JPH04302381A - High-speed image generating device - Google Patents

Abstract

PURPOSE:To generate an image without a sense of incompatibility compared with a natural image recognized by humans at high speed by utilizing the visual recognizing action of humans, generating only a close observation area at high definition and generating the other non-close observation area at non-high accu racy. CONSTITUTION:In an image generation in computer graphics, a close observation point detecting device 10-16 calculates the direction of the signal line of an observer from the data obtained from an eye-ball position detecting device 10-17 and a head part posture detecting device 10-18 and calculates the close observation point on a monitor. As the area of the monitor screen, there are the method to form an area at the area by the fixed radius system circle or the expanded radius system circle with the close observation point as a center and the method to form the area by the object shape having the close observation point. Further, as the method to classify the definition of rendering, the classifying processing to enlarge successively the size of the picture element unit from the center to the periphery is used, the image is processed and generated in accordance with the close observation area classification and the definition classification and outputted to the monitor screen at high speed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-speed image generation device using a high-definition viewing area display method.

More specifically, the present invention relates to image generation in CG (computer graphics).
It aims to speed up the generation process by utilizing the human visual characteristics, and is suitable for fields that require high-speed image generation, such as real-time simulation using CG, real-time animation, etc., and for regular television. The present invention relates to a high-speed image generation device that can be used for CG for high-definition monitors such as high-definition, which has five times the amount of pixel data.

0003

SUMMARY OF THE INVENTION Generally, generating a high-definition CG image requires calculation processing time, and in particular, it takes an enormous amount of time to display the entire monitor screen in high definition.

[0004] The high-speed image generation device according to the present invention uses human visual characteristics, that is, it first takes a rough view of the whole image and then focuses on a certain part of the object of interest. Utilizing the visual recognition behavior of humans who try to look closely, it generates a high-definition image of only the focused area, and generates non-high-definition but high-speed images of other non-gazing areas.
In addition, it can quickly generate images that are more natural than natural images that humans recognize.

[0005]

[Background Art] Recent advances in CG technology have made it possible to generate realistic images that are no different from natural images, and images that are attractive and powerful that are not found in natural images, but it takes a very long time to generate these images. .

[0006] As a document describing image generation using this type of CG technology, there is a paper reporting that changing the rendering method (wire frame method, flat shading method, etc.) for each object in CG image generation makes it easier to understand the image. However, it does not utilize the gaze area.

[0007] In addition, there is a paper report on compressing and displaying an image other than the region of interest in image processing, but this does not involve the use of CG technology.

[0008]

[Problems to be Solved by the Invention] As described above, research has been conducted on various speeding-up technologies related to image generation, but these are based on speeding up hardware and speeding up rendering. It is not a method to speed up image generation using natural visual recognition by incorporating human visual characteristics.

[0009] In other words, by using the visual recognition behavior of humans, who first take a rough view of the whole picture and then focus on a certain part of the object that interests them to take a closer look, There is a need for a device that can quickly generate images that look more natural than natural images.

[0010] Accordingly, an object of the present invention is to create a high-speed image that makes it possible to generate an image that does not give a sense of discomfort compared to a natural image that humans recognize by utilizing the visual characteristics of humans when generating images using CG technology. The purpose is to provide a generating device.

[0011]

[Means for Solving the Problems] The present invention provides an apparatus for generating an image by computer image processing, which includes a regionalization means for specifying a gaze area on a monitor screen, and a region processing unit for specifying a region of interest on a monitor screen, and and a definition setting means for determining the definition of the target image, and generates a high-definition image for the region of attention, and generates a non-high-definition, high-speed image for other non-gazing regions.

Further, in an embodiment of the present invention, a computer
In image generation in graphics (CG), a computer monitor screen (10-13
) Fixed radius circular shape centered on the point of interest above (Figure 2)
The feature is that the area is divided into a circle using an enlarged radius method (Figure 3) or an object shape with a gaze point (Figure 6). A method of rendering by gradually increasing the size of the image from the central area to the peripheral area (Fig. 7), a method of gradually changing the rendering method from the central area to the peripheral area from a high-definition one to a high-speed one (Fig. 8), or
A high-speed image generation device for computer graphics is constructed using a method (FIG. 9) that combines both of them.

[0013]

[Operation] The high-speed image generation device according to the present invention takes advantage of human visual characteristics, that is, it first takes a rough view of the entire image and then focuses on a certain part of the object that interests the user. Utilizes the visual recognition behavior of humans who try to look closely, and generates high-definition images only for the focused region, and generates non-high-definition but high-speed images for other non-gazing regions, and creates natural images that humans can recognize. It generates images that look natural at high speed compared to the previous version.

[0014]

[Embodiments] As a premise for explaining specific embodiments of the present invention, the principle of the present invention will be explained below.

[0015] The high-speed image generation device according to the present invention uses human visual characteristics, that is, it first takes a rough view of the entire image and then focuses on a certain part of the object of interest. Utilizing the visual recognition behavior of humans who try to look closely, it generates a high-definition image of only the focused area, and generates non-high-definition but high-speed images of other non-gazing areas.
In addition, it can quickly generate images that are more natural than natural images that humans recognize.

