CN110083818A - A kind of system according to spatial term animation based on tool body cognition - Google Patents
A kind of system according to spatial term animation based on tool body cognition Download PDFInfo
- Publication number
- CN110083818A CN110083818A CN201810078217.2A CN201810078217A CN110083818A CN 110083818 A CN110083818 A CN 110083818A CN 201810078217 A CN201810078217 A CN 201810078217A CN 110083818 A CN110083818 A CN 110083818A
- Authority
- CN
- China
- Prior art keywords
- semantic
- module
- path
- spatial
- relationship
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T13/00—Animation
- G06T13/20—3D [Three Dimensional] animation
Abstract
The invention discloses a kind of systems according to spatial term animation based on tool body cognition, handle semantics comprehension on natural language using tool body, simulation thought, the conversion of natural language to animation is completed under the premise of supporting flexibly abundant semantic.Its technical solution are as follows: the computer program that the sentence of specific spatial term is converted into animation is realized to the function according to spatial term animation by each system module in this distributed semantic knowledge base of the distributed semantic knowledge base that designs according to semantic type, building and based on these system modules.
Description
Technical field
The present invention relates to the technologies according to spatial term animation, more particularly to by using tool body, simulation think of
Want to handle semantics comprehension on natural language to according to the system of spatial term animation.
Background technique
It is widely applied scene according to the system of spatial term animation, the production in addition to being applied to animation films and television programs
Or earlier production, it can be also used for production game, assisting children education.In addition, this system significantly reduce production it is above-mentioned
The professional threshold of product enables more people to enjoy the enjoyment of creation.
Regrettably, most existing systems can only be according to spatial term static images, or generate specific field
The dynamic animation of scape, such as the simulation of traffic accident.There are three major defects for these systems: closing in the space between 1. main process objects
System, it is difficult to the various movements of expressive object；2. system itself lacks scalability, the picture or animation that can be generated extremely have
Limit, and be difficult to be customized according to the specific demand of user；3. the language of input is the minimum subset of natural language, only consider
Several fixed forms, the semanteme that can be expressed are extremely limited.
And for the correlation theory that natural language is converted animation by support, solution is needed most, is also the largest
Difficulty is semantic understanding, and specific manifestation is: the problem of we are without processing well " hiding semanteme " at present.Currently without assorted
Model can be handled well, such as " I drive turn right, then turn right " illustrate " I drive u-turn/toward revolution " in this way
Semantic paragraph.
Both sides reason causes above-mentioned phenomenon:
First, existing semantic understanding research largely ignores the important function acted in natural language,
And it is primarily focused on the generic relation between process object, the attribute of object or the static state of object.Since there are this
The research of sample is oriented to, and existing some semantic representation methods are not also capable of the semanteme of processing movement well, such as traditional
Formal semantics.
Second, main stream approach has ignored the importance of simulation at present, and this simulation exactly has what body cognition was emphasized.People
Class, which comprehends the meaning of a sentence, to be needed to simulate, this phenomenon can be more accurately described as: for the mankind, being understood a certain specific
The brain region activated when sentence and we actually accomplish the big brain area that movement/behavior that this sentence is characterized is activated
Domain is identical.Many cognition neural researchs all demonstrate this point.And what the Study on Semantic of current mainstream was substantially studied
It is still the relationship between symbol, is not deep into really on the counterpart of objective world represented by symbol.
The method for improving above-mentioned two situations both points to the tool body doctrine of cognitive science.Simulation semantic
Exactly combine the school of the cognition of tool body, computer science and Cognitive Neuroscience.But it is established in structure formula grammer and logical semantics
The fact that on theory, weakens the theoretical attainable height of institute when realizing semantic understanding of this school.
In this school, in terms of the understanding for realizing grounded verb, for describing object in the shifting of physical space
Dynamic and metaphor X-scheme is outstanding.But there is still significant limitations by X-scheme, are in particular in: 1.X-
Scheme can carry out certain extension by metaphor, but since this is one for particular task --- object is in physical space
It is mobile --- model, so it can not be applied to other movements or relationship.2. including generation in the structure of X-scheme
The state on boundary and Agent, but these information and close structure be coupled, it is difficult to be used for other relationships.3. compared to people
Class people knows the modularization idea of system, and X-scheme can independently solve specific problem, but this is but also X-schema is difficult to
It incorporates in bigger model.
Summary of the invention
A brief summary of one or more aspects is given below to provide to the basic comprehension in terms of these.This general introduction is not
The extensive overview of all aspects contemplated, and be both not intended to identify critical or decisive element in all aspects also non-
Attempt to define the range in terms of any or all.Its unique purpose is to provide the one of one or more aspects in simplified form
A little concepts are with the sequence for more detailed description given later.
The purpose of the present invention is to solve the above problem, provide it is a kind of based on tool body cognition according to spatial term
The system of animation handles semantics comprehension on natural language using having body, simulation thought, before supporting flexibly semanteme abundant
It puts and completes the conversion of natural language to animation.
