CN105957046B - The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input - Google Patents
The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input Download PDFInfo
- Publication number
- CN105957046B CN105957046B CN201610289106.7A CN201610289106A CN105957046B CN 105957046 B CN105957046 B CN 105957046B CN 201610289106 A CN201610289106 A CN 201610289106A CN 105957046 B CN105957046 B CN 105957046B
- Authority
- CN
- China
- Prior art keywords
- endpoint
- stroke
- gesture
- group
- coordinate system
- 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.)
- Expired - Fee Related
Links
- 238000007500 overflow downdraw method Methods 0.000 title claims description 13
- 238000000034 method Methods 0.000 claims abstract description 33
- 230000004927 fusion Effects 0.000 claims abstract description 32
- 238000012937 correction Methods 0.000 claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 7
- 239000000284 extract Substances 0.000 claims abstract description 5
- 238000007499 fusion processing Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 2
- 239000013585 weight reducing agent Substances 0.000 claims description 2
- 238000013461 design Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- UDHXJZHVNHGCEC-UHFFFAOYSA-N Chlorophacinone Chemical compound C1=CC(Cl)=CC=C1C(C=1C=CC=CC=1)C(=O)C1C(=O)C2=CC=CC=C2C1=O UDHXJZHVNHGCEC-UHFFFAOYSA-N 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000005267 amalgamation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000000155 melt Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Abstract
The present invention extracts the angle information through being fitted obtained gesture strokes first, and draws out corresponding angular histogram to obtain the angular distribution of gesture strokes, then carries out collimation to gesture according to angular histogram and obtain correction stroke；Then, judge cluster point using variable coefficient tolerance band, point to be fused is subjected to grouping and clustering；Finally, correction stroke progress endpoint is merged using " coordinate system-weight criterion " of the present invention regular.The method of the present invention can be regular to the flat and stereo perspective view progress endpoint fusion drawn online, efficiently solves the existing fusion problem for forcing to mediate endpoint, and carry out efficiently regular processing to Freehandhand-drawing perspective view.
Description
Technical field
The present invention relates to a kind of endpoint fusion methods of online Freehandhand-drawing flat and stereo perspective view, are set for intelligent Freehandhand-drawing input
It is standby, such as handwriting pad, smart mobile phone.
Background technology
Cartographical sketching be it is a kind of from however directly thought alienation and communication exchanges mode, its easy, nature and image,
Be conducive to designer to show Design Thinking and carry out creative activity.One outstanding product designer, will not only have
Design, intention, it is also necessary to be expressed by certain form of expression.Design to it is perceived must be specific by certain
Carrier conversion, this graphic mode of sketch is design of expression intention and design capture memory is most direct, most effective means.Hand
Cartographic sketching draws that speed is fast, lines beautiful nature, gives a kind of strong artistic appeal of people.But since cartographical sketching has wind
The features such as lattice variation, abstract, inaccurate, unstable, these are understood in computer the design idea of user by cartographical sketching
It is very difficult.Mainly there are three aspects for its reason：First, people often have more for the sketch that same object is drawn
Kind of the form of expression, the same object that actually same person is drawn in different times are likely to that there are relatively big differences.The
Two, when expressing threedimensional model using two-dimentional sketch, data dimension reduce and caused by ambiguity；Third, sometimes user couple
Draw the indefinite of object geometric format and caused by randomness.The target of online hand-sketched graphics recognition is exactly to ensure that user is defeated
Under the premise of entering intention and user adaptation, best identification and regular is carried out to sketch.
Currently, being rebuild in the field of 3D models by 2D sketches, generally using based on basic line element in terms of sketch explanation
Manual draw recognition methods.This method solve the Uniqueness of figure description, including stroke segmentation, identification, fitting, endpoint
Fusion and etc., wherein endpoint fusion is to influence a committed step of recognition effect.
For the perspective view of Freehandhand-drawing flat and stereo, be made of a series of single stroke straightways, and the stroke depth of parallelism it is inconsistent,
It is also not perfect at node.Endpoint integration technology have passed through long-term research and have many achievements in research, but predominantly
Endpoint fusion method based on tolerance band：The existing method that tolerance threshold value is obtained based on statistic analysis result；By line element length
Method of one percentage as tolerance threshold value.Obviously, the above tolerance threshold value result can not obtain reliable fusion results.
