CN108846880A - A kind of cigarette quality feature visualization method - Google Patents
A kind of cigarette quality feature visualization method Download PDFInfo
- Publication number
- CN108846880A CN108846880A CN201810376303.1A CN201810376303A CN108846880A CN 108846880 A CN108846880 A CN 108846880A CN 201810376303 A CN201810376303 A CN 201810376303A CN 108846880 A CN108846880 A CN 108846880A
- Authority
- CN
- China
- Prior art keywords
- cigarette
- feature
- quality feature
- cigarette quality
- qualitative characteristics
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/20—Drawing from basic elements, e.g. lines or circles
- G06T11/206—Drawing of charts or graphs
Abstract
The present invention relates to a kind of cigarette quality feature visualization methods, belong to Cigarette evaluation technical field.This method is realized by carrying out qualitative characteristics evaluation marking, qualitative characteristics data point reuse and normalized, the multidimensional scaling characterization of face graph to cigarette to the visual observation of cigarette quality feature, classification, analysis.Finally by all kinds of face graph of comparative observation, analyze the notable feature of all kinds of types of facial makeup in Beijing operas, and according to facial feature map table, intuitive judgment is carried out to the qualitative characteristics of all kinds of samples, this method realizes the visualization to cigarette quality feature, compared with the figures methods of exhibiting such as traditional cigarette quality characteristic evaluation method or radar map by corresponding to different qualitative characteristics indexs to types of facial makeup in Beijing operas different characteristic, more intuitive, the advantages of being more readily understood differentiation is shown with figure.
Description
Technical field
The invention belongs to Cigarette evaluation technical fields, and in particular to a kind of cigarette quality feature visualization method.
Background technique
Cigarette quality be characterized in cigarette smoking impression general performance, different brands, specification cigarette have the product of oneself
Matter feature.At this stage, cigarette quality feature mainly in a manner of sensory evaluation based on, according to sensory evaluation determine cigarette quality spy
Sign.However the index currently used for the description of the qualitative characteristics of cigarette is very more, such as fragrance, throat's drying, throat's stimulation, index
It is numerous cause to cigarette quality understand it is obscure and not intuitively.
Figure helps intuitively to understand the data studied.For conventional two dimension, three-dimensional data, figure holds
It is easy to get to but being highly difficult for more high dimensional data, graphical representation and to the intuitivism apprehension of data structure.By multidimensional number
It is indicated according to planar graph, to reflect the internal feature and regularity of multivariate data, becomes multidimensional data visualization skill
Art problem to be solved.
Face graph is also known as Qie Nuofu face (Chernoff Face), it can be between the Facial expression data of face
Feature.Since the facial expression of people can leave deep impression to people, thus it is easy to distinguish, it is possible to according to expression to data
It is clustered.Face graph includes six essential characteristics:Profile, mouth, nose, eye, pupil, the eyebrow of face.This six features are by 18
Variable expression.According to the technique of painting that Chernof was proposed in 1973, using 18 indexs, the facial characteristics that each index represents is different,
According to the value of each variable, according to certain mathematical function relationship, so that it may determine the profile of face, the position of face, shape etc.,
To draw the entire types of facial makeup in Beijing operas, the difference of the types of facial makeup in Beijing operas is then distinguished using human eye, finally distinguishes different samples.
In view of the above-mentioned problems, how to use data visualization method, cigarette quality feature is intuitively analyzed, with full
The demand of sufficient cigarette product evaluation analysis is the technical problem of this field urgent need to resolve.
Summary of the invention
It is an object of the invention to the disadvantages more obscure for existing cigarette quality characteristic evaluating and description, using the types of facial makeup in Beijing operas
Figure method, provides a kind of cigarette quality feature visualization method, this method by cigarette quality characteristic visualization face graph into
Row description so that the more three-dimensional direct visualization of the evaluation result of cigarette quality feature shows, have extremely strong operability with
Application value.
