CN103577843A - Identification method for handwritten character strings in air - Google Patents
Identification method for handwritten character strings in air Download PDFInfo
- Publication number
- CN103577843A CN103577843A CN201310597539.5A CN201310597539A CN103577843A CN 103577843 A CN103577843 A CN 103577843A CN 201310597539 A CN201310597539 A CN 201310597539A CN 103577843 A CN103577843 A CN 103577843A
- Authority
- CN
- China
- Prior art keywords
- identification
- character string
- module
- finger trace
- gesture
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000000605 extraction Methods 0.000 claims abstract description 24
- 230000009471 action Effects 0.000 claims abstract description 6
- 238000013210 evaluation model Methods 0.000 claims description 24
- 239000012634 fragment Substances 0.000 claims description 12
- 238000012217 deletion Methods 0.000 claims description 10
- 230000037430 deletion Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 238000011156 evaluation Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 3
- 230000003993 interaction Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 7
- 238000012549 training Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000005484 gravity Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Landscapes
- Character Discrimination (AREA)
Abstract
The invention discloses an identification system and method for handwritten character strings in air. The system comprises a gesture identification module, a finger track extraction module, a pretreatment module and a character string identification module, wherein the gesture identification module is connected with the finger track extraction module and used for recognizing and identifying gestures for starting and ending writing; the finger track extraction module is connected with the pretreatment module and used for performing finger track extraction and record on input handwriting actions; the pretreatment module is connected with the in-air handwritten character string identification module and used for pre-treating finger track data; and the character string identification module is connected with the pretreatment module and used for performing character string identification on the finger track data. The invention further discloses the identification method for the handwritten character strings in air. According to the identification system and method, the supported writing mode is novel and convenient, the identification is accurate and fast, a more humanized and intelligent handwriting input mode other than the traditional handwritten mode is provided for people, and the system and the method can be widely applied to human-computer interaction systems such as game operation, television control, a teaching system and the like.
Description
Technical field
The invention belongs to computer utility and mode identification technology, the recognition methods of especially a kind of aerial hand-written character string.
Background technology
Word, as a kind of Information Communication and media of communication generally using, is being played the part of important role in man-machine interactive system.Now widely used character input modes comprises: keyboard, touch screen, handwriting pad etc.These input modes have advantage separately, but also exist deficiency separately, as keyboard is subject to size and the restriction of number of keys, the equipment such as touch screen and handwriting pad be subject to volume size restriction, to write region limited etc.Therefore, design is more natural, convenient, character input modes is an important research direction efficiently.In recent years, people are by combining handwriting identification and computer vision system to have designed the character input modes of a class based on visual gesture, wherein a kind of approach is that people normally make paper using pen write then by the ink marks of writing on camera collection paper, be identified as word, concrete grammar can list of references " M.E.Munich, P.Perona, Visual input for pen-based computers, IEEE Trans.Pattern Analysis and Machine Intelligence, 24 (3): 313-328, 2002 ", but this method has still been subject to the restriction of external condition (size of paper etc.), there is certain limitation.Another kind of approach is directly by finger, to write virtual word aloft, and camera or motion sensor, as the movement locus of the Real-time Collection fingers such as Kinect, are then word by track identification.This character input modes, owing to not write the restriction of region and ways of writing, is compared other character input modes nature and convenient more.But, owing to being difficult to judgement starting writing/starting to write in writing in this type systematic, cause between word cutting very difficult.So this type systematic can only be identified single word at present, specifically can list of references " Jin Lianwen, fourth are triumphant, Yan Hanyu, a kind of virtual character recognition method of visible detection, Chinese Patent Application No. 200810029420, publication number CN10132029A ".
The present invention proposes the recognition methods of a kind of aerial hand-written character string, user uses finger to write aloft, by camera or motion sensor collection finger motion track, by track identification, is then character string.For lacking the information of starting writing/start to write in aerial hand-written character string, cause character string to cross the problem that has unnecessary stroke after cutting, this method has proposed monobasic/binary of stroke section and has deleted evaluation model for calculating the deletion cost of stroke section.Under the framework of integrated cutting-identification, by the output of effective integration Character recognizer, character geometric model and stroke section, delete evaluation model and can effectively remove unnecessary stroke section, thereby improve the accuracy rate of character cutting and identification.
Summary of the invention
Based on above-mentioned problems of the prior art, the invention provides the recognition methods of a kind of aerial hand-written character string, the method is based on Gesture Recognition, finger tracking technique and hand-written character string recognition technology, can identify rapidly and accurately aerial hand-written character string, solve the character string identification problem of inputting by gesture.The present invention is directed to aerial hand-written character string and have the feature that connects pen between word, in the evaluation function of candidate's cutting-identification path, merged stroke section and deleted cost, can effectively remove unnecessary stroke; And based on the irrelevant path interpretational criteria of character cutting length, can effectively utilize dynamic programming algorithm fast search optimal path to obtain recognition result; In addition, under integrated cutting-identification framework, by effective integration character recognition degree of confidence, character geometric model and stroke section, delete evaluation model, further improved the precision of aerial hand-written character string identification.