The value of tolerance threshold should be determined according to " susceptibility " of specific line element endpoint, according to relatively fixed two line of tolerance threshold value pair
The amalgamation of first endpoint is judged that threshold value led to fusion, the too small endpoint that can cause be connected of tolerance threshold value compared with conference
Fail to be connected.Lipson then proposes the method based on average distance, calculates each originating endpoint first to each entity
Average distance, it includes average distance of the endpoint to the entity itself, i.e. the half length of entity line element, and its minimum value is made
For the tolerance radius of endpoint.Wang improves this method later, proposes the endpoint fusion side based on variable coefficient tolerance threshold
Method.After obtaining the tolerance band of originating endpoint, it polymerize if originating endpoint two-by-two is mutually fallen into the tolerance band of other side,
The two nodes are merged again if constituting two nodes and can also constitute new fusion each other, and will be in each cluster point
The center of originating endpoint be denoted as node.Both the above method, the former tolerance threshold value are relatively fixed；And the latter is according to whole
The size of a entity determines tolerance threshold value, it is made to have more adaptability.But the two does not all account for respectively in fusion process
The relative position relation of sketch stroke, treats each entity in isolation, cluster point is forced to pinch to be combined together, to a certain degree
On destroy connecting each other between each entity, also illustrate the simple simple and easy to do income effect for considering to merge regular algorithm
Not robust.
Invention content
Technical problems to be solved
To solve the problems, such as that existing endpoint integration technology exists, the present invention proposes a kind of for the flat of intelligent Freehandhand-drawing input
The endpoint fusion method of face axonmetric chart according to " context relation " of skeletonizing, and combines and is based on adaptability tolerance band
Fusion method, to sketch stroke carry out endpoint fusion it is regular.
Technical solution
A kind of endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input, inputs as identified fitting institute
Gesture strokes are obtained, are made of N straightway；It is characterized in that steps are as follows：
Step 1：Extract each gesture strokes { Li；1≤i≤N } angle information value：{θi；0≤θi≤180°}；
Step 2：Collimation is carried out to gesture using the method based on angular histogram：
Step 2.1：Using gesture wire size i as abscissa, angle information θiFor ordinate, make histogram, to obtain gesture
The angular distribution of stroke；
Step 2.2：Clustering is carried out to gesture according to the angular distribution and angle threshold Δ δ, even | θi-θj|≤
Δ δ, then by gesture strokes Li、LjIt is divided into one group, finally obtains several groups gesture；
Step 2.3：Assuming that s group strokes are by nsGesture composition, calculates the angular average of every group of gesture strokes, then has
Wherein t, t+1 ..., nsIt indicates stroke wire size in the group of s group gestures, and works asOrOrOrWhen,
Then, by every gesture by angle mean value in respective groupCarry out angle correct so that angle is consistent in group, i.e.,So just achieve the purpose that collimation, and obtained stroke is known as correction stroke；
Step 3：Fusion treatment is carried out to correction stroke using the method based on " coordinate system-weight criterion ", corrects stroke
Sequence { li(θi)；1≤i≤N}；
Step 3.1：All endpoint { d of extraction correction stroke firstie；1≤i≤N, e=0,1 }, i.e., point to be fused, wherein
di0Indicate stroke starting point, di1Indicate stroke terminal；
Step 3.2：The cluster of point to be fused is carried out using the endpoint clustering method based on variable coefficient tolerance band：If several wait for
Merging point is mutually fallen into the tolerance circle of other side, then by these points to be fused as a cluster, per group cluster point number
For K；Finally same group cluster point will be replaced by a node in endpoint fusion process, and gained number of nodes R is
And sequence node is expressed as { Pu；1≤u≤R}；
Step 3.3：R number of nodes is projected, projection node is { Pm；1≤m≤R }, projection node is by three group of strokes
At constituting R coordinate system；Traverse coordinate system sequence { Ci；1≤i≤R }, the consistent coordinate system in direction is divided into one group, then