To achieve the above object, the technical solution adopted by the present invention is as follows:
A kind of cigarette quality feature visualization method, includes the following steps:
Step (1) evaluates P qualitative characteristics of N number of cigarette sample by being no less than the evaluation group that 13 people form;
Wherein, 1 P≤12 <;
The cigarette sensory evaluation data of P qualitative characteristics of N number of cigarette sample is read in Matlab software, root by step (2)
The characteristic sequence in Matlab software is adjusted according to cigarette quality feature and the feature sequencing of facial feature map table,
And carry out data normalization pretreatment;
Step (3) draws cigarette sample face graph to pretreated data call Matlab function glyphplot,
Input parameter ' standardize ' value ' off ' of glyphplot;
Step (4) makees the multidimensional scaling two-dimensional projection face graph of N number of cigarette sample using multidimensional scaling;Root later
According to the distance relationship and types of facial makeup in Beijing operas appearance of face graph, classify in conjunction with clustering to sample;
Step (5), all kinds of face graph of comparative observation analyze the notable feature of all kinds of types of facial makeup in Beijing operas, and according to cigarette quality spy
Sign and facial feature map table carry out intuitive judgment to the qualitative characteristics of all kinds of samples.
It is further preferred that when evaluation group evaluates P qualitative characteristics of N number of cigarette sample, by YC/T
497-2014《Cigarette Chinese-style cigarette style sensory evaluation method》It is evaluated.
It is further preferred that P=12,12 qualitative characteristics are respectively fragrance, throat is dry, throat's stimulation, enriches
Property, oral cavity residual/dry sensation, exquisiteness/soft/mellow and full, oral stimulation/tongue calcination, flue gas concentration, convergence, miscellaneous gas, nasal cavity thorn
Sharp and strength.
It is further preferred that cigarette quality feature and facial feature map table such as table 1;
1 cigarette quality feature of table and facial feature map table
It is further preferred that multidimensional scaling two-dimensional projection face graph draws process:(1) to pretreated data,
Its dissimilarity matrix is constructed using the pdist function of matlab;(2) call the mdscale function of matlab to dissimilarity matrix
Carry out multi-dimentional scale transformation;(3) call glyphplot function, ' Centers ' parameter setting be the transformed number of multi-dimentional scale
According to two-dimensional projection's face graph can be made.
It is further preferred that the clustering uses K-means algorithm.
The present invention has following compared with the figures methods of exhibiting such as traditional cigarette style characteristic evaluation method or radar map
Advantage:
1) types of facial makeup in Beijing operas figure is shown more intuitive, is more readily understood differentiation, by being shown to the corresponding shape of face of concern index, Ji Keqing
Its index size of Chu;
2) cigarette sample of different-style feature simple and clear can be excavated by multidimensional scaling two-dimensional projection face graph
The cigarette sample of similar style and features.
3) it is characterized using types of facial makeup in Beijing operas method of the Matlab to cigarette style characteristic, can be by whole process code, calculating speed is fast,
Significant increase working efficiency and time.
4) for the present invention compared with the evaluation method of dressing face graph or exclusive use face graph, classification right judging rate improves 20-
25%;Compared with K-means types of facial makeup in Beijing operas drawing method, classification right judging rate improves 15% or so.
Detailed description of the invention
Fig. 1 is that embodiment 1 draws obtained cigarette sample face graph;
Fig. 2 is the multidimensional scaling two-dimensional projection face graph of 10 cigarette samples of embodiment 1;
Fig. 3 is that embodiment 2 draws obtained cigarette sample face graph;
Fig. 4 is the multidimensional scaling two-dimensional projection face graph of 10 cigarette samples of embodiment 2.
Specific embodiment
Below with reference to embodiment, the present invention is described in further detail.
It will be understood to those of skill in the art that the following example is merely to illustrate the present invention, and it should not be regarded as limiting this hair
Bright range.In the examples where no specific technique or condition is specified, described technology or conditions according to the literature in the art
Or it is carried out according to product description.Production firm person is not specified in material therefor or equipment, is that can be obtained by purchase
Conventional products.
0 is set as to non-evaluation index value in the embodiment of the present invention, i.e., the direction of pupil, the length of nose, mouth hang down
Straight position, mouth shape, mouth length default value be all 0.
