portable recording medium such as a floppy disk or a floppy (registered trademark) disk, or stored in a storage device of a computer connected via a network, and transferred to another computer via the network. it can. When the program is executed by the computer, the program is stored in a hard disk device or the like in the computer, and is loaded into the main memory and executed.

[0042]

As described above, in the image transformation apparatus of the present invention, a predetermined polygon corresponding to an object is generated, a pre-filter coefficient is calculated, and a texture address for mapping an image to the object is set. It is set and stored together with the designation information on the selection of the method of the deformation processing. Pre-filtering of the image data is performed according to the calculated pre-filter coefficient, and the image data is stored in the image data storage unit. Further, image data generated externally is stored in the image data storage means as needed.
Subsequently, the image data in the image data storage means is read out based on the texture address, and interpolation is performed by using a method according to the designated information, and mapping is performed.

As described above, since the designation information on the method of the deformation processing can be added according to the texture address, the method of the deformation processing can be appropriately selected according to the texture address. Further, a texture mapping processing method is appropriately selected according to each method. As a result, different methods such as the pre-filter method and the mipmap method can be mixed on the same image.

According to the image deformation method of the present invention, a predetermined polygon corresponding to an object is generated, a pre-filter coefficient corresponding to a polygon reduction ratio is calculated, and
A texture address for mapping an image to an object is set and stored together with designation information on selection of a deformation processing method. The designation information is stored in association with the object together with the texture address. Pre-filtering of the original image data is performed according to the calculated pre-filter coefficients, and the pre-filtering is stored in a predetermined storage area. Also, image data generated externally can be stored. Subsequently, image data is read out based on the texture address, and interpolation is performed using a method according to the designated information, and mapping is performed.

As described above, since the designation information on the method of the deformation process can be added according to the texture address, the method of the deformation process can be appropriately selected according to the texture address. Further, a texture mapping processing method is appropriately selected according to each method. As a result, different methods such as the pre-filter method and the mipmap method can be mixed on the same image.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image transformation device according to an embodiment of the present invention.

FIG. 2 is a configuration diagram of a texture address space of the image transformation device according to an embodiment of the present invention.

FIG. 3 is an example of an image special effect.

FIG. 4 is a layout example of a texture address;

FIG. 5 is a configuration diagram of an interpolation portion of the image transformation device according to an embodiment of the present invention.

[Explanation of symbols]

100: address processing block, 110: host, 12
0: modeling unit, 130: DDA, 140: address buffer, 200: video processing block, 210: pre-filter processing unit, 211: H-direction pre-filter, 212
HV scan conversion, 213 V direction prefilter, 214 field / frame conversion, 220 texture memory control, 230 texture memory, 240 interpolation

Claims (6)

[Claims] 1. An image transformation apparatus for performing a transformation process on a predetermined image to generate a special effect, wherein a vertex operation is performed on a polygon-by-polygon basis for the object to be transformed, and a pre-processing is performed according to a reduction ratio of the polygon. Modeling means for calculating a filter coefficient; texture address setting means for setting a texture address for mapping a predetermined image predetermined on the object; and designation information on the selection of the deformation processing method together with the texture address. Texture address storage means for storing in association with the object, pre-filter processing means for inputting image data of the predetermined image as necessary, and performing pre-filter processing in accordance with the pre-filter coefficient; Image data processed by the means, and Image data storage means for storing image data generated externally as required based on the texture address; and image data processed by the pre-filter processing means and image data generated externally. Image data control means for controlling writing to the image data storage means and reading of image data stored in the image data storage means; and storing the image data read out via the image data control means based on the texture address. A texture mapping means for performing mapping by interpolating the image data of the means using a method of deformation processing according to the designation information. 2. The designation information stored in the texture address storage means includes, in the image data storage means, an image obtained by pre-filtering a method of a deformation process performed by the texture mapping means in accordance with an image reduction ratio. A pre-filter system for performing image deformation using the same, or a mipmap system for storing images of predetermined different reduction ratios in the image data storage means and performing image deformation using the same. 2. The image transformation apparatus according to claim 1, wherein the selection is made as follows. 3. The pre-filter type interpolation means for reading out image data stored in the image data storage means and interpolating neighboring point image data according to the pre-filter method, wherein the texture mapping means comprises: Mipmap type interpolation means for reading out the image data stored in the data storage means and interpolating neighboring point image data according to the mipmap method; and the pre-processing based on the designation information stored together with the texture address. 3. The image transformation apparatus according to claim 2, further comprising: a selection unit that selects one of a filter type interpolation unit and the mipmap type interpolation. 4. The specification information stored in the texture address storage means together with the texture address is set in an area of the texture address storage area that is unused in storing the texture address. The image transformation device according to claim 1, wherein 5. The image transformation apparatus according to claim 1, wherein the pre-filter processing means generates image data of a predetermined reduction ratio in advance when performing texture mapping of a mipmap method, if necessary. 2. The image deforming apparatus according to claim 1, wherein said means stores image data of different reduction rates generated by said pre-filter processing means in said image data storage means. 6. An image deformation method for generating a special effect by performing a deformation process on a predetermined image, comprising: calculating a vertex in units of polygons for the object to be image-deformed; and a pre-filter coefficient according to a reduction ratio of the polygon. Calculating a texture address for mapping a predetermined image to the object, storing designation information relating to the selection of the deformation processing method in association with the object together with the texture address, and, if necessary, The image data of the image is input and subjected to a pre-filtering process in accordance with the pre-filter coefficient, and the pre-filtered image data and image data generated externally as necessary are based on the texture address. The pre-filter processing based on the texture address. An image transformation method comprising the steps of: reading out the obtained image data and performing mapping by performing interpolation processing using a transformation processing method according to the designation information.