To several groups coordinate system, every group of coordinate system number is wj(1≤j≤R), and same group of coordinate system sees a coordinate system as, presses
According to weight w shared by coordinate system groupjSize, obtain new coordinate system sequence { Vx；1≤x≤q }, respectively indicate 1 grade of coordinate system,
2 grades of coordinate systems, q represent total rank number of coordinate system；
Step 3.4：Successively according to weight wjThe sequence of reduction carries out each node the fusion treatment of cluster point：
Step 3.4.1：By the new coordinate system sequence { V of the order traversal of weight reductionx；1≤x≤q}；
Step 3.4.2：Judge and determines VxIn participated in merging regular correction stroke number M being 0 or 1 or 2 or 3, and
Stroke was ranked up, i.e. l according to " participating in merging regular preferential storage principle "1、l2Always preferential storage has participated in merging
Regular correction stroke；
Step 3.4.3：If M=3, and K=3, then l1、l2It asks and meets at P points, work as e=0, stroke l3On it is fused regular
Endpoint d0It is constant, d1It coincides with P points；Work as e=1, stroke l3On merged regular endpoint d1It is constant, d0It coincides with P points, this
In the case of kind, since three strokes both participated in processing, stroke endpoint is connected, then inevitably just will produce fusion error；
Step 3.4.4：If M=0,1,2, and K=2, then l1、l2It asks and meets at P points；
Step 3.4.5：If M=0,1,2, and K=3, then l1、l2It asks and meets at P points；And work as e=0, by l3Move to d0With point
P is overlapped；Work as e=1, by l3Move to d1It is overlapped with point P, and keeps angle constant.
Δ δ=20 ° in step 2.2.
The endpoint clustering method based on variable coefficient tolerance band uses document " Shuxia W, Suihuai in step 3.2
Y.Endpoint fusing of freehand 3D object sketch with Hidden-part-draw[C]
.Computer-Aided Industrial Design&Conceptual Design,2009.CAID&CD 2009.IEEE
10th International Conference on,2009:Method disclosed in 586-590 ".
Before endpoint fusion, parallel processing is carried out to gesture according to " contextual information " of stroke, it is defeated to retain user in time
The initial drafting entered is intended to；The reservation of stroke parallel relation is so that regular result is more in line with its projection with 3D solid simultaneously
Relationship and convenient for being coupled with existing three-dimensional rebuilding method.
This method is divided into two stages：Error free fusing stage, when just proceeding by endpoint fusion, what preceding coordinate system
In stroke be almost in untreated state, M=0 in this stage, 1,2, then the fusion in the stage it is regular only ask friendship,
Translate two kinds of forms.Therefore, which is that zero error is regular；Band error fusing stage, at this moment M=3, i.e., all correction strokes are all
Fusion treatment is taken part in, therefore when handling to be fused, inevitably corresponding stroke can carry out non-translation, the angle of stroke
Therefore it changes, therefore error also accumulates grade coordinate system to the end.
The angle for correcting stroke is known as ideal regular angle value θd, the stroke angle of gained is known as real after endpoint merges
The regular angle value θ in borderaIf stroke angle is still θ after regulard, i.e. θa=θd, then being exactly best regular result.But by
In the presence of fusion error, that is, merge front and rear angles difference Δ θ=| θa-θd|.It is therefore proposed that by Freehandhand-drawing perspective view in regular completion
Generated differential seat angle Δ θ merges an excellent standard as endpoint is judged afterwards.
Advantageous effect
The present invention extracts the angle information through being fitted obtained gesture strokes first, and draws out corresponding angular histogram
To obtain the angular distribution of gesture strokes, then collimation is carried out to gesture according to angular histogram and obtains correction stroke；So
Afterwards, judge cluster point using variable coefficient tolerance band, point to be fused is subjected to grouping and clustering；Finally, " the coordinate of the present invention is utilized
It is regular that system-weight criterion " carries out endpoint fusion to correction stroke.The method of the present invention can throw the flat and stereo that can be drawn online
Shadow figure carry out endpoint fusion it is regular, efficiently solve it is existing force mediate endpoint fusion problem, and to Freehandhand-drawing perspective view into
The efficient regular processing of row.