The technical solution of the present invention is as follows: present invention discloses a kind of based on tool body cognition according to spatial term animation
System, including semantic knowledge-base, system module group and a computer program, in which:
Semantic knowledge-base is divided into multiple semantic word banks according to semantic type, and each semanteme word bank includes that one kind is semantic simultaneously
Information including the project in a corresponding semantic domain, each semanteme word bank includes the corresponding table of index/semanteme, semantic content；
System module group, is established on semantic knowledge-base, including multiple system modules, wherein each semanteme word bank corresponding one
A individual system module, the system module corresponding to semantic word bank realize the semanteme in corresponding semantic word bank, others system
Module of uniting completes pretreatment to the natural language speech of input, sequence of movement control, 3D model management and to 3D engine
Encapsulation；
Computer program is turned specific natural language sentence generated by controlling the interaction between each system module
Turn to animation, including simple sentence reasoning process and simple sentence flash demo process.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, semanteme are known
The semantic word bank for knowing library includes broad sense Action Semantic library, narrow sense Action Semantic library, narrow sense movement-spatiality semantic base, muscle
Control semantic base, spatial relation semantics library, static relation semantic base, path semantic base, space path semantic base.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, composition system
Each system module of system module group includes and the one-to-one path module of each semantic word bank, broad sense action module, narrow sense
Movement-spatiality module, narrow sense action module, muscle control module, space path module, spatial relationship module, static pass
It is module and time/event module, frame/input module, object module, analog module.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, narrow sense are dynamic
The content for making the narrow sense Action Semantic of semantic base is stored in the muscle control semantic base and narrow sense movement-spatiality semanteme of lower layer
In library, narrow sense Action Semantic refers to the semanteme that will lead to object space state change, and wherein object space state change includes phase
For some object or relative to objective world itself, narrow sense Action Semantic is therein by a quadruple notation
Caused by four elements respectively indicate character string for activating object, the object of initiation movement, the object of receiving movement, movement
The variation of spatiality.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, space is closed
It is the semanteme that semantic base is description spatial relationship, spatial relation semantics indicate that five elements are language respectively by a five-tuple
Justice index, in spatial relationship as the object of object of reference, in spatial relationship as the object of target, for judging object of reference
Whether object and target object have the discriminant function of particular space relationship, under the premise of object of reference object has existed
Building one and object of reference object have the constructor of the target object of particular space relationship.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, static state are closed
It is the semanteme that semantic base is description static relation, static relation includes indicating that an object is present in the specific position in space
There are relationship, there are the semantemes of relationship by a four-tuple mark, and four elements are semantic index respectively, in spatial relationship
It is middle all to be deposited as the object of object of reference, in spatial relationship as the empty object of target, in object of reference object and target empty object
Under the premise of make object of reference object position be equal to target empty object position constructor.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, space road
Diameter semantic base describes the semanteme of space path, and the semanteme for describing space path includes semanteme for constructing a space path,
Wherein space path describes the spatial position on an object formerly latter two time point, passes through a quadruple notation space
Path is semantic, wherein four elements respectively indicate semantic indexing, moved on path object, the target empty pair as terminal
As, the set in path that is made of initial position and final position.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, path language
Yi Ku includes the semanteme of description narrow sense path of motion, indicates narrow sense path of motion by a triple, wherein three elements point
Not Biao Shi semantic indexing, action subject, adjacency matrix, wherein adjacency matrix include movement be related to model each part,
And variation of these parts during the entire process of acting execution relative to last moment spatial position.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, path mould
Block is used to handle the animation control of snap action, narrow sense movement and space path.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, the time/
Event module is used to record the sequencing of different movements, and the execution of animation is controlled during animation simulation.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, frame/
Input module is for pre-processing the natural language for the system that is input to, wherein pretreated content includes subordinate sentence, extracts spy
Determine object, participle.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, object mould
Block is for the relationship between recording indexes and 3D object, and by calling analog module to be loaded into specific 3D model, generation is used for
Control the object of the 3D model.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, simulates mould
Block is packaged the part needed in 3D engine for system, is supplied to the control of system other modules to 3D object, scene
The filling of object involved in method and responsible relationship, movement.
One embodiment of the system according to spatial term animation according to the present invention based on tool body cognition, computer
Program is run in systems to execute following steps:
Step 1: the sentence of natural language being split, object analysis and word segmentation processing, pushed away into subsequent sentence justice
The reason stage；
Step 2: untreated index is judged whether there is, if thening follow the steps 3, if without thening follow the steps 12；
Step 3: the current index of judgement is object or broad sense movement, if object thens follow the steps 4, if broad sense acts
Then follow the steps 5；
Step 4: the corresponding objects in activation analog module fill second object of relationship/movement, return step 2；
Step 5: judging type of action, static relation thens follow the steps 6 if type of action is building, if type of action is structure
It builds spatial relationship and thens follow the steps 7, narrow sense movement thens follow the steps 8 if type of action is building, if type of action is that building is empty
Between path then follow the steps 9；
Step 6: building particular static relationship continues to execute step 10；
Step 7: building particular space relationship continues to execute step 11；
Step 8: constructing specific narrow sense movement, continue to execute step 10；
Step 9: building space path continues to execute step 10；
Step 10: constructing corresponding basis path BasePath in path module；
Step 11: filling first object of relationship/movement, return step 2；
Step 12: each path is reconstructed in path module, and sequence of movement is then reconstructed in time/event module, it is complete
At simple sentence reasoning；
Step 13: simple sentence flash demo is realized in completing simple sentence reasoning process.
The present invention, which compares the prior art, to be had following the utility model has the advantages that the present invention passes through the distribution that designs according to semantic type
Semantic knowledge-base is constructed each system module in this distributed semantic knowledge base and is incited somebody to action based on these system modules
The program that the sentence of specific spatial term is converted into animation realizes the function of spatial term animation.
Compared to existing system, main advantage of the invention has can convert animation for natural language at following 4 points: 1.,
Rather than just still image；2. system itself has powerful scalability, one side user can simply add oneself and need
The semanteme wanted, another aspect system can be used as the basis of more complicated semantic understanding system；3. the natural language of input is more clever
It is living, it is smaller to the limitation of user；4. having used advanced theoretical model, richer flexible semanteme can be indicated ensure that
Under the premise of, use model more reliable in engineering.
Detailed description of the invention
After the detailed description for reading embodiment of the disclosure in conjunction with the following drawings, it better understood when of the invention
Features described above and advantage.In the accompanying drawings, each component is not necessarily drawn to scale, and has similar correlation properties or feature
Component may have same or similar appended drawing reference.
Fig. 1 shows an embodiment of the system according to spatial term animation of the invention based on tool body cognition
Schematic diagram.
Fig. 2 shows in an embodiment of the system according to spatial term animation of the invention based on tool body cognition
Computer program operational flow diagram.
Fig. 3 shows the flow diagram using system of the invention.
Specific embodiment
Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.Note that below in conjunction with attached drawing and specifically real
The aspects for applying example description is merely exemplary, and is understood not to carry out any restrictions to protection scope of the present invention.
Before describing system principle, some technical terms being related to are defined first.
It is semantic to define 1:Itself it is also an object for the specific operation of special object, can be used as other semantemes
Operation object.
In program realization, as soon as each semanteme is a class, so that semanteme is played a role and need to carry out example to this class
Change.This class includes the object that one or two need to fill, and acts on some operations on these objects.These
Operation includes but is not limited to: building object changes the state of the inside of object, deletes object.
Define 2 semantic domains:The set of similar semantic.
In program realization, each semantic domain is a class.Semantic domain is by system example.Semantic domain is provided to language
The building and delete operation of adopted object.(operation for defining in 1 " changing object internal state " is provided by object itself) semantic domain
An index and semantic mapping table are also typically included.The content of this table can be used as data and be stored in outside program
Portion can also write direct in program.It is specifically chosen which kind of method is mainly determined by the semantic quantity for being included in target semantic domain
It is fixed.If quantity is few, so that it may consider inside write-in program.If quantity is more, it is stored in outside, in semantic domain instantiation
When be loaded into.
Define 3 entity objects: represent the object of a specific 3D entity.
In a program, 3D entity is provided or is constructed by the 3D engine that program uses, including but not limited to 3D model, 2D mould
Type, node object etc..
Define 4 empty objects: also referred to as PlaceHolder represents one and only contains the 3D object of volume and location information.
In a program, the object implementatio8 for the node type that empty object is provided by engine, be it is sightless, be not directed to specific
3D object.
Define 5 activation: keep target semantic or the operation of entity object instantiation.
This definition is realized for program.
Define 6 indexes: it is usually a character string, index is used to activate certain semantic or entity object, and index itself can
To be activated by other indexes.
In systems, " vocabulary " being just known as is indexed.Semantic domain would generally provide activation (Active) operation, and
Index is the parameter of Active operation.The purpose of Active operation is closed according to index is corresponding with index/semanteme in semantic domain
It is the specific semantic object of table exampleization.