According to an aspect of the present invention, disclose a kind of aerial hand-written character string recognition system, this system comprises gesture identification module, finger trace extraction module, pretreatment module, character string identification module, wherein:
Described gesture identification module is connected with described finger trace extraction module, starts the gesture of writing and finishing to write for identification marking;
Described finger trace extraction module is connected with described pretreatment module, at described gesture identification Module recognition after the gesture that starts to write, for the hand-written action of input, carry out extraction and the record of finger trace;
Described pretreatment module is connected with described aerial hand-written character string identification module, for the finger trace data that described finger trace extraction module is extracted, carries out pre-service;
Described character string identification module is connected with described pretreatment module, for the pretreated finger trace of process is carried out to character string identification.
According to a further aspect in the invention, also disclose the recognition methods of a kind of aerial hand-written character string, the method comprises the following steps:
Step 201, detects the gesture that whether exists sign to start to write;
Step 202, if the gesture that sign starts to write detected, forwards step 203 to, otherwise returns to step 201;
Step 203, real-time follow-up also gathers the track that finger is write aloft, preserves the finger trace data that collect;
Whether step 204, detect and exist sign to finish the gesture of writing, if detected, preserves described finger trace data, forwards step 205 to, otherwise return to step 203;
Step 205, carries out pre-service to preserving the finger trace data that obtain;
Step 206, based on carrying out character string identification through pretreated finger trace data, and output string recognition result;
Step 207, if the gesture that sign starts to write detected again, returns to step 201, otherwise finishes.
Aerial hand-written character string provided by the invention recognition methods, its beneficial effect is:
(1) user uses finger Free Writing aloft, makes to use gesture to control beginning and the end of writing, and by camera or motion sensor, gathers finger trace, by track identification, is character string, nature and the hommization more of the conventional book of comparing WriteMode;
(2) based on character string identification, more convenient than the input mode based on individual character identification;
(3) spend the cutting stage, considering the feature of aerial hand-written character string, effectively possible company's pen is being disconnected, do not affecting follow-up identification simultaneously;
(4) for aerial hand-written character string, cross the feature that has unnecessary stroke after cutting, proposed stroke section and deleted evaluation model, comprise that monobasic is deleted evaluation model and binary is deleted evaluation model, for calculating the deletion cost of stroke section;
(5) effectively merged Character recognizer, character geometric model and stroke section monobasic/binary and deleted evaluation model, character cutting and accuracy of identification are improved;
(6) Character recognizer of the present invention has adopted and has had the high and Nearest prototype sorter that computation complexity is low of discrimination, thereby the present invention can be met in practical application taking the requirement of the little and quick identification of storage space.
In sum, aerial hand-written character string recognition system of the present invention and method, ways of writing novelty is convenient, and identification is accurately rapid, for people provide a kind of hommization more and intelligentized handwriting input mode outside traditional handwriting mode.The present invention can be widely used in man-machine interactive system, as game operation, TV control, tutoring system etc.
Accompanying drawing explanation
Fig. 1 is the structural representation of the aerial hand-written character string of the present invention recognition system;
Fig. 2 is the process flow diagram of the aerial hand-written character string of the present invention recognition methods;
Fig. 3 carries out the example of hand-written character string identification according to the present invention.
Embodiment
For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.
Fig. 1 is the structural representation of the aerial hand-written character string of the present invention recognition system, and as shown in Figure 1, described system comprises gesture identification module, finger trace extraction module, pretreatment module, character string identification module, wherein:
Described gesture identification module is connected with described finger trace extraction module, starts the gesture of writing and finishing to write for identification marking;
In an embodiment of the present invention, for " starting to write " and " end is write " these two states have defined respectively gesture trigger event: in the face of camera, the action of " stretching out palm " represents to start to write, the action of " holding palm " represents to finish once to write, when described gesture identification Module recognition arrives this two kinds of gestures, will enter into corresponding state.These two kinds of gestures of the present embodiment definition are simple, efficient, are convenient to user and grasp and use.
Described finger trace extraction module is connected with described pretreatment module, at described gesture identification Module recognition after the gesture that starts to write, for the hand-written action of input, carry out extraction and the record of finger trace;
Described gesture identification Module recognition enters after the state of writing to the gesture that starts to write, the finger of described finger trace extraction module identification mark user writing, then in whole writing process, mobile track data pointed aloft in real-time follow-up record, comprise (X, Y) coordinate and the corresponding system time of finger in space.When described gesture identification module detects the gesture of " holding palm ", once write end, the finger trace data of described recognition system keeping records, for follow-up character string identification.