Description of the drawings
Fig. 1：The endpoint of online cartographical sketching merges regular flow chart
Fig. 2：Cluster point tolerance justifies schematic diagram：(a) 2 points of clusters；(b) 3 points of clusters；
Fig. 3：The endpoint of embodiment Freehandhand-drawing flat and stereo perspective view merges regular process：(a) sketch stroke；(b) gesture pen
It draws；(c) stroke is corrected；(d) fusion results
Fig. 4：Embodiment gesture angular histogram
Fig. 5：The front and back Δ θ of embodiment endpoint fusion changes line chart
Fig. 6：The many cases experimental result picture of the endpoint fusion method of online Freehandhand-drawing flat and stereo perspective view
Specific implementation mode
In conjunction with embodiment, attached drawing, the invention will be further described：
The endpoint fusion method based on angular distribution used in the present embodiment, using following steps：
If Fig. 3-a show embodiment sketch, a kind of endpoint of flat and stereo perspective view for intelligent Freehandhand-drawing input melts
Conjunction method includes the following steps：
Step 1：The known embodiment Freehandhand-drawing flat and stereo perspective view is made of 18 gestures.First, the gesture sequence is traversed
Arrange { Li；1≤i≤18 }, extract the angle information value { θ of every gesturei；0≤θi≤180°}；
Step 2：Collimation is carried out to gesture using the method based on angular distribution；
Step 2.1：With gesture wire size i (1≤i≤18) for abscissa, angle information θiFor ordinate, make histogram, from
And the angular distribution of gesture strokes is obtained, as shown in Figure 4；
Step 2.2：According to step 2.1 gained angular distribution, the grouping of gesture is carried out.When | θi-θj|≤Δ δ (Δ δ here
=20 °), by gesture strokes Li、LjIt is divided into one group, finally obtains four groups of gesture Qx(1≤x≤4)：{Q1:L1,L4,L10,L14,
L17}；{Q2:L2,L3,L9,L13,L18}；{Q3:L5,L8}；{Q4:L6,L7,L11,L12,L15,L16}。
Step 2.3：The angular average of every group of gesture strokes is calculated according to the above groupingSo
Have
Then, by every gesture by angle mean value in respective groupCarry out angle correct so that the angle of every group of stroke
Unanimously, to achieve the purpose that collimation, and correction stroke { l is obtainedi(θi)；1≤i≤18}；
Step 3：Direct regular object in endpoint fusion process is correction strokes sequence { li(θi)；1≤i≤18 }, still
In fact, it does not meet at node strictly, it is therefore desirable to be attached processing to them.To correction stroke obtained as above,
Fusion treatment is carried out using the method based on " coordinate system-weight criterion "；
Step 3.1：All endpoint { d of extraction correction stroke firsti,e；1≤i≤18, e=0,1 }, i.e., point to be fused,
Middle di0Indicate stroke starting point, di1Indicate stroke terminal；
Step 3.2：Using variable coefficient tolerance band, treats merging point and clustered.If several points to be fused are mutually fallen into pair
In the tolerance circle of side, then by these points to be fused as a cluster, as shown in Fig. 3-c, 12 groups of point clusters are finally obtained
Dg(1≤g≤12)：D1={ d1,0,d2,0,d16,0, D2={ d1,1,d3,0,d12,0, D3={ d2,1,d4,0,d11,0, D4={ d3,1,
d5,0,d6,0, D5={ d4,1,d5,1,d7,0, D6={ d6,1,d8,0,d9,0, D7={ d7,1,d8,1,d10,0, D8={ d9,1,d10,1,
d15,0, D9={ d11,0,d14,0,d18,1, D10={ d12,1,d13,0,d17,1, D11={ d13,1,d14,1,d15,1, D12={ d16,1,
d17,0,d18,0}.And finally same group cluster point will be replaced saving to get to R=12 by a node in endpoint fusion process
Point.
The endpoint cluster process based on variable coefficient tolerance band uses document " Shuxia W, Suihuai
Y.Endpoint fusing of freehand 3D object sketch with Hidden-part-draw[C]
.Computer-Aided Industrial Design&Conceptual Design,2009.CAID&CD 2009.IEEE
10th International Conference on,2009:Method disclosed in 586-590 ", basic principle are：It calculates
The average distance of each straightway head and the tail point, the measurement method include from each point to be fused to all straightway sequences
The average distance that endpoint is put to straightway head and the tail where itself, the i.e. half of length of straigh line.By line minimum in average distance
Tolerance radius of the property function as the endpoint, i.e.,
Wherein λ is tolerance radius factor, takes 1,2 ..., m；I=0,1 ..., n is wire size；J=0,1 ..., k are period,
dijIt is j-th point of i-th line member in all straightways head and the tail point sum of the distance in all straightway sequences minimum one
It is a.The size of the obtained tolerance circle of this method can change with the variation of the Global Information of input sketch, have good
Adaptability；
Step 3.3：Know that a flat and stereo has R vertex, theoretically, perspective view is centainly also corresponding, and there are R
Node { Pm；1≤m≤R }, and each node is made of three strokes, here it is considered that there are a coordinates at each node
System then shares R coordinate system.And respectively the direction of the straightway stroke where cluster point respectively represents the coordinate system at each node
Three directions.