Define 7 animations: the process that scene described in speech is showed with the form of 3D animation.
In systems, animation not only includes complete scene and role animation (including existing animation in model data
Skeleton cartoon/frame animation broadcasting of realization, the movement and rotation of the whole position of model, the movement of video camera are calculated in systems
Deng), further comprise: scenario building, the addition of cartoon role/object, removal process.
Fig. 1 shows an embodiment of the system according to spatial term animation of the invention based on tool body cognition
Principle.Referring to Figure 1, here is the detailed description to the implementation principle of the system of the present embodiment.The system of the present embodiment includes
Semantic knowledge-base, system module group, computer program.
Semantic knowledge-base is multiple semantic word banks according to semantic Type division, each semantic word bank includes a kind of semantic
And correspond to a semantic domain in program.The information that project in each semanteme word bank is included is similar: index/semanteme
Corresponding table, semantic content (number of its related object, type, the required operation to object).It is realized from program
For angle, all the elements of semantic base are included in context when program is run.In some sense, program is exhausted most
Digital-to-analogue block can be realized by the communication between index manipulation and module also by index.
System supports the animation of the sentence based on grounded words, that is, those contents characterized that can pass through
The animation of the sentence of vision and Auditory Perception.Such as " his front at her ", " Xiao Ming's dancing ", " Xiao Wang go to tree side ", " small
It is bright to beat Xiao Wang " and " Xiao Wang says: ' I will go to park tomorrow ' " this kind of sentence.
System mould group is established on semantic knowledge-base, includes multiple system modules, and in systems, each semantic base is right
Ying Yuyi individual system modules, but system module and not only realize semanteme in corresponding semantic base.For example, one of semantic
The path semantic base of word bank has recorded object progress spatial movement when realizing the movement of different narrow senses and is formed by path.But with it
Path module in corresponding system module will also handle snap action, realize narrow other than realizing the semanteme of this part
Synchronous (if necessary) that justice movement animation is displaced with object itself.Other modules of system are responsible for other of simulation human brain
Cognitive function realizes necessary purpose in some engineerings, and just without corresponding semantic base.But these module working mechanisms
Still there are many similarities with the module of realization semantic base, for example activates specific 3D object etc. using index.In numerous systems
In module, other than realizing the semanteme in semantic base, the pretreatment to the natural language speech of input, sequence of movement are also completed
Control, 3D model management and the encapsulation to 3D engine.
The semantic word bank of composition semantic knowledge-base includes broad sense Action Semantic library, narrow sense Action Semantic library, narrow sense movement-
Spatiality semantic base, muscle control semantic base, spatial relation semantics library, static relation semantic base, path semantic base and sky
Between path semantic base.Referring to Figure 1, above-mentioned each semantic word bank is illustrated.
Narrow sense Action Semantic library itself does not include any real semantic information, and the particular content of narrow sense Action Semantic is deposited
Storage is in Liang Ge lower layer semantic base, it may be assumed that muscle controls semantic base and narrow sense movement-spatiality semantic base.Narrow sense Action Semantic
The presence in library is more to unify in logic to be more clear the level of model realizing.In system realization, narrow sense movement
The effect of module is directly to will pass to index (the i.e. above-mentioned two lower layer's semantic base) transmitting downwards of module.
Narrow sense Action Semantic includes that those will lead to the semanteme of object space state change (movement only in this way can
Pass through animation --- visual imagery shows), these semantemes are classified as the semanteme of description movement in linguistics.These become
Change can be relative to some object, be also possible to relative to objective world itself.Movement relative to some object is such as
" stretching (hand) ", " turning round (head) "；Relative to objective world itself such as " race ", " jump ".It is noted herein that described in above
The target of " relative to " is referring generally to being.
It should be noted that " races " such movement, although being also related to part relative to the other parts of entirety or whole
The spatiality of body itself changes, but this variation is not the key that This move.Such as: during running, leg relative to
The variation of trunk position relationship, it is not important compared to for " making entire body that change in location occur ".Opposite, " stretching (hand) "
Such movement, partially the spatiality variation relative to whole other parts or entirety itself is the key that movement.
In addition, " movement " is often difficult to differentiate between itself with " result of movement " for narrow sense movement.Such as it is right
In " stretching (hand) " this movement, the movement of mobile arm and " position of arm is made to change ", this was the result is that be closely related
's.Any point-in-time in " stretching (hand) " action process has the variation of corresponding spatiality.
Each semanteme in narrow sense Action Semantic includes following the description:
1) type of the object of initiation movement.It is usually specific, such as: the object of mankind's type, the hand object (mankind couple
A part for elephant).
2) type of the object of receiving movement.It is usually specific, such as: the object of mankind's type, objective world.
3) variation of spatiality caused by acting.This variation is indicated by the chained list of a simple path structure.Often
A path structure is a binary group { Start, End }.Start and End respectively represents two points in space, path be then by
The two put the straightway being formed by connecting.It arbitrarily complicated path can be close with a series of set of the first connected straightways
Like expression, each line segment is shorter, and degree of approximation is higher.
The realization of narrow sense Action Semantic was generally completed in the animation stage, in other words the dummy run phase.
To sum up, each narrow sense Action Semantic can be expressed as following four-tuples: Index, Object-A, Object-B,
Path}.Wherein, Index is the character string for activating object, and Object-A is the object of initiation movement, and Object-B is to hold
The object acted, Path are binary group mentioned above { Start, End }.
Spatial relation semantics library includes the semanteme for describing spatial relationship, is that relationship most basic in natural language is semantic.It is empty
Between be instantaneous relationship in relation nature, i.e. instantaneous state of the observer between the object at the time point for making description.This state
It itself is likely to be strictly instantaneous, but it could also be possible that continues.Spatial relationship is lasting, it is meant that in a period of time
It is interior, it goes to destroy this relationship without other movements.
Spatial relation semantics can be expressed as five-tuple: { Index, Object-A, Object-B, fc, fm }.Its
In, Index is this semantic index, and Object-A is the object in spatial relationship as object of reference, Object-B be
As the object of target in spatial relationship, fc is discriminant function for judging whether two objects have specific spatial relationship,
Fm is that constructor is used under the premise of Object-A has existed, and building one and Object-A have particular space relationship
Object-B.This Object-B is usually empty object.
The core of discriminant function fc and constructor fm are judgement/tectonic models of relationship.This model is to relationship one
Kind numerical value description, using the form of fuzzy set.The form for why using fuzzy set is because semanteme is mould to a certain extent
Paste, but it has prototype again.
For example, in systems, the position of object is specified by three-dimensional vector (x, y, z), and PosA is the position of Object-A
It sets, PosB is the position of Object-B, dis=PosB-PosA.Dis.x indicates the value of the x of this vector of dis.
So, just there is judgment rule (describing using pseudo-code) for the relationship semanteme indexed by " left side ":
If dis.x=-3Then Object-A and Object-B are typical " left side " relationships.