Embodiment 1:
(1) lotus king (hard), Yellow Crane Tower (soft refined rhythm), Bright Yellow (soft great Jin circle), double happiness (flourishing age), Soviet Union's cigarette (soft gold are taken
Sand), Chinese (soft), Zhongnanhai (blue prevailing custom), white sand (and all over the world), Mount Huang (brand-new Anhui cigarette) and benefit group (red proboscis) 10 not
Same brand, different type cigarette sample, the evaluation group being made of 13 people is by YC/T 497-2014《Cigarette Chinese-style cigarette style
Sensory evaluation method》Sensory evaluation is carried out respectively to its 12 qualitative characteristics, evaluation result takes mean value, and evaluation result is shown in Table 2;
2 Chinese-style cigarette representative sample Analyses Methods for Sensory Evaluation Results of table
(2) the cigarette sensory evaluation data of 12 qualitative characteristics of 10 cigarette samples is read in into Matlab software, according to
Cigarette wind qualitative characteristics and the feature sequencing of facial feature map table are adjusted former characteristic sequence, and carry out data and return
One changes pretreatment;The data normalization preprocess method is minimax normalization method;
(3) cigarette sample face graph is drawn to pretreated data call Matlab function glyphplot,
Input parameter ' standardize ' value ' off ' of glyphplot;Obtained face graph is shown in Fig. 1;
(4) multidimensional scaling is utilized, the multidimensional scaling two-dimensional projection face graph of 10 cigarette samples, such as Fig. 2 are made;Later
According to the distance relationship of face graph, types of facial makeup in Beijing operas appearance similarity degree, classify in conjunction with Kmean algorithm to sample;
By Fig. 1,2 it is found that Zhongnanhai (blue prevailing custom) is used as one kind, Mount Huang (brand-new Anhui cigarette) is alone as one kind.Double happiness
As one kind, the two is closer in multidimensional scaling figure for (flourishing age) and Bright Yellow (soft great Jin circle).Lotus king (hard), Yellow Crane Tower
As one kind, three is closer in multidimensional scaling figure for (soft refined rhythm) and Soviet Union's cigarette (soft gold sand).Sharp group's (red proboscis) sample list
It is solely used as one kind, the width that other sample eyes are compared in face graph becomes larger, dense with flue gas known to control facial feature map table
It spends more significant.
Both Chinese (soft) and white sand (and world) are closer, face graph is more as one kind in multidimensional scaling figure
It is similar, show as being bold in face graph, eye-level is higher, width is larger between two, upper half face ellipse eccentricity is larger,
Lower half face ellipse eccentricity is larger, upper lower half face relative length is high compared with oxeye upright position, and control facial feature map table can
Know, oral stimulation/tongue calcination, oral cavity residual/dry sensation, throat dry, throat stimulation etc. fragrant preferably with fragrance, rich table
Comfort index performance is preferable.Faint scent, fruity, Xin Xiang compare other samples more with fragrant and sweet known to control facial feature map table
Significantly.
From the face graph of different size cigarette quality feature description from the point of view of, using face graph method can very visual pattern will
The qualitative characteristics of different size cigarette are described.
Embodiment 2:
(1) cloud and mist (soft big heavy nine), cloud and mist (impression), cloud and mist (soft treasure), cloud and mist (purple), cloud and mist (soft purple), jade are chosen
It is typical that small stream (villa garden), Yuxi (boundary), Yuxi (soft), Hongta (classics 100) and 10 clouds of Hongta (soft classics) produce cigarette
Representative sample, the evaluation group being made of 20 people is by YC/T 497-2014《Cigarette Chinese-style cigarette style sensory evaluation method》It is right
Its 12 qualitative characteristics carry out sensory evaluation respectively, and evaluation result takes mean value, and sensory evaluation data is shown in Table 3,4;
3 Yuxi of table, Hongta representative sample Analyses Methods for Sensory Evaluation Results
4 cloud and mist brand representative sample Analyses Methods for Sensory Evaluation Results of table
(2) the cigarette sensory evaluation data of 12 qualitative characteristics of 10 cigarette samples is read in into Matlab software, according to
Cigarette quality feature and the feature sequencing of facial feature map table are adjusted former characteristic sequence, and carry out data normalizing
Change pretreatment;The data normalization preprocess method is Zscore standardized method;
(3) cigarette sample face graph is drawn to pretreated data call Matlab function glyphplot,
Input parameter ' standardize ' value ' off ' of glyphplot;Obtained face graph is shown in Fig. 3;
(4) multidimensional scaling is utilized, makees the multidimensional scaling two-dimensional projection face graph of 10 cigarette samples, such as Fig. 4, later
According to the distance relationship of face graph, types of facial makeup in Beijing operas appearance similarity degree, classify in conjunction with K-means algorithm to sample;
By Fig. 3,4 it is found that cloud and mist (soft big heavy nine), cloud and mist (impression), Yuxi (boundary) and Yuxi (villa garden) as a kind of,
Four are closer in multidimensional scaling figure, the width larger, between two of larger, the upper lower half face ellipse eccentricity of face compared with
Greatly, the vertical range at the straight face center of eyes is larger, illustrates that these product fragrance are less than normal compared with horn of plenty, throat's stimulation and drying,
Rich preferable, qualitative characteristics are overall more excellent.