First, a method of dividing the gaze area will be explained.

[0017] As shown in FIG. 1, a method is used in which the region is created in a circle centered on the point of interest, or a method is used in which the region is created in the shape of an object where the point of interest is located, as shown in FIG.

The method of forming a circular region around the point of gaze shown in FIG. 1 is a method that utilizes a circular region of gaze that is suitable for human visual characteristics. The size of the circular area can be efficiently selected from the following two methods according to the movement of the line of sight.

■Fixed radius method As shown in FIG. 2, this method creates a circular area with a predetermined radius regardless of the length of the gaze time, and is used when there is a lot of movement of the line of sight.

■ Expanding radius method As shown in FIG. 3, this is a method in which the radius of the circle is expanded as the gaze time increases, and is used during gaze when the movement of the line of sight is small.

Furthermore, as a method for rendering a circular area (coloring each pixel on the monitor screen), a scanning method shown in FIG. 4 and a spiral scanning method shown in FIG. 5 are used. This spiral scan method is based on the enlarged radius method described above (Fig. 3
) suitable for

Next, a method (FIG. 6) of dividing the object shape into regions having a gaze point will be explained.

As shown in (2) of FIG. 2, when there are multiple objects close to each other in the front and back direction, even if the objects are within a circular range centered on the point of gaze, objects other than the object of interest may be ignored. Considering that high-definition display may actually make it difficult to recognize, only the object shape at which the gaze point is located is displayed in high-definition, as shown in FIG. This utilizes the fact that image data used for CG image generation is hierarchically related to each object.

Next, a method for classifying the degree of definition will be explained.

After dividing the gaze area by the above method, the definition of each area is divided as follows.

1) As shown in FIG. 7, the method uses a method that distinguishes the size of each pixel.

A single rendering method is used, but the size of each pixel is small in a high-definition area near the point of interest (for example, one pixel is taken as a pixel unit), and the size of each pixel becomes smaller as the distance from the point of interest increases. A method is used in which the number of pixels is increased (for example, 4, 9, or 16 pixels are used as a pixel unit).

2) As shown in FIG. 8, the classification is performed using a rendering (coloring) method.

A high-definition rendering method is used for the region of attention, and a relatively low-definition but high-speed rendering method is used for the non-gazing region.

[0030] High-definition rendering methods include the ray tracing method, radiosity method, Phong shading method, etc., and high-speed rendering methods include:
There are Gouraud shading methods, flat shading methods, etc.

3) A method combining the above 1) and 2) is used. That is, as shown in FIG. 9, it depends on the combination of the size of the pixel unit and the rendering method.

[0032] This is a method that aims at new effects by combining the methods 1) and 2) described above.

In other words, it is possible to prevent gaps in coloring caused by different rendering methods inside and outside the gaze area, and it is also effective to make the image generation processing time per area the same at each location on the monitor screen. , images can be generated and displayed naturally.

Next, an embodiment of the present invention will be described with reference to FIG. 10, FIG. 11, and FIG. 12. Here, FIG.
A block diagram showing an example of the entire high-speed image generation device using the attention area high-definition display method, and FIGS. 11 and 12 are flowcharts showing the procedure of high-speed image generation using the attention area high-definition display method.

In the block diagram shown in FIG.
-1 is a ROM, which stores computer start-up processing programs and the like.

A RAM 10-2 is used as a work area for programs, data processing, etc.

10-3 is an external interface;
An external interface such as Ethernet or RS232C is used.

10-4 is a CPU (central processing unit of a computer), which performs calculation processing of the gaze area, rendering (coloring) calculation processing, and the like.

10-5 is an internal interface;
An internal interface such as VME Bus is used.

A display list buffer 10-6 stores a display program generated by the CPU.

A display processor 10-7 interprets the contents of the display program and performs the following raster and text processing.

A graphics processor 10-8 performs three-dimensional image generation processing such as hidden line and hidden surface processing.

A frame buffer 10-9 uses a two-dimensional memory to store color information for each point on the display.

A display controller 10-10 reads out the contents of the frame buffer and controls the video signal in accordance with the synchronization signal of the display.

Reference numeral 10-11 is a color lookup table in which a table of colors used in the display is stored.

A D/A converter 10-12 converts the video signal from digital to analog.

10-13 is a color display;
Outputs the result of calculation processing by the CPU.

A keyboard 10-14 is used to input letters, numbers, and the like.

Reference numeral 10-15 is a mouse, through which coordinate information is input by hand movements.

[0050] Reference numeral 10-16 is a gaze point detection device, which uses eyeball position data obtained from the eyeball position detection device 10-17 and head center position and inclination data obtained from the head posture detection device 10-18. , the direction of the observer's line of sight is calculated, and the position of the gaze point on the monitor screen is calculated.

Reference numeral 10-17 is an eyeball position detecting device which irradiates the eye with infrared rays or the like and detects the position of the eyeball based on the difference in reflectance from the white and iris of the eye.

Reference numeral 10-18 is a head posture detection device, which has an ultrasonic sensor attached to the head and detects the center position, inclination, etc. of the head.