If dis.x<0&&dis.x>=- 6Then Object-A and Object-B are " left side " relationships.
If dis.x < -6Then Object-A and Object-B is not " left side " relationship.
The fc for realizing above-mentioned rule is provided in " left side " semanteme, for judging whether two objects have " left side " relationship.
The construction of the relationship semanteme of " left side " index can be realized with operations described below and (be described using pseudo-code):
1. obtaining the position PosA of Object-A.
2. making PosB=PosA+ (0,0, -3)
3. constructing empty object Object-B, the position that it is arranged is PosB.
The fm for realizing above-mentioned rule is provided in " left side " semanteme, for constructing an Object-B, makes itself and Object-A
With " left side " relationship.
In systems, the spatial relationship quantity for needing to realize is seldom, has: on, under, it is left, it is right, it is preceding, after, east, west, south, north,
It blocks, accommodates.
Spatial relationship and the maximum difference of narrow sense movement are that spatial relation semantics are just completed during sentence comprehension
The process of semantic realization.But the realization of narrow sense Action Semantic, it is necessary to could be completed in animation simulation.
Static semantic relationship library includes the semanteme of description static relation.Illustrate by taking the semanteme indexed comprising " " as an example,
" ", which has indexed, indicates that an object is present in the specific position in space there are relationship, and semanteme can be expressed as one four
Tuple: { Index, Object-A, Object-B, fm }.Wherein, Index is this semantic index, and Object-A is in space
As the object of object of reference in relationship, Object-B is the empty object in spatial relationship as target, and fm is constructor, is used
In under the premise of Object-A and Ojbect-B is existing, so that the position of Object-A is equal to the position of Object-B.
Fm can be described with pseudo-code are as follows:
1. obtaining the position PosB of Object-B.
2. obtaining the position PosA of Object-A.
3. making PosA=PosB.
Space path semantic base includes the semanteme of description space path.Illustrate by taking the semanteme of " to " index as an example, " arriving " rope
The semanteme for constructing a space path is drawn, this paths itself were described again on an object formerly latter two time point
Spatial position.
Space path semanteme can be indicated with a triple: { Index, Object-A, Object-B, path }.Wherein,
Index is this semantic index, and Object-A is the object moved on path, and Object-B is the target pair as terminal
As, it is necessary to it is an empty object, path is a partial ordering set, each element pathElement of set is one
The binary group of PathElement type, this binary group can be described with { start, end }.Wherein, start is initial position,
End is final position.In systems, start and end can be indicated with three-dimensional vector, and start is generally only (0,0,0).
Similar to narrow sense Action Semantic library, broad sense Action Semantic library itself does not include any real semantic information, extensively
The semanteme of justice movement is contained in the semantic base of junior as shown in Figure 1.
Path semantic base includes the semanteme of description narrow sense path of motion.Narrow sense movement refers to that those will lead to object space state
The movement of variation, as soon as a track can be left in space by executing the object of these movements, the semantic record in the semantic base of path
These tracks.
From the angle of engineering, according to the difference that animation is realized, it is semantic that there are two types of mode realizing routes: if 1. animation informations
It is previously stored inside model, then path semanteme only includes path itself；If realized 2. animation passes through rotation bone,
Path semanteme not only includes path, it is also necessary to comprising the corresponding moment to the rotation information of all bones to take exercises.The two is simultaneously
Different without essence, first method is more simple, and second method is more flexible accurate.
The definition and realization for the path semanteme that first method described herein needs.
The path semanteme of narrow sense movement can be described by triple: { Index, Object-A, object-path-
graph}.Wherein, Index is this semantic index.Object-A is the main body of movement, such as " Xiao Ming walks ", " small
It is bright " it is exactly the main body walked.Object-path-graph is an adjacency matrix.Adjacency matrix includes the mould that movement is related to
Variation of each part and these parts of type during the entire process of acting execution relative to last moment spatial position.Example
Such as, for the first implementation method of " walking " this movement, animation information is comprised in inside model, then object-path-
Just only remained in graph model it is whole and it " walking " animation implementation procedure displacement.
The realization of the semantic knowledge-base in system is described above, the principle of each module in system module group is carried out below
Explanation.
The system module of composition system module group includes: dynamic with the one-to-one path module of each semanteme word bank, broad sense
Make module, narrow sense movement-spatiality module, narrow sense action module, muscle control module, space path module, spatial relationship
Module, static relation module and time/event module, frame/input module, object module, analog module.
Path module can only be activated by the index that is transmitted by lower layer, this be system module corresponding with other semantic bases not
Same place.Path module is responsible for handling snap action (InstantPath, by static relation module activation), narrow sense movement
(NarrowActionPath is activated by narrow sense movement-spatiality module and muscle control module) and space path (SPath,
By the space path module activation of lower layer) animation control.These three path objects have a common parent BasePath.It closes
It had been described in the semantic base of path in the information that NarrowActionPath is stored and indicated.
Algorithm (the path objects building method, path objects weight of several major functions of path module offer are provided
Structure method, path elements recalculate method, animation execute method, path objects adding method) introduction and pseudo-code realize.This
A little algorithms are all called by path module, but the position of its specific implementation is determined by the structure and abstraction hierarchy of projects, big portion
Quartile is in the class of lower layer.
The purpose of path objects building method be will the information from other lower modules it is regular for several types with side
Just the processing after, and corresponding routing information is added to NarrowActionPath.Pseudo-code is described as follows:
1.1 obtain ActionId from the object A for calling activation primitive.It is obtained according to ActionId currently in path module
In corresponding objects Target.
If 1.2 this object are not Null, A is inserted into Target.
If 1.3 this object are Null, according to the different Target of the type structure of A, and A is inserted
Target。
1.3.1 if A is constructed from muscle control module or narrow sense movement-spatiality module
The Target of NarrowActionPath type after inserting A, obtains the OriPath of needs.
1.3.2 if A constructs the Target of SPath type from space path module, and A is inserted.
1.3.3 if A constructs the Target of InstantPath type from static relation module, and A is inserted.
The purpose of path objects reconstructing method is the space path object that processing is activated by indexing " to ".Space path object
Practical function can be varied according to the content of sentence.For example " Xiao Ming, which goes to tree, exists " and " Xiao Ming is to when setting " is likely to raw
At entirely different animation, substantially it is believed that for previous sentence, " to " provides only the description to path, will run to " race "
This movement is limited to bounded movement；And for latter sentence, " to " will be also responsible for activating the setting to animated actions, for example be arranged
For " walking ", " race " even " is jumped ", and the routing information for later providing itself is added up.