(soft big heavy nine), cloud and mist (impression), Yuxi (villa garden), Yun Xi (boundary) are as a kind of, four entirety sense organs for cloud and mist
Quality is higher, especially has outstanding performance on fragrance, fine and smooth soft and miscellaneous gas.
As one kind, three entirety aesthetic quality is general for Hongta (classics 100), cloud and mist (purple), Hongta (soft classics),
Especially score is slightly lower in stimulation and residual index.
Cloud and mist (soft treasure), Yuxi are (soft) as a kind of, and the two entirety aesthetic quality is higher, fragrance and it is rich on compared with
It is prominent.
Cloud and mist (soft purple) is alone one kind, is mainly manifested in that flue gas concentration is slightly weak, and dry sensation is slightly strong.
From the face graph of different size cigarette quality feature description from the point of view of, using face graph method can very visual pattern will
The qualitative characteristics of different size cigarette are described.
The basic principles, main features and advantages of the present invention have been shown and described above.The technology of the industry
Personnel are it should be appreciated that the present invention is not limited to the above embodiments, and the above embodiments and description only describe this
The principle of invention, without departing from the spirit and scope of the present invention, various changes and improvements may be made to the invention, these changes
Change and improvement all fall within the protetion scope of the claimed invention.The claimed scope of the invention by appended claims and its
Equivalent thereof.
Claims (6)
1. a kind of cigarette quality feature visualization method, which is characterized in that include the following steps:
Step(1), P qualitative characteristics of N number of cigarette sample are evaluated by being no less than the evaluation group that 13 people form;Its
In, 1 P≤12 <;
Step(2), the cigarette sensory evaluation data of P qualitative characteristics of N number of cigarette sample is read in into Matlab software, according to volume
Tobacco quality feature and the feature sequencing of facial feature map table are adjusted the characteristic sequence in Matlab software, go forward side by side
The pretreatment of row data normalization;
Step(3), cigarette sample face graph is drawn to pretreated data call Matlab function glyphplot,
Input parameter ' standardize ' value ' off ' of glyphplot;
Step(4), using multidimensional scaling, make the multidimensional scaling two-dimensional projection face graph of N number of cigarette sample;Later according to face
The distance relationship and types of facial makeup in Beijing operas appearance of spectrogram classify to sample in conjunction with clustering;
Step(5), all kinds of face graph of comparative observation analyzes the notable feature of all kinds of types of facial makeup in Beijing operas, and according to cigarette quality feature with
Facial feature map table carries out intuitive judgment to the qualitative characteristics of all kinds of samples.
2. cigarette quality feature visualization method according to claim 1, which is characterized in that evaluation group is to N number of cigarette
When P qualitative characteristics of sample are evaluated, by YC/T 497-2014《Cigarette Chinese-style cigarette style sensory evaluation method》Into
Row evaluation.
3. cigarette quality feature visualization method according to claim 1, which is characterized in that P=12,12 qualitative characteristics
Respectively fragrance, throat's drying, throat's stimulation, rich, oral cavity residual/dry sensation, exquisiteness/soft/mellow and full, oral stimulation/
Tongue calcination, flue gas concentration, convergence, miscellaneous gas, nasal cavity stimulation and strength.
4. cigarette quality feature visualization method according to claim 1, which is characterized in that cigarette quality feature and face
Feature Mapping table such as table 1;
1 cigarette quality feature of table and facial feature map table
。
5. cigarette quality feature visualization method according to claim 1, which is characterized in that multidimensional scaling two-dimensional projection
Face graph draws process:(1)To pretreated data, its dissimilarity matrix is constructed using the pdist function of matlab;(2)
The mdscale function of matlab is called to carry out multi-dimentional scale transformation to dissimilarity matrix;(3)Glyphplot function is called, '
Centers ' parameter setting is the transformed data of multi-dimentional scale, can make two-dimensional projection's face graph.