Described pretreatment module is connected with described aerial hand-written character string identification module, for the finger trace data of extracting for described finger trace extraction module, carries out pre-service;
Described pretreated object is the noise reducing in finger trace data, to facilitate follow-up character string identification.Described pre-service comprises: first described finger trace data are carried out to smooth operation, to reduce the impact that in finger tracking, the factor such as illumination variation, finger shake produces tracing point position; Subsequently, remove the redundant points in described finger trace data, make only to retain on the same position in described finger trace a tracing point.
Described character string identification module is connected with described pretreatment module, for carrying out character string identification for the pretreated finger trace of process.
Described character string identification module is word by pretreated finger trace data identification.In an embodiment of the present invention, the identification of described character string is based on candidate's cutting-identification grid, during identification, first finger trace data were carried out to cutting and obtained a series of basic fragments, by adjacent one or more basic fragment, be merged into candidate characters again, with Character recognizer, candidate characters is identified and obtained candidate's cutting-identification grid, finally in grid, search for optimum cutting-identification path and obtain final character cutting and recognition result.
Fig. 2 is the process flow diagram of the aerial hand-written character string of the present invention recognition methods, as shown in Figure 2, said method comprising the steps of:
Step 201, detects the gesture that whether exists sign to start to write;
In an embodiment of the present invention, the definition of gesture that sign starts to write is " stretching out palm ".
Step 202, if the gesture that sign starts to write detected, shows that user prepares written character string, forwards step 203 to, otherwise returns to step 201;
Step 203, real-time follow-up also gathers the track that finger is write aloft, preserves the finger trace data that collect;
User, by pointing Free Writing aloft, also gathers by camera or motion sensor real-time follow-up the track that finger is write aloft.Described finger trace data comprise: current finger point is at (X, the Y) in space coordinate figure and current system time, and described finger trace data are identified for character string afterwards.
Whether step 204, detect and exist sign to finish the gesture of writing, if detected, shows that user finishes once to write, and preserves described finger trace data, forwards step 205 to, otherwise return to step 203;
In an embodiment of the present invention, the definition of gesture that sign finishes to write is " holding palm ".
Step 205, carries out pre-service to preserving the finger trace data that obtain;
Pre-service can reduce the impact of noise in track gatherer process, thereby improves the precision of character string identification, and described pre-service specifically comprises:
1) for described finger trace data, carry out smooth operation, remove noise spot, be about to each tracing point with its on average the replacing of all tracing points in certain neighborhood around, can reduce like this impact being produced as illumination variation, finger shake etc.;
2) remove the redundant points in described finger trace data, if there are a plurality of tracing points at the same position place on track, only retain a tracing point.
Step 206, based on carrying out character string identification through pretreated finger trace data, and output string recognition result;
Described step 206 is further comprising the steps:
Step 2061, carried out cutting based on a plurality of cut-offs of crossing to the pretreated finger trace data of process, thereby obtained a series of stroke piece, be i.e. basic fragment;
Wherein, the described cut-off of crossing is comprised of following 3 class points:
1) Y coordinate Local Extremum: comprise Y coordinate Local modulus maxima and minimum point.By the Y coordinate comparison of all tracing points in certain distance before and after each tracing point and its, if the y coordinate of this point is maximal value or minimum value, select this point to cross cut-off for candidate.
2) local minimum point of horizontal projection value: first fill the white space between each tracing point by the method for linear interpolation, then projection in the horizontal direction, the projection value at coordinate Y place is that all horizontal ordinates are the number of the tracing point of X, and the point of choosing projection value and be local minimum is candidate's cut-off.
3) grow in stroke section to the join end to end point of distance maximum of straight line of stroke section: obtaining above-mentioned 1), 2) after class candidate's cut-off, choose the stroke section that course length between adjacent cut-off is greater than a certain threshold value (here threshold value be made as character string estimate 1.2 times of height), calculate in long stroke section this stroke section and join end to end the point of distance maximum of straight line as candidate's cut-off.
The method that must be cut-off described here takes full advantage of features of shape and the time feature of aerial hand-written character string, the cut-off of crossing obtaining can, by character in even pen place disconnection, not affect character recognition (fracture because of stroke does not affect the feature extraction of character global shape) below simultaneously.
Step 2062, a series of basic fragment based on obtaining builds candidate's cutting-identification grid;
Described step 2062 is further comprising the steps:
Step 20621, builds candidate characters pattern by one or more adjacent basic fragment combination;
In this step, the rule that builds candidate characters pattern is as follows: first, the candidate segment number that forms candidate characters is no more than at most 5, and basic fragment number thinks it is non-character over the candidate characters of 5, does not carry out character recognition; Secondly, the width of candidate characters can not surpass some predetermined thresholds, and the width of setting candidate characters here can not be greater than character string and estimate highly 2 times.
Wherein, estimate that the method for character string height is: travel through all tracing points, find maximal value and the minimum value of Y coordinate, both differences estimate height as character string.
Step 20622, using Character recognizer (sorter) is each candidate characters mode assignments candidate characters classification, thereby obtains candidate's cutting-identification grid.