After calling path objects reconstructing method, the SPath from space path module should all be located
Reason.In path module, NarrowActionPath and InstantPath is only left according to the information that current sentence obtains.Method
Pseudo-code be described as follows:
1. taking out all SPath in the all_Paths that all BasePath activated from simple sentence are constituted, and it is put into set
temp_SPath。
2. successively extracting first object A that each SPath is included, (firstObject appears in " arriving " in sentence
The left side object).The NarrowActionPath found in current all_paths using A as first object is (normal
In the case of be at most only able to find one).And by second object of this NarrowActionPath (secondObject, it is former
Initial value should be Null) with second object value of current SPath.Path elements are called to recalculate method (below later
It can mention).Finally current SPath is deleted.
3. each remaining SPath object in couple temp_SPath calls path objects adding method (can mention below
To), it deletes later.
It is to recalculate path elements to agree with by sentence and existing analog loop that path elements, which recalculate the purpose of method,
The spatial position for the object that border determines jointly.Why to recalculate, there are two types of reason is usual: 1. are determined by path elements
Path length (distance of initial position to final position) is not equal to the path length of (being more than or less than) needs；2. by path member
There is barrier on the path that element determines.
Two algorithms are given below, respectively correspond aforementioned two kinds of situations, wherein second algorithm will call the first algorithm.
In addition, the algorithm due to go to action is relatively simple, can voluntarily be constructed with reference to following two algorithm.
Algorithm 1:
1. single path elements information is one and is constituted (Start and End) by two vectors by pathElement object
Complex object.Multiple pathElement constitute a queue OriPath.By the End of the pathElement in this queue
It is successively added, obtains the distance Ori_dis of original path.
2. indicating destination path (being provided by " to ", be a single pathElement) with Target_Path, calculate
The difference of Start and End in Target_path obtains the distance Target_dis of destination path.
3. calculating Target_dis/Ori_dis, scale is obtained.If scale < 2, it is 1 (for controlling that times, which is arranged,
Brake the number of repetition of picture), and by each End of each pathElement in OriPath multiplied by scale.If scale >
=2, then scale is rounded up, is denoted as scale '.Scale/scale ' is calculated, note result is new_scale.Times is set
For scale ', and and by each End of each pathElement in OriPath multiplied by new_scale.
Algorithm 2:
1. generating the fullpath of cut-through object by any heuristic path generating algorithm.And this paths is counted again
It calculates, generates by the approximate alternative scheme of multiple straightways.Each straightway is indicated with pathElement object.These straight lines
The number of section is n.
2. replicating n OriPath, by the corresponding expression straightway of each OriPath to pathElement, and adjust
The algorithm 1 provided with this section.All pathElement in all OriPath recalculated are extracted, are sequentially added
One queue, forms final OriPath.It is 1 that times, which is arranged,.
The purpose that animation executes method is to execute animation, and most of realize is located at muscle control module and narrow sense movement-
In class under spatiality module.Pseudo-code is described as follows:
In the case that if 1.1 pathElementNum are -1, and achieved mark is false, then to control animation
Variable is calculated.First pathElement is obtained, End/maxPathNum is calculated, result is recorded as
currentPathSeg.It is 1 that pathElementNum, which is arranged, and setting currentPathSegNum is 1.And allow path packet
The muscleControl object contained starts to execute animation.
1.2 for other situations:
The length that 1.2.1 if pathElmentNum is equal to PathElement queue subtracts 1, and
CurrentPathSegNum is equal to maxPathNum, then achieved is set to true.
The length that 1.2.2 if pathElmentNum is equal to PathElement queue subtracts 1, and
CurrentPathSegNum be less than maxPathNum, then the NAGC object control target object for allowing path to be included according to
CurrentPathSeg is mobile, and currentPathSegNum is added 1.
The length that 1.2.3 if pathElmentNum is less than PathElement queue subtracts 1, and
CurrentPathSegNum is equal to maxPathNum.Then pathElementNum is made to add 1, is obtained according to pathElemnetNum
Current pathElement simultaneously calculates currentPathSeg.It is 1 that currentPathSegNum, which is arranged,.And the path is allowed to be included
MuscleControl object start the animation of next stage.
The length that 1.2.4 if pathElmentNum is less than PathElement queue subtracts 1, and
CurrentPathSegNum is less than maxPathNum.The NAGC object control target object for so allowing path to be included according to
CurrentPathSeg is mobile, and currentPathSegNum is added 1.
The purpose of path objects adding method is in the case where only " arriving " without required movement in sentence, and creation is filled out
The NarrowActionPath to substitute the bad for the good.System can movement selected to use according to different conditions, here to use " walking " dynamic
As example.Pseudo-code is described as follows:
1. cloning the NarrowActionPath of one " walking " movement, and its OriPath is set.
2. obtaining action object A and target object B by SPath.
3. constructing the MuscleControl object Target_MuscleControl and NAGC object Target_ of " walking "
The firstObject of two objects is assigned a value of A by Nagc, and secondObject is assigned a value of B.
4. Target_MuscleControl and Target_Nagc is inserted NarrowActionPath, and call path
Element recalculates method.
5. NarrowActionPath is added in the temporary queue under path module.
Time/event module is used to record the sequencing of different movements, and holding for animation is controlled during animation simulation
Row.The index of each movement can obtain corresponding junior's index in broad sense action module and obtain an ActionId.Extensively
Each ActionId can be sent to time/event module while activating lower module by adopted action module.Time/event mould
There are the queues that one records ActionId in block, under the ActionId information storage for will be sent by broad sense action module
Come to use later.
In time/event module, there are also other two queues.One queue be used to record current sentence bounded (by
NarrowActionPath is provided) and instantaneous (being provided by InstantPath) movement, another queue, which is then recorded in, works as foreword
In language understanding task it is already existing it is all it is unbounded movement (equally provided by NarrowActionPath, but these
Whether NarrowActionPath's be the flag bit IsUnbounded of unbounded movement is true).
Two algorithms are executed in time/event module: movement method for reordering and simple sentence animation execute method.Below with puppet
Code illustrates the two algorithms respectively.
After path module has executed path objects reconstructing method, at this time the BasePath in path module often and
There is very big difference when sentence inputting, reason is that SPath has been not present, and there may be newly-generated
NarrowActionPath.So, it is therefore necessary to which the ActionId being located in time/event module is arranged again
Sequence places them into correctly processing queue, finally deletes those ActionId being not present.
The pseudo-code of movement method for reordering is expressed as follows:
1. all ActionId of current sentence are temporarily stored in temp_list before treatment, temp_list is checked
In all ActionId whether path module (this check algorithm provided by path module) is all provided.Delete those
ActionId through being not present.
2. whether the IsUnbounded flag bit for checking path objects corresponding to remaining ActionId is true.It will
All IsUnbounded flag bits are that the ActionId of true is sequentially added in time/object module for storing unbounded movement
Queue in, and former ActionId is deleted from temp_list.
The purpose that animation executes method is the successive suitable of object action characterized in natural language sentences according to input
Sequence executes necessary operation.Animation executes method and executes after movement method for reordering executes, at this time all ActionId
Sequence has met correct sequence, and bounded movement and unbounded movement have been completed differentiation, are placed into different queues
In the middle.
