JPH0239379A - Image deformation processing method - Google Patents

Abstract

PURPOSE:To easily generate a desired image by obtaining the desired image by a deformation processing of a fundamental image by changing a shape and the position relation of each partial image by a macro-deformation instruction, with respect to the fundamental image constituted by combining the partial images. CONSTITUTION:For instance, in the case of an arm, a partial image 8 is constituted of each unit image 9 of an arm, a hand, a finger, etc. The unit image being shape data fundamentally such as a sphere and a cube, etc. required for assembling a fundamental image is taken out from a unit image data base, and can be put together by performing operation is movement, rotation, enlargement and reduction, based on connecting information. In most of cases, this fundamental image is different from an operator's desired image. Therefore, in a deformation processing part 5, the fundamental image is brought to deformation processing to a pose intended by the operator. A deformation instruction is a macro-instruction consisting of a deformation image instructing part for designating the fundamental image to be deformed, a deformation instructing part for designating what kind of deformation is executed, and an object image instructing part for designating an object which follows up a shape, and inputted from a command input part 1.

Description

[Detailed description of the invention] (Technical field to which the invention pertains) The present invention is CG (computer graphics).

This invention relates to a method for transforming images that are being created or have already been created in a CAD (Computer Aided Design) 9 image database or the like.

(Prior art) A conventional image transformation method involves manually sequentially performing operations such as moving, rotating twice, and enlarging/reducing the individual partial images that make up one image using an input device such as a mouse or keyboard. there were. However, in this method, when an image consists of many partial images, it is difficult to specify the partial image to be manipulated.
It is necessary to manually create instructions for the operation content, specify the amount of operation, etc. for each partial image one by one, so the number of input operations to the computer is enormous, and there is a drawback that it requires a great deal of effort and time. Ta.

Further, this work requires patience and is painful for the operator, and is also tiring because the operator must stare at the display for a long time.

Also, as a way to solve such shortcomings.

There is a method in which partial image transformation processing is registered in advance as an editing command and the command is sequentially called (metry-driven animation, NCL86-18). However, although this method is effective for deforming a single partial image, if multiple partial images are to be transformed while interacting with each other, the user must be aware of their relative relationships when deforming the images. This adjustment work requires a great deal of time and effort, which is a drawback.

(Object of the Invention) An object of the present invention is to add motion-related data to partial images in advance, in order to solve the drawbacks of conventional image transformation methods such as CG and CAD, and to The computer autonomously transforms multiple partial images using only macro transformation commands, reducing the burden on the operator.
To enable even a person who is not good at drawing to easily create a desired image.

(Structure of the Invention) (Characteristics of the Invention and Differences from the Prior Art) The present invention transforms a basic image composed of a combination of partial images into shapes and shapes of individual partial images using macro deformation commands.
It is characterized by changing the positional relationship and obtaining a desired image by deforming the basic image. The conventional technology involves instructing the transformation of an image using natural language transformation commands, and
The difference is that the interaction between multiple basic images is also automatically controlled by a computer.

(Embodiment) FIG. 1 shows a hardware configuration diagram of an embodiment of the present invention, and in the figure, 1 is a command input section.

2 is a system control unit, 3 is a basic image generation unit, 4 is a partial image dictionary unit, 5 is a transformation processing unit, 6 is a transformation instruction dictionary F unit, and 7 is a display unit, which performs the processing shown in FIG. Work according to flow. The details of the processing will be described below.

First, a command to create a basic image necessary for constructing a desired image is inputted from the command input section 1.

The basic image generation section 3 is activated via the system control section 2. The basic image generation section 3 generates a basic image by combining partial images made up of unit images stored in the partial image dictionary section 4 according to basic image generation information registered in advance.

FIG. 3 is an example of a basic image of a "person" composed of partial images, and is composed of partial images 8 of a head, torso, arms, nine legs, and each partial image 8 is, for example, an arm, an arm, It is composed of unit images 9 of hands, fingers, etc. Specifically, it is extracted from the Ilt image unit image database, which is basically shape data of spheres, cubes, etc. necessary to assemble the basic image, and is moved and enlarged/reduced based on the connection information. It is assembled by applying

This basic image has the shape of a predetermined pose, and is often different from the image desired by the operator. Therefore, the transformation processing unit 5 transforms the basic image into the pose intended by the operator.

The transformation command in the present invention is a macro consisting of a transformed image instruction section that specifies the basic image to be transformed, a transformation instruction section that specifies what kind of transformation is to be performed, and a target image instruction section that specifies the object associated with the shape. It is a command, and the command human resources department 1
(for example, a keyboard). Macro commands mainly focus on action commands that involve objects, such as "grab -1,""sit on -1,""shake hands with -1," and the like.

However, it goes without saying that it is also possible to transform operation commands that do not involve a target, such as "run,""bend," and "bow."

The modification command dictionary section 6 outputs a driving data set for controlling the specific movement of each partial image forming the basic image to the modification processing section 5 from the macro modification commands. This deformation processing section 5
converts this drive data set and basic image data into a plurality of micro-deformation instructions, calculates deformation information based on the shape data, and executes a predetermined deformation process. and,
Repeat until all transformation processing is completed.

Note that the basic image is a collection of partial images, and has the movable conditions of the partial images as an attribute. Further, a partial image is a collection of unit images, and has the movable condition of the partial image itself as an attribute. in particular.