Due to the aerial hand-written character string information of not starting writing/start to write, therefore after crossing cutting, there is unnecessary stroke, if these unnecessary stroke are not deleted the identification that may lead to errors, so in candidate's cutting-identification grid, need to judge that each candidate segment belongs to a candidate characters or unnecessary stroke, that is to say each basic fragment or belong to a candidate characters or be marked as unnecessary stroke.
In this step, the sorter of employing is to have higher recognition correct rate and non-character refusal ability, simultaneously memory space and all less Nearest prototype sorter of calculated amount; Nearest prototype sorter is by training and obtain having in a large number on the hand-written character sample of classification mark.In prototype sorter, each classification has one or more prototypes (representing with eigenvector), during classification, calculates input pattern (representing with eigenvector) to the distance of each prototype, and the minimum prototype of distance provides the classification of input pattern; Candidate's classification that the minimum a plurality of classifications of distance are input pattern.
Wherein, the feature extraction of candidate characters pattern (being expressed as eigenvector) method adopts conventional using partial stroke direction Histogram drawing method in hand script Chinese input equipment character recognition, because it is to disclose known method, here do not describe in detail, can list of references: Cheng-Lin Liu, Xiang-Dong Zhou, Online Japanese character recognition using trajectory-based normalization and direction feature extraction, Proc.10th International Workshop on Frontiers of Handwriting Recognition, La Baule, France, 2006, pp.217-222.
Step 2063, identifies at the enterprising line character string of described candidate's cutting-identification grid: according to path evaluation function, calculate the mark in each cutting-identification path, with dynamic programming algorithm search, obtain optimum cutting-recognition combination, obtain final character string recognition result.
Wherein, described path evaluation function has merged Character recognizer, character geometric model and stroke section deletes evaluation model, can adjust the effect size of each several part, and described weights can be set by experience, also can obtain by training by weights; Meanwhile, use dynamic programming method to obtain the optimal path in a plurality of cutting-identifications path by fast search, obtain final character string recognition result.
Described stroke section is deleted evaluation model and is divided into two classes: monobasic deletes evaluation model and binary is deleted evaluation model, wherein, monobasic is deleted evaluation model and is used two class linear support vector machines (SVM) to deleting the geometric properties modeling of stroke section, calculate the deletion cost of stroke section, by observing, in character string, need the stroke section of deleting to compare obvious geometric properties with other stroke section, such as their boundary rectangle frame the ratio of width to height is larger, the slope of tracing point fitting a straight line is less, horizontal projection value average center of gravity less and stroke section is less to the distance at its boundary rectangle frame center, extract accordingly four class monobasics and delete geometric properties: the ratio of width to height of the boundary rectangle frame of stroke section, the slope of fitting a straight line, horizontal projection value average and center of gravity are to the distance at boundary rectangle frame center, deletion stroke section and normal stroke section sample training that the parameter of described monobasic deletion evaluation model can be crossed with a large amount of marks obtain, binary deletion evaluation model is effectively supplementary as monobasic deletion evaluation model, deletes the binary geometric relationship (such as deletion stroke section is common and its adjacent character degree of overlapping is less) of stroke section and its adjacent character pattern can correct the deletion of part stroke section mistake by utilization.Based on this, binary delete geometric properties comprise delete stroke section and adjacent character separately width ratio, the distance between coboundary/lower boundary/left margin/right margin of boundary rectangle frame, width, catercorner length, the square root of area and 11 category features such as logarithm of the ratio of width to height of boundary rectangle frame overlapping region, described binary is deleted evaluation model and is used two class Linear SVMs to carry out modeling to the binary geometric properties of stroke section and its adjacent character pattern, and the stroke section that its model parameter serviceable indicia is crossed and the training of adjacent character equal samples obtain.
Whether described character geometric model is used for evaluating a candidate characters as a normal character.Character geometric model extracts nine features from candidate characters: the X/Y coordinate of character center of gravity, the width of boundary rectangle frame, catercorner length, the square root of area, the logarithm of the ratio of width to height and boundary rectangle frame up/down border are to the distance of character horizontal center line.Then use quadric discriminant function (QDF) modeling, give a mark to each character class.
Described character geometric model mark, character recognition mark are evaluated candidate's cutting-identification path together with deleting evaluation model mark with stroke section, can play the effect that improves character cutting and recognition performance.