The pseudo-code that animation executes method is described as follows:
1.1, which are equipped with boundary (and instantaneous) movement, all has been placed in boundedAndInstantList, and unbounded movement is then put into
In unboundedList.
1.2 check boundedAndInstantList, and whether there are also remaining ActionId, if so, then:
1.2.1 taking out first.Corresponding path objects are found from path module.
1.2.2 check whether this path objects has been finished, if be finished: from
Corresponding ActionId is deleted in boundedAndInstantList, and corresponding path objects are deleted from path module.
1.2.3 if it is not, the animation for executing this object executes method (see aforementioned).It traverses later
Each ActionId in unboundedList finds the corresponding path objects of this ActionId in path module, and executes
Its animation executes method (see aforementioned).
Frame/input module for pre-processing to the natural language for the system that is input to, wrap in sequence by process content
Include following three step:
1. subordinate sentence.By the sentence of input according to ".","！" and "? " it is split.
2. extracting special object.System allows using different 3D objects.It is right by this to need before using 3D object
As associating with specific index, that is, the name of object.Later, in the language of input description animation, use is specific
Symbol comes out these object identities.For example " [] " is used to surround the name of object.System will be right by these special symbols
The index of elephant extracts, and uses for treatment process later.In addition, if it is desired to clearly set object location information, be
System also provides the format of input specific location information." (a, b, c) " is used to indicate specific position.Wherein, " a ", " b ", " c " are
Three positive integers.Here provide an example: [Xiao Ming] is in (0,0,0).Tree is in (20,0, -15).[Xiao Ming] Xiang Shu.[Xiao Ming]
Go to tree side.
3. participle.Since system can only handle stored vocabulary, so after the input of all vocabulary, as long as adopting
Participle task can be completed with simplest maximum matching process.More complicated method can certainly be used or using existing
Participle tool segmented.
Object module has recorded relationship between index (name) and 3D object.According to the index of input, object module is called
Specific 3D model is loaded by the corresponding method of analog module, and generates the object for controlling this 3D model.Object module is also
Empty object can be constructed with same process, is used for process later according to the location information of input.
Analog module is the encapsulation that system carries out the part needed in 3D engine.Analog module is in addition to needing to provide other
Module to the control method of 3D object, scene outside, be also responsible for the filling of object involved in relationship, movement.This process is related to
Two kinds of relationship filling algorithms, first object (object of reference object or agent object) being related to respectively for relationship/movement and the
Two objects (target object or word denoting the receiver of an action object).After the execution of relationship filling algorithm, relationship is filled, so that it may be started subsequent
Treatment process.
Analog module determines the activation sequence of each object using intensity of activation mechanism at any time.Intensity of activation mechanism
Be described as follows: each object is endowed highest intensity of activation and (can set, be defaulted as 7) when activation.Swashing
Before any new object living, if there is the object activated in analog module, the activation of all these objects is strong
Degree, which subtracts a fixed value, (can set, be defaulted as 1).When an object has existed, and its index plays work again
With when, this object is specific to influence to be that its intensity of activation is re-set as peak by " reactivation ".For swashing again
Journey is lived through, before reactivation starts, the intensity of activation of other objects will also subtract a fixed value and (can set, default
For 1).
It should be noted that the algorithm for obtaining first object is located at analog module, the relationship tune filled by needs
With.It is located at for the relationship filling algorithm of second object and needs the class of the relationship of filling, is called by analog module.
Relationship filling algorithm (first object) is to obtain to can be used for first object in filling relationship from analog module
Object.Pseudo-code is described as follows:
1.1 obtain the object of the currently active maximum intensity in module.
1.2 judge that following conditions of this object are no while meeting: intensity of activation is greater than -1, " isUsed " of object
Mark is true, and object is not empty.
1.2.1 if it is satisfied, then obtaining the object of smaller than the intensity of activation currently set 1 intensity of activation.
1.2.2 if conditions are not met, then existing object is the target object needed.
The pseudo-code of relationship filling algorithm (second object) is described as follows:
1. each relationship/action module has the relationship/action object team for being used to store currently processed sentence of oneself
Column, if this queue is operatorList.Each element of queue is relationship/action object.Obtain changing for this queue
For device, and by iterator be directed toward queue end.If this iterator has previous object, this object is passed through into iteration
Device takes out, and judges whether this object is filled.
If 1.1 this object are not filled.Then fill second object of this object.
1.1.1 if fill successfully judge this for filling object whether be sky object (placeholder).
It 1.1.1.1 is true by the mark position " isUsed " of this object if it is empty object.
As shown in figure 3, system of the invention first receives the sentence of input and puts it into sentence list to be processed.From
Current sentence is taken out in list to be handled, and for each sentence, first implements simple sentence reasoning process, then implement simple sentence animation and drill
Show process, until the pending sentence of institute is disposed in list.
And the specific steps of simple sentence reasoning and simple sentence flash demo are as shown in Figure 2.
Step S101: the sentence of natural language is split, object analysis and word segmentation processing.
Step S102: judging whether the sentence there are also non-reasoning, if the sentence there are also non-reasoning thens follow the steps S103,
Otherwise process terminates.
Step S103: enter the sentence justice reasoning stage.
Step S104: judge whether to execute step if there are also untreated index there are also untreated index (vocabulary)
Rapid S105, it is no to then follow the steps S115.
Step S105: the current index of judgement is object or broad sense movement, thens follow the steps S106 if it is object, if
It is that broad sense movement thens follow the steps S108.
Step S106: the corresponding objects in activation analog module.
Step S107: filling second object of relationship/movement, return step S104.
Step S108: judging the type of action of broad sense movement, if type of action is building, spatial relationship is thened follow the steps
S109, if type of action is building, static relation thens follow the steps S110, if type of action is the specific narrow sense movement of building
S111 is thened follow the steps, space path thens follow the steps S112 if type of action is building.
Step S109: then building particular space relationship executes step S114.
Step S110: then building particular static relationship executes step S113.
Step S111: specific narrow sense movement is constructed, step S113 is then executed.
Step S112: then building space path executes step S113.
Step S113: corresponding BasePath is constructed in path module.
Step S114: first object of relationship/movement is filled, then return step S104.
Step S115: each path is reconstructed in path module.
Step S116: sequence of movement is reconstructed in time/event module.
Step S117: simple sentence reasoning is completed, for flash demo, and the return step S102 after carrying out flash demo.
Although for simplify explain the above method is illustrated to and is described as a series of actions, it should be understood that and understand,
The order that these methods are not acted is limited, because according to one or more embodiments, some movements can occur in different order
And/or with from it is depicted and described herein or herein it is not shown and describe but it will be appreciated by those skilled in the art that other
Movement concomitantly occurs.