Traverse coordinate system sequence { Ci；1≤i≤12 }, the consistent coordinate system in direction is divided into one group, then obtains 3 groups of coordinates
System, and every group of coordinate system number wj={ 8,2,2 }.According to wjSize (i.e. the size of weight shared by coordinate system group), can be by institute
There is coordinate system to be divided into 2 ranks：V1={ Ci；I=1,2,3,8,9,10,11,12 }；V2={ Ci；I=4,6；5,7}；
Step 3.4：Since weight shared by 1 grade of coordinate system (i.e. 1 grade of node) is maximum, first to the cluster in this group of coordinate system
Endpoint progress fusion treatment is the most reasonable, and the sequence then reduced successively according to weight carries out each node at the fusion of cluster point
Reason.And the principle is referred to as " coordinate system-weight criterion "；
Step 3.4.1：New coordinate system sequence { V is traversed by weight orderx；X=1,2,2 }；
Step 3.4.2：Judge CiIn participated in merging regular correction stroke number M=？(M=0,1,2,3), and to pen
It draws and was ranked up, i.e. l according to " participating in merging regular preferential storage principle "1、l2Always preferential storage has participated in merging regular
Correction stroke；
Step 3.4.3：If M=3 (K=3), l1、l2It asks and meets at P points.Work as e=0, stroke l3On it is fused regular
Endpoint d0It is constant, d1It coincides with P points；Work as e=1, stroke l3On merged regular endpoint d1It is constant, d0With { point coincides.This
In the case of kind, since three strokes both participated in processing, stroke endpoint is connected, then inevitably just will produce fusion error；
Step 3.4.4：If M=0,1,2, and K=2, then l1、l2It asks and meets at P points.
Step 3.4.5：If M=0,1,2, and K=3, then l1、l2It asks and meets at P points.And work as e=0, by l3Move to d0With point
P is overlapped；Work as e=1, by l3Move to d1It is overlapped with point P, and keeps angle constant, finally merged regular result such as Fig. 3-d
Middle dark color stroke representation.
The obtained stroke for being given by different endpoint fusion methods in order to be compared, in the present embodiment it is regular as a result,
Such as light stroke representation in Fig. 3-d.Before fusion is regular, every group of correction stroke angle, θdIt is consistent, it is desirable to regular merging
Stroke angle is still θ afterwardsd, therefore as ideal regular angle value θd.In fact, the angle of stroke claims after endpoint merges
For practical regular angle value θa, but due to merging the presence of error, produce fusion front and rear angles error delta θ=| θa-θd|,
And the standard that present aspect is excellent as endpoint blending algorithm is judged.
As shown in figure 5, providing the front and back generated angular error Δ θ's of endpoint fusion based on two kinds of algorithms of different respectively
Change line chart.By information in figure it is found that the Freehandhand-drawing flat and stereo perspective view is made of 18 strokes, blending algorithm rule of the present invention
Angular error Δ θ after whole is that middle blue broken line indicates；The algorithm of Wang is that grey broken line indicates in figure.Two kinds of sides in figure
Maximum angle error is consistent obtained by method, i.e., max (Δ θ)=5 °.Carefully analyze, optimal regular result should be Δ θ=|
θa-θd|=0 °, inventive algorithm acquired results major part stroke angular error is 0 °~1 °, and only an example is 5 °；And the calculation of Wang
Too big, the only two stroke l of Δ θ deviations fluctuating obtained by method13、l17Angular error is 0 °, seriously affects the readable of whole picture perspective view
Property and reliability, and acquired results do not meet the projection relation with three-dimension object more, are not easy to and existing Three-dimensional Gravity yet
Construction method is coupled.