6. cigarette quality feature visualization method according to claim 1, which is characterized in that the clustering uses
K-means algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810376303.1A CN108846880A (en) | 2018-04-25 | 2018-04-25 | A kind of cigarette quality feature visualization method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810376303.1A CN108846880A (en) | 2018-04-25 | 2018-04-25 | A kind of cigarette quality feature visualization method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108846880A true CN108846880A (en) | 2018-11-20 |
Family
ID=64212220
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810376303.1A Pending CN108846880A (en) | 2018-04-25 | 2018-04-25 | A kind of cigarette quality feature visualization method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108846880A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111709627A (en) * | 2020-06-05 | 2020-09-25 | 江苏中烟工业有限责任公司 | Method for judging matching degree of efficacy and style characteristics of cigarette brand formula |
CN112415208A (en) * | 2020-11-17 | 2021-02-26 | 北京航空航天大学 | Method for evaluating quality of proteomics mass spectrum data |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6990238B1 (en) * | 1999-09-30 | 2006-01-24 | Battelle Memorial Institute | Data processing, analysis, and visualization system for use with disparate data types |
US20060136478A1 (en) * | 2004-12-21 | 2006-06-22 | Kathrin Berkner | Dynamic document icons |
CN104573333A (en) * | 2014-12-22 | 2015-04-29 | 长江大学 | Method for optimizing of model selection based on clustering analysis |
CN105427046A (en) * | 2015-11-24 | 2016-03-23 | 云南中烟工业有限责任公司 | Facial makeup identification method for cigarette quality features |
-
2018
- 2018-04-25 CN CN201810376303.1A patent/CN108846880A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6990238B1 (en) * | 1999-09-30 | 2006-01-24 | Battelle Memorial Institute | Data processing, analysis, and visualization system for use with disparate data types |
US20060136478A1 (en) * | 2004-12-21 | 2006-06-22 | Kathrin Berkner | Dynamic document icons |
CN104573333A (en) * | 2014-12-22 | 2015-04-29 | 长江大学 | Method for optimizing of model selection based on clustering analysis |
CN105427046A (en) * | 2015-11-24 | 2016-03-23 | 云南中烟工业有限责任公司 | Facial makeup identification method for cigarette quality features |
Non-Patent Citations (1)
Title |
---|
邓淑萍: "基于多维标度法的聚类分析", 《数理医药学杂志》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111709627A (en) * | 2020-06-05 | 2020-09-25 | 江苏中烟工业有限责任公司 | Method for judging matching degree of efficacy and style characteristics of cigarette brand formula |
CN112415208A (en) * | 2020-11-17 | 2021-02-26 | 北京航空航天大学 | Method for evaluating quality of proteomics mass spectrum data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6091836A (en) | Method for classifying features and a map representing the features | |
WO2017016059A1 (en) | Text mining-based attribute analysis method for internet media users | |
CN106021330B (en) | A kind of method for searching three-dimension model for mixing under contour line view | |
CN105046730B (en) | Written handwriting rendering method and device applied to writing brush | |
CN108846880A (en) | A kind of cigarette quality feature visualization method | |
CN105678235B (en) | Three-dimensional face expression recognition methods based on representative region various dimensions feature | |
CN110367980A (en) | EEG signals Emotion identification method based on polynary empirical mode decomposition | |
CN104899607B (en) | A kind of automatic classification method of traditional moire pattern | |
CN106960181A (en) | A kind of pedestrian's attribute recognition approach based on RGBD data | |
CN103578481A (en) | Method for recognizing cross-linguistic voice emotion | |
CN109102538B (en) | Method for extracting land characteristic points and land characteristic lines by using contour line data and weather map analysis method | |
CN108596486A (en) | A kind of cigarette style characteristic method for visualizing | |
CN106485266A (en) | A kind of ancient wall classifying identification method based on extraction color characteristic | |
Li et al. | Speech-driven cartoon animation with emotions | |
CN106709448A (en) | Face convolutional neural network-based traditional Chinese medicine constitution recognition method | |
JP3614783B2 (en) | Face classification | |
Vayadande et al. | Facial Emotion Based Song Recommendation System | |
CN110197193A (en) | A kind of automatic grouping method of multi-parameter stream data | |
Marzinotto et al. | Age and gender characterization through a two layer clustering of online handwriting | |
Zalewski et al. | Synthesis and recognition of facial expressions in virtual 3d views | |
Li et al. | Multi-network fusion based on cnn for facial expression recognition | |
CN105427046A (en) | Facial makeup identification method for cigarette quality features | |
CN106127149B (en) | A kind of groups of method and apparatus of flow chart stroke based on eye movement data | |
CN107908999A (en) | A kind of tired expression recognition method of architectural feature stratification | |
CN111310546B (en) | Method for extracting and authenticating writing rhythm characteristics in online handwriting authentication |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181120 |