Based on connection information between partial images, a basic image is assembled by performing a transformation operation on the unit images that constitute the partial images. Accordingly, shape data such as the coordinate values and size of the center of each partial image are updated in accordance with the deformation operation during assembly.

Figure 4 shows an example of image generation of a human hand grasping a ball.The transformation command "grasp the ball with your hand" is given as the basic image, "hand" as the basic image to be transformed, and "ball" as the target image. It is currently being executed. The transformation processing flow in this case is shown in FIG. First, move your hand in the direction of the target object, calculate the distance r between "1 hand" and "ball" based on the shape data of this target object and the shape data of the basic image,
Move the “hand” to the “ball” until the distance is below a certain threshold
move it to Next, rotate the thumb around the first joint until it touches the ball. Further, rotation is performed around the second joint until the tip of the rm finger touches. Similarly, rotation processing is performed for all fingers.

In this way, with this method, the hand can automatically grasp the target ball according to the size and position of the ball, which is easier to operate than using a pointing device such as a mouse to instruct the shape one by one. Good.

The flow shown in Figure 5 is a specific processing flow for the verb "grasp" (transformation command), but the specific processing flow for such a macro transformation command is similar to the transformation command dictionary section in Figure 1. 6, the basic image can be freely transformed within the deformable area of this storage section 6, and the degree of deformation is calculated based on the shape data of the target object in each processing flow. Therefore, an image with a desired pose can be obtained regardless of the object.

Figure 6 shows the basics for the same verb "grasp" (transformation command) depending on the type and shape of the object, such as (a) a pen, (b) a tennis ball, and (c) a Bing Pong ball. This shows that the transformation process for "one move," which is an image, changes. In this case, for the verb "grasp", in addition to the drive data set that controls the normal movement, multiple drive data sets such as "grasp with two fingers" are prepared, and the basic image of the object is created. Based on the shape data, an optimal drive data set may be selected to perform the deformation process.

The macro deformation command is called by the deformation command dictionary section 6 and is finally replaced with operations such as movement of the unit image by nine rotations, enlargement/reduction, etc.

FIG. 7 is an example of a table in which a driving data set is called from the verb "tsukame". If the shape data of the target image is, for example, "height 109 width 1. slope 60", rule 1 in the table is applied and driving data set 1 is created.
is selected, and a transformation process is executed in which the object is rotated with one hand so as to be equal to the inclination of the object and then grasped.

By the way, in the above explanation, the action of "grabbing" something with one hand was explained, but when generating an image of a person sitting on a chair, first, using the partial image dictionary unit 4, the action of "grabbing" something is explained.
A basic image of a ``chair'' is generated, and the ``person'' is moved to the ``chair'' until the distance between the ``person'' and the ``chair'' becomes less than a certain threshold.

Next, bend your knees until your hips or feet touch the chair. If the chair has a backrest, the transformation process is performed by selecting a drive data set that rotates the chair around the waist until the "torso (back)" comes into contact.

Similarly, the present invention can be applied to all kinds of actions involving a single person or an object such as an animal, such as "hitting", "kicking", "opening the second section door", etc.

In addition, although the above explanation deals with the case where the target image is not transformed, it is also possible to transform the target image at the same time as the basic image.
The present invention is applicable. For example, when generating an image of people shaking hands, the ``person'' and the target ``person'' are alternately moved until the distance is less than a certain threshold, and then Stretch your "arm" forward, and when the distance between your "hand" and your opponent's "hand" becomes less than a certain threshold, move it a certain distance towards the position of your opponent's "1 hand" until they make contact. Perform this transformation alternately. And the opponent's 1
Bend your fingers until they touch your hand.

Further, there may be a large number of basic images. Furthermore, although a command format is assumed as the input method, it is also possible to input scenario-like information using linguistic information, or input rough sketch information using a tablet or the like as an input device.

(Effects of the Invention) As explained above, according to the present invention, even when the base image and the target image interact with each other to perform transformation processing, the computer automatically adjusts the degree of transformation. , processing time and labor can be significantly reduced.

[Brief explanation of the drawing]

FIG. 1 is a hardware configuration diagram according to an embodiment of the present invention;
Figure 2 is a processing flow diagram of Figure 1, Figure 3 is an explanatory diagram of configuring a basic image from partial images, Figure 4 is a diagram showing transformation processing of a basic image, and Figure 5 is a processing flow diagram of transformation commands. , FIG. 6 is a diagram explaining that the deformation process differs depending on the object,
FIG. 7 is an explanatory diagram of calling a driving data set from a macro transformation instruction. 1... Command input section, 2... System control section,
3... Basic image generation section, 4... Partial double image dictionary section, 5... Transformation processing section, Command dictionary section, 7... Display section. 6...Deformation

Claims (1)

[Claims] When creating an image using a computer or the like, data regarding the shapes of a large number of partial images and data representing characteristics regarding the movement of the partial images are stored in advance in a predetermined memory,
A basic image is created by one or a combination of the partial images, and the basic image is transformed by changing the positional relationship of each partial image that constitutes the basic image in accordance with a macro transformation command. When a transformation command involves a target object, the basic image is created by selecting or combining predetermined data from among a large number of movement-related data stored in the transformation command dictionary according to the shape, size, and position of the target image. An image deformation processing method characterized in that a desired image is created by controlling the degree of deformation of the image.