This method improves the existing endpoint fusion method of online Freehandhand-drawing perspective view.This method is carried out by corresponding example
It verifies and is compared with existing algorithm, the results show is more preferable using the effect that present aspect method obtains, and regular knot
The input that fruit also further meets user is intended to.The man-machine friendship of a new generation of conceptual design is supported in the research of this method for developing
Mutual system has certain reference value, improves the regular precision of sketch recognition, also to further realize Freehandhand-drawing three-dimension object
Certain basis has been established in the reconstruct of perspective view.
Claims (2)
1. a kind of endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input, inputs as obtained by identified fitting
Gesture strokes are made of N straightway；It is characterized in that steps are as follows：
Step 1：Extract each gesture strokes { Li；1≤i≤N } angle information value { θi；0≤θi≤180°}；
Step 2：Collimation is carried out to gesture using the method based on angular histogram：
Step 2.1：Using gesture wire size i as abscissa, angle information θiFor ordinate, make histogram, to obtain gesture strokes
Angular distribution；
Step 2.2：Clustering is carried out to gesture according to the angular distribution and angle threshold Δ δ, even | θi-θj|≤Δ δ,
Then by gesture strokes Li、LjIt is divided into one group, finally obtains several groups gesture；
Step 2.3：Assuming that s group strokes are by nsGesture composition, calculates the angular average of every group of gesture strokes, then has
Wherein t, t+1 ..., nsIt indicates stroke wire size in the group of s group gestures, and works asOr
OrOrWhen,0 °, 30 °, 90 °, 150 ° of value is distinguished successively；
Then, by every gesture by angle mean value in respective groupCarry out angle correct so that angle is consistent in group, i.e.,So just achieve the purpose that collimation, and obtained stroke is known as correction stroke, correction
Strokes sequence { li(θi)；1≤i≤N}；
Step 3：Fusion treatment is carried out to correction stroke using the method based on " coordinate system-weight criterion "：
Step 3.1：All endpoint { d of extraction correction stroke firstie；1≤i≤N, e=0,1 }, i.e., point to be fused, wherein di0Table
Show stroke starting point, di1Indicate stroke terminal；
Step 3.2：The cluster of point to be fused is carried out using the endpoint clustering method based on variable coefficient tolerance band：If several to be fused
Point is mutually fallen into the tolerance circle of other side, is K per group cluster point number then by these points to be fused as a cluster；
Finally same group cluster point will be replaced by a node in endpoint fusion process, and gained number of nodes R is
And sequence node is expressed as { Pu；1≤u≤R}；
Step 3.3：R number of nodes is projected, projection node is { Pm；1≤m≤R }, projection node is made of three strokes, is formed
R coordinate system；Traverse coordinate system sequence { Cm；1≤m≤R }, the consistent coordinate system in direction is divided into one group, then is obtained several
Group coordinate system, every group of coordinate system number are wz, 1≤z≤R, and same group of coordinate system sees a coordinate system as, according to coordinate system
The shared weight w of groupzSize, obtain new coordinate system sequence { Vx；1≤x≤q }, indicate 1~q grades of coordinate systems；
Step 3.4：Successively according to weight wzThe sequence of reduction carries out each node the fusion treatment of cluster point：
Step 3.4.1：By the new coordinate system sequence { V of the order traversal of weight reductionx；1≤x≤q}；
Step 3.4.2：Judge and determines VxIn participated in merging regular correction stroke number M being 0 or 1 or 2 or 3, and to stroke
It was ranked up, i.e. l according to " participating in merging regular preferential storage principle "1、l2Always preferential storage has participated in merging regular
Correct stroke；
Step 3.4.3：If M=3, and K=3, then l1、l2It asks and meets at P points, work as e=0, stroke l3On fused regular endpoint
d30It is constant, d31It coincides with P points；Work as e=1, stroke l3On merged regular endpoint d31It is constant, d30It coincides with P points, this
In the case of kind, since three strokes both participated in processing, stroke endpoint is connected, then inevitably just will produce fusion error；
Step 3.4.4：If M=0,1,2, and K=2, then l1、l2It asks and meets at P points；
Step 3.4.5：If M=0,1,2, and K=3, then l1、l2It asks and meets at P points；And work as e=0, by l3Move to d30With point P weights
It closes；Work as e=1, by l3Move to d31It is overlapped with point P, and keeps angle constant.