In this step, the account form of path mark is as follows: by each candidate's cutting-identification path representation, be a pair of candidate characters pattern (eigenvector) sequence (O=o
1, L, o
n) and candidate characters classification sequence (C=c
1, L, c
n), the length of n delegated path wherein, the number of candidate characters namely, owing to may there being unnecessary stroke, so candidate characters classification c
imay belong to the set of significant character classification (is c
i∈ { ω
1, L, ω
m, M is the character class number that sorter determines) or belong to idle character classification (c
i=ω
0, i.e. unnecessary stroke);
Definition D (O, C) is path evaluation function, optimum cutting-identification path (O
*, C
*) be:
Path evaluation function D (O, C) has merged Character recognizer, character geometric model and stroke section and has deleted evaluation model.Being defined in the set of deleting stroke section on a path candidate is R, and not deleting the set of stroke section is N, and D (O, C) is:
In above formula, D (o
i, c
i) and
represent respectively Character recognizer output distance and character geometric model output distance;
it is the output distance that monobasic stroke section is deleted evaluation model;
With
Be the output distance that binary is deleted evaluation model, both measure respectively and delete stroke section o
iwith its adjacent strokes section o
i-1, o
i+1between binary geometric relationship; { λ
1, λ
2, λ
3three weights, and be used for different model roles in balance D (O, C), can experience set, also can obtain by training; k
ito form candidate characters c
istroke piece number; By stroke piece number, identification mark weighting is made in the evaluation function of path, be added total item and have nothing to do with cutting route length (n), thereby available dynamic programming method fast search obtains optimal path.
Step 207, if the gesture that sign starts to write detected again, returns to step 201, otherwise finishes.
According to the present invention, carry out the example of hand-written character string identification as shown in Figure 3, Figure 30 1 reads the initial trace file of an aerial hand-written character string the result that it is shown on computer screen, Figure 30 2 is the pre-service results to character string shown in Figure 30 1, Figure 30 3 crosses cutting result to the character string of Figure 30 2, the position of scheming short-and-medium vertical line sign was the position at cut-off place, Figure 30 4 is that character string is crossed candidate's cutting-identification grid example after cutting, the top display candidate characters of each square frame, square frame bottom is two candidate characters classification results, symbol " * " represents that stroke section is unnecessary stroke, should be deleted, thick line sign be the optimal path obtaining by dynamic programming algorithm search.From example, can find out, the present invention can effectively identify aerial hand-written character string.
Above-described specific embodiment; object of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (10)
1. an aerial hand-written character string recognition system, is characterized in that, this system comprises gesture identification module, finger trace extraction module, pretreatment module, character string identification module, wherein:
Described gesture identification module is connected with described finger trace extraction module, starts the gesture of writing and finishing to write for identification marking;
Described finger trace extraction module is connected with described pretreatment module, at described gesture identification Module recognition after the gesture that starts to write, for the hand-written action of input, carry out extraction and the record of finger trace;
Described pretreatment module is connected with described aerial hand-written character string identification module, for the finger trace data that described finger trace extraction module is extracted, carries out pre-service;
Described character string identification module is connected with described pretreatment module, for the pretreated finger trace data of process are carried out to character string identification.
2. system according to claim 1, is characterized in that, the finger trace data that described finger trace extraction module extracts comprise coordinate figure and the corresponding system time of finger in space.
3. system according to claim 1, is characterized in that, described pre-service comprises: for the smooth operation of finger trace data, and remove the redundant points in described finger trace data.
4. system according to claim 1, is characterized in that, first described character string identification module carried out finger trace data cutting and obtain a series of basic fragments; By adjacent one or more basic fragment, be merged into candidate characters; With Character recognizer, candidate characters is identified and obtained candidate's cutting-identification grid again; Finally in grid, search for optimum cutting-identification path and obtain final character cutting and recognition result.
5. an aerial hand-written character string recognition methods, is characterized in that, the method comprises the following steps:
Step 201, detects the gesture that whether exists sign to start to write;
Step 202, if the gesture that sign starts to write detected, forwards step 203 to, otherwise returns to step 201;
Step 203, real-time follow-up also gathers the track that finger is write aloft, preserves the finger trace data that collect;
Whether step 204, detect and exist sign to finish the gesture of writing, if detected, preserves described finger trace data, forwards step 205 to, otherwise return to step 203;
Step 205, carries out pre-service to preserving the finger trace data that obtain;
Step 206, based on carrying out character string identification through pretreated finger trace data, and output string recognition result;
Step 207, if the gesture that sign starts to write detected again, returns to step 201, otherwise finishes.
6. method according to claim 5, is characterized in that, described step 206 is further comprising the steps:
Step 2061, carried out cutting based on a plurality of cut-offs of crossing to the pretreated finger trace data of process, thereby obtained a series of stroke piece, be i.e. basic fragment;
Step 2062, a series of basic fragment based on obtaining builds candidate's cutting-identification grid;
Step 2063, identifies at the enterprising line character string of described candidate's cutting-identification grid: according to path evaluation function, calculate the mark in each cutting-identification path, with dynamic programming algorithm search, obtain optimum cutting-recognition combination, obtain final character string recognition result.
7. method according to claim 6, is characterized in that, the described cut-off of crossing comprises: the Local Extremum of Y coordinate, in the local minimum point of horizontal projection value and long stroke section to the join end to end point of distance maximum of straight line of stroke section.