Those skilled in the art will further appreciate that, the various illustratives described in conjunction with the embodiments described herein
Logic plate, module, circuit and algorithm steps can be realized as electronic hardware, computer software or combination of the two.It is clear
Explain to Chu this interchangeability of hardware and software, various illustrative components, frame, module, circuit and step be above with
Its functional form makees generalization description.Such functionality be implemented as hardware or software depend on concrete application and
It is applied to the design constraint of total system.Technical staff can realize every kind of specific application described with different modes
Functionality, but such realization decision should not be interpreted to cause departing from the scope of the present invention.
General place can be used in conjunction with various illustrative logic plates, module and the circuit that presently disclosed embodiment describes
Reason device, digital signal processor (DSP), specific integrated circuit (ASIC), field programmable gate array (FPGA) other are compiled
Journey logical device, discrete door or transistor logic, discrete hardware component or its be designed to carry out function described herein
Any combination is realized or is executed.General processor can be microprocessor, but in alternative, which, which can be, appoints
What conventional processor, controller, microcontroller or state machine.Processor is also implemented as calculating the combination of equipment, example
As DSP and the combination of microprocessor, multi-microprocessor, the one or more microprocessors to cooperate with DSP core or it is any its
His such configuration.
The step of method or algorithm for describing in conjunction with embodiment disclosed herein, can be embodied directly in hardware, in by processor
It is embodied in the software module of execution or in combination of the two.Software module can reside in RAM memory, flash memory, ROM and deposit
Reservoir, eprom memory, eeprom memory, register, hard disk, removable disk, CD-ROM or known in the art appoint
In the storage medium of what other forms.Exemplary storage medium is coupled to processor so that the processor can be from/to the storage
Medium reads and writees information.In alternative, storage medium can be integrated into processor.Pocessor and storage media can
It resides in ASIC.ASIC can reside in user terminal.In alternative, pocessor and storage media can be used as discrete sets
Part is resident in the user terminal.
In one or more exemplary embodiments, described function can be in hardware, software, firmware, or any combination thereof
Middle realization.If being embodied as computer program product in software, each function can be used as one or more item instructions or generation
Code may be stored on the computer-readable medium or be transmitted by it.Computer-readable medium includes computer storage medium and communication
Both media comprising any medium for facilitating computer program to shift from one place to another.Storage medium can be can quilt
Any usable medium of computer access.It is non-limiting as example, such computer-readable medium may include RAM, ROM,
EEPROM, CD-ROM or other optical disc storages, disk storage or other magnetic storage apparatus can be used to carrying or store instruction
Or data structure form desirable program code and any other medium that can be accessed by a computer.Any connection is also by by rights
Referred to as computer-readable medium.For example, if software is using coaxial cable, fiber optic cables, twisted pair, digital subscriber line
(DSL) or the wireless technology of such as infrared, radio and microwave etc is passed from web site, server or other remote sources
It send, then the coaxial cable, fiber optic cables, twisted pair, DSL or such as infrared, radio and microwave etc is wireless
Technology is just included among the definition of medium.Disk (disk) and dish (disc) as used herein include compression dish
(CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and blu-ray disc, which disk (disk) are often reproduced in a manner of magnetic
Data, and dish (disc) with laser reproduce data optically.Combinations of the above should also be included in computer-readable medium
In the range of.
Offer is to make any person skilled in the art all and can make or use this public affairs to the previous description of the disclosure
It opens.The various modifications of the disclosure all will be apparent for a person skilled in the art, and as defined herein general
Suitable principle can be applied to other variants without departing from the spirit or scope of the disclosure.The disclosure is not intended to be limited as a result,
Due to example described herein and design, but should be awarded and principle disclosed herein and novel features phase one
The widest scope of cause.
Claims (14)
1. it is a kind of based on tool body cognition the system according to spatial term animation, which is characterized in that including semantic knowledge-base,
System module group and a computer program, in which:
Semantic knowledge-base is divided into multiple semantic word banks according to semantic type, and each semanteme word bank includes a kind of semantic and right
A semantic domain is answered, the project in each semanteme word bank includes the information including the corresponding table of index/semanteme, semantic content；
System module group, is established on semantic knowledge-base, including multiple system modules, wherein each corresponding list of semanteme word bank
Only system module, the system module corresponding to semantic word bank realize the semanteme in corresponding semantic word bank, other system moulds
Block completes the pretreatment to the natural language speech of input, sequence of movement control, 3D model management and the encapsulation to 3D engine；
Computer program is converted specific natural language sentence generated to by controlling the interaction between each system module
Animation, including simple sentence reasoning process and simple sentence flash demo process.
2. the system according to spatial term animation according to claim 1 based on tool body cognition, which is characterized in that
The semantic word bank of semantic knowledge-base includes broad sense Action Semantic library, narrow sense Action Semantic library, narrow sense movement-spatiality semanteme
Library, muscle control semantic base, spatial relation semantics library, static relation semantic base, path semantic base, space path semantic base.
3. the system according to spatial term animation according to claim 2 based on tool body cognition, which is characterized in that
Each system module of composition system module group includes and the one-to-one path module of each semanteme word bank, broad sense act mould
Block, narrow sense movement-spatiality module, narrow sense action module, muscle control module, space path module, spatial relationship module,
Static relation module and time/event module, frame/input module, object module, analog module.
4. the system according to spatial term animation according to claim 3 based on tool body cognition, which is characterized in that
The content of the narrow sense Action Semantic in narrow sense Action Semantic library is stored in the muscle control semantic base and narrow sense movement-space shape of lower layer
In state semantic base, narrow sense Action Semantic refers to the semanteme that will lead to object space state change, wherein object space state change
Including relative to some object or relative to objective world itself, narrow sense Action Semantic by a quadruple notation,
Four elements therein respectively indicate object, the movement that character string, the object of initiation movement, receiving for activating object act
The variation of caused spatiality.
5. the system according to spatial term animation according to claim 4 based on tool body cognition, which is characterized in that
Spatial relation semantics library is the semanteme for describing spatial relationship, and spatial relation semantics are indicated by a five-tuple, five elements point
Be not semantic indexing, in spatial relationship as the object of object of reference, in spatial relationship as the object of target, for judging
Whether object of reference object and target object have the discriminant function of particular space relationship, for already existing in object of reference object
Under the premise of building one and object of reference object have particular space relationship target object constructor.
6. the system according to spatial term animation according to claim 5 based on tool body cognition, which is characterized in that
Static relation semantic base be describe static relation semanteme, static relation include indicate an object be present in it is specific in space
Position there are relationship, there are the semantemes of relationship by a four-tuple mark, and four elements are semantic index respectively, in sky
Between in relationship as the object of object of reference, in spatial relationship as the empty object of target, in object of reference object and target empty pair
The position of object of reference object is made to be equal to the constructor of the position of target empty object under the premise of as existing for all.
7. the system according to spatial term animation according to claim 6 based on tool body cognition, which is characterized in that
Space path semantic base describes the semanteme of space path, and the semanteme for describing space path includes for constructing a space path
Semanteme, wherein space path describes the spatial position on an object formerly latter two time point, passes through a four-tuple table
Show space path semanteme, wherein four elements respectively indicate semantic indexing, moved on path object, the target as terminal
Empty object, the path being made of initial position and final position set.