2. a kind of endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input according to claim 1,
It is characterized in that angle threshold Δ δ=20 ° in step 2.2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610289106.7A CN105957046B (en) | 2016-05-04 | 2016-05-04 | The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610289106.7A CN105957046B (en) | 2016-05-04 | 2016-05-04 | The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105957046A CN105957046A (en) | 2016-09-21 |
| CN105957046B true CN105957046B (en) | 2018-11-13 |
Family
ID=56913775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610289106.7A Expired - Fee Related CN105957046B (en) | 2016-05-04 | 2016-05-04 | The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105957046B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7991561B2 (en) * | 2005-09-29 | 2011-08-02 | Roche Molecular Systems, Inc. | Ct determination by cluster analysis with variable cluster endpoint |
| US8837849B2 (en) * | 2007-06-26 | 2014-09-16 | Google Inc. | Method for noise-robust color changes in digital images |
| CN101944175B (en) * | 2009-07-07 | 2013-01-30 | 华中师范大学 | Dynamic geometric virtualization method of physical kinematics experiment |
| JP5642032B2 (en) * | 2011-08-17 | 2014-12-17 | 日本特殊陶業株式会社 | Spark plug |
| CN103902100B (en) * | 2014-04-01 | 2016-08-17 | 西北工业大学 | Stroke segmentation method based on velocity characteristic for intelligence Freehandhand-drawing input |
| CN103927056B (en) * | 2014-04-28 | 2016-08-17 | 西北工业大学 | Stroke segmentation method based on composite character for intelligence Freehandhand-drawing input |
| CN105046287B (en) * | 2015-06-29 | 2018-05-18 | 西北工业大学 | A kind of online more strokes repeat the cluster and approximating method of skeletonizing |
-
2016
- 2016-05-04 CN CN201610289106.7A patent/CN105957046B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN105957046A (en) | 2016-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100559398C (en) | Automatic deepness image registration method | |
| CN101807114B (en) | Natural interactive method based on three-dimensional gestures | |
| CN110555412B (en) | End-to-end human body gesture recognition method based on combination of RGB and point cloud | |
| CN102567703B (en) | Hand motion identification information processing method based on classification characteristic | |
| CN108921926A (en) | A kind of end-to-end three-dimensional facial reconstruction method based on single image | |
| CN101853523A (en) | Method for adopting rough drawings to establish three-dimensional human face molds | |
| CN102567716B (en) | Face synthetic system and implementation method | |
| CN109766866B (en) | Face characteristic point real-time detection method and detection system based on three-dimensional reconstruction | |
| CN102881047B (en) | Automatic non-closed implicit curved surface reconstruction method | |
| CN100369064C (en) | Human body posture deforming method based on video content | |
| Liu et al. | Facial expression recognition using pose-guided face alignment and discriminative features based on deep learning | |
| CN104123746A (en) | Calculating method for three-dimension scanning point cloud real-time normal vectors | |
| CN100585638C (en) | Curved body three-dimensional boundary representation model reconstruction method and device based on line boundary | |
| CN103500010A (en) | Method for locating fingertips of person through video | |
| CN100487732C (en) | Method for generating cartoon portrait based on photo of human face | |
| CN104134236A (en) | Boolean operation method of three-dimensional planar solids | |
| CN100504911C (en) | Method for automatically extracting face area from three-dimensional scanning original data of head and shoulder | |
| CN103093488A (en) | Virtual haircut interpolation and tweening animation producing method | |
| CN105243137A (en) | Draft-based three-dimensional model retrieval viewpoint selection method | |
| Liu et al. | BuildingSketch: Freehand mid-air sketching for building modeling | |
| CN102289661A (en) | Method for matching three-dimensional grid models based on spectrum matching | |
| CN105957046B (en) | The endpoint fusion method of flat and stereo perspective view for intelligent Freehandhand-drawing input | |
| Li et al. | DPG-Net: Densely progressive-growing network for point cloud completion | |
| Ran et al. | Normal estimation algorithm for point cloud using KD-Tree | |
| CN103218611B (en) | Based on the human body motion tracking method of distributed collaboration study |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181113 |