8. method according to claim 6, is characterized in that, described step 2062 is further comprising the steps:
Step 20621, builds candidate characters pattern by one or more adjacent basic fragment combination;
Step 20622, is used Character recognizer, and sorter, is each candidate characters mode assignments candidate characters classification, thereby obtains candidate's cutting-identification grid.
9. method according to claim 8, is characterized in that, described path evaluation function comprises Character recognizer, character geometric model and stroke section deletion evaluation model.
10. method according to claim 9, is characterized in that,
Described stroke section is deleted evaluation model and is comprised that monobasic is deleted evaluation model and binary is deleted evaluation model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310597539.5A CN103577843B (en) | 2013-11-22 | 2013-11-22 | A kind of aerial hand-written character string recognition methods |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310597539.5A CN103577843B (en) | 2013-11-22 | 2013-11-22 | A kind of aerial hand-written character string recognition methods |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103577843A true CN103577843A (en) | 2014-02-12 |
| CN103577843B CN103577843B (en) | 2016-06-22 |
Family
ID=50049591
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310597539.5A Active CN103577843B (en) | 2013-11-22 | 2013-11-22 | A kind of aerial hand-written character string recognition methods |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103577843B (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103984943A (en) * | 2014-05-30 | 2014-08-13 | 厦门大学 | Scene text identification method based on Bayesian probability frame |
| CN104951083A (en) * | 2015-07-21 | 2015-09-30 | 石狮市智诚通讯器材贸易有限公司 | Remote gesture input method and input system |
| CN104978010A (en) * | 2014-04-03 | 2015-10-14 | 冠捷投资有限公司 | Three-dimensional space handwriting trajectory acquisition method |
| CN105094544A (en) * | 2015-07-16 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Acquisition method and device for emoticons |
| WO2015184760A1 (en) * | 2014-06-03 | 2015-12-10 | 深圳Tcl新技术有限公司 | Air gesture input method and apparatus |
| CN106446757A (en) * | 2016-05-20 | 2017-02-22 | 北京九艺同兴科技有限公司 | Human body motion data similarity automatic evaluation method |
| CN106723722A (en) * | 2016-12-26 | 2017-05-31 | 东莞理工学院 | For ring, system and the application method kept a diary |
| CN107092902A (en) * | 2016-02-18 | 2017-08-25 | 富士通株式会社 | The recognition methods of character string and system |
| CN107291215A (en) * | 2016-04-01 | 2017-10-24 | 北京搜狗科技发展有限公司 | A kind of body-sensing input message processing method and device |
| CN107305630A (en) * | 2016-04-25 | 2017-10-31 | 腾讯科技（深圳）有限公司 | Text sequence recognition methods and device |
| CN107330430A (en) * | 2017-06-27 | 2017-11-07 | 司马大大（北京）智能***有限公司 | Tibetan character recognition apparatus and method |
| CN107390880A (en) * | 2017-09-15 | 2017-11-24 | 西安建筑科技大学 | One kind is based on the contactless multi-angle input equipment of shadow and input method |
| CN107562203A (en) * | 2017-09-14 | 2018-01-09 | 北京奇艺世纪科技有限公司 | A kind of input method and device |
| CN107992792A (en) * | 2017-10-16 | 2018-05-04 | 华南理工大学 | A kind of aerial handwritten Chinese character recognition system and method based on acceleration transducer |
| CN109344793A (en) * | 2018-10-19 | 2019-02-15 | 北京百度网讯科技有限公司 | Aerial hand-written method, apparatus, equipment and computer readable storage medium for identification |
| CN109493652A (en) * | 2018-11-05 | 2019-03-19 | 广州南洋理工职业学院 | Practicing teaching system based on VR technology |
| CN109992124A (en) * | 2018-01-02 | 2019-07-09 | 北京搜狗科技发展有限公司 | Input method, device and machine readable media |
| CN110196635A (en) * | 2019-04-28 | 2019-09-03 | 浙江大学 | A kind of gesture input method based on wearable device |
| CN111580649A (en) * | 2020-04-24 | 2020-08-25 | 佛山科学技术学院 | Deep learning-based air handwriting interaction method and system |
| CN112215178A (en) * | 2020-10-19 | 2021-01-12 | 南京大学 | Chemical experiment recording system based on pen type interaction |
| CN112462935A (en) * | 2020-11-05 | 2021-03-09 | 深圳市易平方网络科技有限公司 | Text input processing method, device, terminal and medium based on remote gestures |
| CN112686134A (en) * | 2020-12-29 | 2021-04-20 | 科大讯飞股份有限公司 | Handwriting recognition method and device, electronic equipment and storage medium |