8. the system according to spatial term animation according to claim 7 based on tool body cognition, which is characterized in that
Path semantic base includes the semanteme of description narrow sense path of motion, narrow sense path of motion is indicated by a triple, wherein three
Element respectively indicates semantic indexing, action subject, adjacency matrix, and wherein adjacency matrix includes each of the model that movement is related to
Partially and these parts act execute during the entire process of variation relative to last moment spatial position.
9. the system according to spatial term animation according to claim 8 based on tool body cognition, which is characterized in that
Path module is used to handle the animation control of snap action, narrow sense movement and space path.
10. the system according to spatial term animation according to claim 9 based on tool body cognition, feature exist
In time/event module is used to record the sequencing of different movements, and the execution of animation is controlled during animation simulation.
11. the system according to spatial term animation according to claim 10 based on tool body cognition, feature exist
In frame/input module is for pre-processing the natural language for the system that is input to, wherein pretreated content includes point
Sentence extracts special object, participle.
12. the system according to spatial term animation according to claim 11 based on tool body cognition, feature exist
In, object module is for the relationship between recording indexes and 3D object, by calling analog module to be loaded into specific 3D model,
Generate the object for controlling the 3D model.
13. the system according to spatial term animation according to claim 12 based on tool body cognition, feature exist
The part needed in 3D engine is packaged for system in, analog module, is supplied to system other modules to 3D object, field
The filling of object involved in the control method of scape and responsible relationship, movement.
14. the system according to spatial term animation according to claim 13 based on tool body cognition, feature exist
In computer program is run in systems to execute following steps:
Step 1: the sentence of natural language being split, object analysis and word segmentation processing, into subsequent sentence justice reasoning rank
Section；
Step 2: untreated index is judged whether there is, if thening follow the steps 3, if without thening follow the steps 12；
Step 3: the current index of judgement is object or broad sense movement, if object thens follow the steps 4, if broad sense movement is then held
Row step 5；
Step 4: the corresponding objects in activation analog module fill second object of relationship/movement, return step 2；
Step 5: judging type of action, static relation thens follow the steps 6 if type of action is building, if type of action is that building is empty
Between relationship then follow the steps 7, if type of action be building narrow sense movement then follow the steps 8, if type of action be building space road
Diameter thens follow the steps 9；
Step 6: building particular static relationship continues to execute step 10；
Step 7: building particular space relationship continues to execute step 11；
Step 8: constructing specific narrow sense movement, continue to execute step 10；
Step 9: building space path continues to execute step 10；
Step 10: constructing corresponding basis path BasePath in path module；
Step 11: filling first object of relationship/movement, return step 2；
Step 12: reconstructing each path in path module, sequence of movement is then reconstructed in time/event module, complete single
Sentence reasoning；
Step 13: simple sentence flash demo is realized in completing simple sentence reasoning process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810078217.2A CN110083818A (en) | 2018-01-26 | 2018-01-26 | A kind of system according to spatial term animation based on tool body cognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810078217.2A CN110083818A (en) | 2018-01-26 | 2018-01-26 | A kind of system according to spatial term animation based on tool body cognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110083818A true CN110083818A (en) | 2019-08-02 |
Family
ID=67412669
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810078217.2A Pending CN110083818A (en) | 2018-01-26 | 2018-01-26 | A kind of system according to spatial term animation based on tool body cognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110083818A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1447265A (en) * | 2002-03-25 | 2003-10-08 | 白涛 | Basic method for converting character type sentence to corresponding animation cartoon |
| CN1952928A (en) * | 2005-10-20 | 2007-04-25 | 梁威 | Computer system to constitute natural language base and automatic dialogue retrieve |
| US20140160134A1 (en) * | 2011-08-31 | 2014-06-12 | Timur Nuruahitovich Bekmambetov | Visualization of a natural language text |
| CN107193904A (en) * | 2017-05-11 | 2017-09-22 | 浙江唯见科技有限公司 | A kind of books VR and AR experience interactive system |
-
2018
- 2018-01-26 CN CN201810078217.2A patent/CN110083818A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1447265A (en) * | 2002-03-25 | 2003-10-08 | 白涛 | Basic method for converting character type sentence to corresponding animation cartoon |
| CN1952928A (en) * | 2005-10-20 | 2007-04-25 | 梁威 | Computer system to constitute natural language base and automatic dialogue retrieve |
| US20140160134A1 (en) * | 2011-08-31 | 2014-06-12 | Timur Nuruahitovich Bekmambetov | Visualization of a natural language text |
| CN107193904A (en) * | 2017-05-11 | 2017-09-22 | 浙江唯见科技有限公司 | A kind of books VR and AR experience interactive system |
Non-Patent Citations (2)
| Title |
|---|
| 李海英 等: "具身在人工智能导师系统中隐身何处？", 《华南师范大学学报(社会科学版)》 * |
| 郭键: "自然语言脚本生成动画脚本的关键技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Evans et al. | Versu—a simulationist storytelling system | |
| US8508537B2 (en) | System and method for dependency graph evaluation for animation | |
| CN106503034B (en) | A kind of method and device for motion picture soundtrack | |
| León et al. | Creativity in Story Generation From the Ground Up: Non-deterministic Simulation driven by Narrative. | |
| Hahn et al. | Persistent realtime building interior generation | |
| CN107767438A (en) | A kind of method and apparatus that user mutual is carried out based on virtual objects | |
| Borovikov et al. | Winning isn’t everything: Training agents to playtest modern games | |
| Bertrand et al. | Controlling a population | |
| Flórez-Puga et al. | Query-enabled behavior trees | |
| CN110083818A (en) | A kind of system according to spatial term animation based on tool body cognition | |
| CN113673672A (en) | Curling game strategy generation method based on Monte Carlo reinforcement learning | |
| Abbadi et al. | Casanova: A simple, high-performance language for game development | |
| Poulakos et al. | Towards an accessible interface for story world building | |
| Evans et al. | The AI Architecture of Versu | |
| CN108961428A (en) | A kind of Style Transfer method of three-dimensional motion, medium, device and calculate equipment | |
| Stanley | Art in the sciences of the artificial | |
| Espinosa Leal et al. | Reinforcement learning for extended reality: designing self-play scenarios | |
| Lumbreras et al. | Hyperstories: A model to specify and design interactive educational stories | |
| CN114155325A (en) | Virtual character animation generation method and system | |
| Sarkar et al. | Procedural content generation using behavior trees (PCGBT) | |
| Slater et al. | A review of agent emotion architectures | |
| Goudoulakis et al. | Evaluation of planning algorithms for digital interactive storytelling | |
| Wang | Mastering the game of Gomoku without human knowledge | |
| Chao et al. | Quantum Enhancements for AlphaZero | |
| Palazzo et al. | Spyke3d: A new computer games oriented bdi agent framework |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190802 |