Next, the operation of this embodiment will be explained.

1) An observer equipped with an eyeball position detection device 10-17 and a head posture detection device 10-18 watches the monitor screen (step S1).

2) The eyeball position detection device 10-17 irradiates the eyes with infrared rays, etc., and detects the position of the eyeballs based on the difference in reflectance from the whites of the eyes and the iris (step S2).
.

3) The head posture detection device 10-18 detects the center position and inclination of the head using an ultrasonic sensor attached to the head (step S3).

4) The gaze point detection device 10-16 uses the eyeball position data obtained from the eyeball position detection device 10-17 and the head center position and inclination data obtained from the head posture detection device 10-18. The direction of the observer's line of sight is calculated, and the position of the gaze point on the monitor screen is calculated (step S4).

5) Select an appropriate method from the following three types of gaze point region segmentation processing to segment the monitor screen.

■Method of creating a circular region using a fixed radius method centered on the gaze point (FIG. 2) (Steps S5, S7, S8
).

[0060] ■ A method of creating a circular area using an enlarged radius method centered on the gaze point (Fig. 3) (Steps S5, S7, S9)
).

[0061] ① Method of regionalizing based on the object shape where the gaze point is located (FIG. 6) (steps S6, S10).

6) Select an appropriate method from the following three types of definition classification processing to classify the rendering of the monitor screen.

■Method of dividing the size of each pixel (Fig. 7
) (Step S11).

■ A method of classification using a rendering (coloring) method (FIG. 8) (step S12).

■ Combination of pixel unit size and rendering method (FIG. 9) (step S13).

7) An image is generated according to the above-described gaze area classification processing and definition classification processing, and is output to the monitor screen (step S14).

[0067] The above-described high-definition viewing area display method allows image generation to be performed at high speed.

[0068]

Effects of the Invention The present invention aims to speed up the image generation process in CG by utilizing human visual characteristics, and is suitable for fields that require high-speed image generation, such as CG. Prototype production of CG software can improve work efficiency by using it for real-time simulations, real-time animation, etc., and for CG for high-definition monitors such as high-definition monitors, which have five times the amount of pixel data as regular televisions. As a result, it is easy to do and the efficiency of checking the contents is increased.
It is possible to create more attractive CG software.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of a method of dividing a screen into circular regions centered on a point of interest within a monitor screen.

[Figure 2] A fixed radius circular area centered on the point of interest (time elapses from left to right, but even if the point of interest (star mark in the figure) moves, high resolution is achieved. FIG.

[Figure 3] A method of creating a circular region using an expanding radius method centered on the gaze point (as time elapses from left to right, the gaze point (
When the star mark in the figure) does not move much, as shown in the figure,
FIG. 3 is an explanatory diagram of the range of the high-definition area expanding with the gaze time.

FIG. 4 is a diagram showing a horizontal method employed on a normal monitor screen as a scanning method within a circular area.

[Figure 5] Shows a spiral scanning method as a scanning method within a circular area (in the case of the enlarged radius method in Figure 3, it is possible to speed up image generation because there is no need to recalculate pixels that have already been rendered). It is a diagram.

FIG. 6 is an explanatory diagram of a method of regionalizing based on the shape of an object with a gaze point (a method that utilizes the fact that image data has a hierarchical structure for each object).

[Figure 7] As a method of dividing the definition, it is divided by the size of the pixel unit (the pixel unit is made smaller toward the center, and the pixel unit becomes larger toward the periphery.The smaller the pixel unit is, the longer the rendering calculation time is. This is an explanatory diagram of shortening the image generation time for the entire monitor screen by utilizing this fact.

[Figure 8] As a method of classifying the definition, there is a method of dividing by different rendering (coloring) methods (high-definition rendering method (longer image generation time) towards the center, faster rendering method (lower definition) towards the periphery. FIG. 3 is an explanatory diagram of how the image generation time for the entire monitor screen can be shortened by using this method.

[Figure 9] As a method of dividing the definition, a method of dividing by combining the method based on pixel size in Figure 7 and the high-definition rendering method in Figure 8 (the same rendering method is used inside and outside the gaze area, so there is no gap In addition, since the pixel unit becomes larger toward the periphery, the image generation speed can be increased, thereby shortening the image generation time for the entire monitor screen.

FIG. 10 is a block diagram showing an embodiment of the present invention.

FIG. 11 is a flowchart showing the processing procedure of this embodiment.

FIG. 12 is a flowchart showing the processing procedure of this embodiment.

[Explanation of symbols]

10-3 External interface 10-4 CPU 10-13 Color display 10-16 Gaze point detection device 10-17 Eyeball position detection device 10-18 Head posture detection device

Claims (1)

[Claims] Claim 1: A device that generates an image by computer image processing, comprising a region forming means for specifying a region of interest on a monitor screen, and a degree of definition of an image to be processed in the inner and outer regions of the region-formed division. 1. A high-speed image generation device, comprising: definition setting means, which generates a high-definition image for the region of attention, and generates a high-speed, non-high-definition image for other non-gazing regions.