| CN113391703A (en) * | 2021-06-16 | 2021-09-14 | 咙咙信息技术(沈阳)有限公司 | System for operating air writing based on media application |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009040322A1 (en) * | 2007-09-25 | 2009-04-02 | Continental Automotive Gmbh | Method and apparatus for the contactless input of characters |
| CN101853126A (en) * | 2010-05-12 | 2010-10-06 | 中国科学院自动化研究所 | Real-time identification method for on-line handwriting sentences |
| CN102520790A (en) * | 2011-11-23 | 2012-06-27 | 中兴通讯股份有限公司 | Character input method based on image sensing module, device and terminal |
| CN103150019A (en) * | 2013-03-12 | 2013-06-12 | 深圳市国华识别科技开发有限公司 | Handwriting input system and method |
| CN103226388A (en) * | 2013-04-07 | 2013-07-31 | 华南理工大学 | Kinect-based handwriting method |
-
2013
- 2013-11-22 CN CN201310597539.5A patent/CN103577843B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009040322A1 (en) * | 2007-09-25 | 2009-04-02 | Continental Automotive Gmbh | Method and apparatus for the contactless input of characters |
| CN101853126A (en) * | 2010-05-12 | 2010-10-06 | 中国科学院自动化研究所 | Real-time identification method for on-line handwriting sentences |
| CN102520790A (en) * | 2011-11-23 | 2012-06-27 | 中兴通讯股份有限公司 | Character input method based on image sensing module, device and terminal |
| CN103150019A (en) * | 2013-03-12 | 2013-06-12 | 深圳市国华识别科技开发有限公司 | Handwriting input system and method |
| CN103226388A (en) * | 2013-04-07 | 2013-07-31 | 华南理工大学 | Kinect-based handwriting method |
Non-Patent Citations (1)
| Title |
|---|
| 赖英超，曾剑铭，沈海斌: "基于连笔消除的空间手写字符识别方法", 《计算机工程》 * |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104978010A (en) * | 2014-04-03 | 2015-10-14 | 冠捷投资有限公司 | Three-dimensional space handwriting trajectory acquisition method |
| CN103984943B (en) * | 2014-05-30 | 2018-06-19 | 厦门大学 | A kind of scene text recognition methods based on Bayesian probability frame |
| CN103984943A (en) * | 2014-05-30 | 2014-08-13 | 厦门大学 | Scene text identification method based on Bayesian probability frame |
| WO2015184760A1 (en) * | 2014-06-03 | 2015-12-10 | 深圳Tcl新技术有限公司 | Air gesture input method and apparatus |
| CN105320248A (en) * | 2014-06-03 | 2016-02-10 | 深圳Tcl新技术有限公司 | Mid-air gesture input method and device |
| CN105320248B (en) * | 2014-06-03 | 2018-12-07 | 深圳Tcl新技术有限公司 | Aerial gesture input method and device |
| CN105094544A (en) * | 2015-07-16 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Acquisition method and device for emoticons |
| CN105094544B (en) * | 2015-07-16 | 2020-03-03 | 百度在线网络技术（北京）有限公司 | Method and device for acquiring characters |
| CN104951083A (en) * | 2015-07-21 | 2015-09-30 | 石狮市智诚通讯器材贸易有限公司 | Remote gesture input method and input system |
| CN107092902B (en) * | 2016-02-18 | 2021-04-06 | 富士通株式会社 | Character string recognition method and system |
| CN107092902A (en) * | 2016-02-18 | 2017-08-25 | 富士通株式会社 | The recognition methods of character string and system |
| CN107291215A (en) * | 2016-04-01 | 2017-10-24 | 北京搜狗科技发展有限公司 | A kind of body-sensing input message processing method and device |
| CN107305630B (en) * | 2016-04-25 | 2021-03-19 | 腾讯科技（深圳）有限公司 | Text sequence identification method and device |
| CN107305630A (en) * | 2016-04-25 | 2017-10-31 | 腾讯科技（深圳）有限公司 | Text sequence recognition methods and device |
| CN106446757A (en) * | 2016-05-20 | 2017-02-22 | 北京九艺同兴科技有限公司 | Human body motion data similarity automatic evaluation method |
| CN106723722B (en) * | 2016-12-26 | 2018-08-07 | 东莞理工学院 | Ring, system for keeping a diary and application method |
| CN106723722A (en) * | 2016-12-26 | 2017-05-31 | 东莞理工学院 | For ring, system and the application method kept a diary |
| CN107330430A (en) * | 2017-06-27 | 2017-11-07 | 司马大大（北京）智能***有限公司 | Tibetan character recognition apparatus and method |
| CN107330430B (en) * | 2017-06-27 | 2020-12-04 | 司马大大（北京）智能***有限公司 | Tibetan character recognition device and method |
| CN107562203A (en) * | 2017-09-14 | 2018-01-09 | 北京奇艺世纪科技有限公司 | A kind of input method and device |
| CN107390880A (en) * | 2017-09-15 | 2017-11-24 | 西安建筑科技大学 | One kind is based on the contactless multi-angle input equipment of shadow and input method |
| CN107992792A (en) * | 2017-10-16 | 2018-05-04 | 华南理工大学 | A kind of aerial handwritten Chinese character recognition system and method based on acceleration transducer |
| CN109992124B (en) * | 2018-01-02 | 2024-05-31 | 北京搜狗科技发展有限公司 | Input method, apparatus and machine readable medium |
| CN109992124A (en) * | 2018-01-02 | 2019-07-09 | 北京搜狗科技发展有限公司 | Input method, device and machine readable media |
| CN109344793A (en) * | 2018-10-19 | 2019-02-15 | 北京百度网讯科技有限公司 | Aerial hand-written method, apparatus, equipment and computer readable storage medium for identification |
| US11423700B2 (en) | 2018-10-19 | 2022-08-23 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method, apparatus, device and computer readable storage medium for recognizing aerial handwriting |
| CN109493652B (en) * | 2018-11-05 | 2021-12-24 | 广州南洋理工职业学院 | Practice teaching system based on VR technique |
| CN109493652A (en) * | 2018-11-05 | 2019-03-19 | 广州南洋理工职业学院 | Practicing teaching system based on VR technology |
| CN110196635A (en) * | 2019-04-28 | 2019-09-03 | 浙江大学 | A kind of gesture input method based on wearable device |
| CN111580649A (en) * | 2020-04-24 | 2020-08-25 | 佛山科学技术学院 | Deep learning-based air handwriting interaction method and system |
| CN111580649B (en) * | 2020-04-24 | 2023-04-25 | 佛山科学技术学院 | Deep learning-based aerial handwriting interaction method and system |
| CN112215178A (en) * | 2020-10-19 | 2021-01-12 | 南京大学 | Chemical experiment recording system based on pen type interaction |
| CN112215178B (en) * | 2020-10-19 | 2024-05-28 | 南京大学 | Chemical experiment recording system based on pen type interaction |
| CN112462935A (en) * | 2020-11-05 | 2021-03-09 | 深圳市易平方网络科技有限公司 | Text input processing method, device, terminal and medium based on remote gestures |
| CN112686134A (en) * | 2020-12-29 | 2021-04-20 | 科大讯飞股份有限公司 | Handwriting recognition method and device, electronic equipment and storage medium |
| CN112686134B (en) * | 2020-12-29 | 2023-12-01 | 科大讯飞股份有限公司 | Handwriting recognition method, handwriting recognition device, electronic equipment and storage medium |
| CN113391703A (en) * | 2021-06-16 | 2021-09-14 | 咙咙信息技术(沈阳)有限公司 | System for operating air writing based on media application |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103577843B (en) | 2016-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103577843B (en) | A kind of aerial hand-written character string recognition methods | |
| CN106354252B (en) | A kind of continuation character gesture track recognition method based on STDW | |
| KR102460737B1 (en) | Method, apparatus, apparatus and computer readable storage medium for public handwriting recognition | |
| CN103226388B (en) | A kind of handwriting sckeme based on Kinect | |
| US10074186B2 (en) | Image search system, image search apparatus, and image search method | |
| CN103065134B (en) | A kind of fingerprint identification device and method with information | |
| CN102200834B (en) | Television control-oriented finger-mouse interaction method | |
| CN102831404B (en) | Gesture detecting method and system | |
| CN103226835B (en) | Based on method for tracking target and the system of online initialization gradient enhancement regression tree | |
| CN101853126B (en) | Real-time identification method for on-line handwriting sentences | |
| CN103150019A (en) | Handwriting input system and method | |
| CN101299236B (en) | Method for recognizing Chinese hand-written phrase | |
| CN103246891A (en) | Chinese sign language recognition method based on kinect | |
| CN103984943A (en) | Scene text identification method based on Bayesian probability frame | |
| Joshi et al. | A random forest approach to segmenting and classifying gestures | |
| CN103336967A (en) | Hand motion trail detection method and apparatus | |
| CN105022843B (en) | A kind of exchange method and system based on online handwriting | |
| Sahoo et al. | Handwritten Bangla word recognition using negative refraction based shape transformation | |
| Jin et al. | Visual gesture character string recognition by classification-based segmentation with stroke deletion | |
| CN111738167A (en) | Method for recognizing unconstrained handwritten text image | |
| CN104915009A (en) | Gesture prediction method and system | |
| US9250802B2 (en) | Shaping device | |
| CN107168633A (en) | A kind of gesture interaction query event construction method based on data dependence | |
| Li et al. | A dynamic hand gesture recognition model based on the improved dynamic time warping algorithm | |
| Tsai et al. | Reverse time ordered stroke context for air-writing recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190425 Address after: 100080 Floor 11104-2, Building 1, 66 Zhongguancun East Road, Haidian District, Beijing Patentee after: Beijing Zhongke Yueshen Technology Co., Ltd. Address before: 100190 Zhongguancun East Road, Haidian District, Haidian District, Beijing Patentee before: Institute of Automation, Chinese Academy of Sciences |