CN105277663A - Facial identification method for cigarette flavor characteristics - Google Patents

Abstract

The invention discloses a facial identification method for cigarette flavor characteristics. The facial identification method comprises the following steps: 1) acquiring sensory evaluation data; 2) reading the sensory evaluation data into Matlab software and performing normalization conversion; 3) drawing cigarette sample facial graphs according to the data after the normalization pretreatment; and 4) classifying the samples according to the similarity between the cigarette sample facial graphs. The facial identification method for cigarette flavor characteristics has the advantages that the facial graphs are vivid and direct; the cigarette flavor characteristics can be vividly and visibly converted into the facial characteristics; the mental trait of human body sensitive to the tiny change of the facial characteristics is utilized to directly identify the similarity of the flavor/fragrance characteristics of the cigarette products, so that the requirement for contrastive analysis and category construction of the cigarette products is met.

Description

A kind of face recognition method of flavor type of cigarette feature

Technical field

The invention belongs to cigarette sensory evaluation technique field, be specifically related to a kind of face recognition method of flavor type of cigarette feature.

Background technology

Cigarette is a kind of specialities for people's consumption.Because the final consumption form of cigarette product is sucked exactly, the style and features of cigarette and fragrance quality will could provide impression to consumer through burning, what therefore, the most directly, the most in time, can represent consumer's will is exactly organoleptic analysis to cigarette and evaluation.The overall tasks of cigarette organoleptic analysis and evaluation and object judge exactly and describe fragrance quality level and style characteristic, be also for R&D personnel, marketing personnel provide effectively, information reliably, to make correct product and marketing decision.Due to the demand in market and the development of product, the A+E of cigarette has become tobacco business important field of research and technical work.

For the evaluation of cigarette style characteristic, tobacco business establishes corresponding evaluation criterion YC/T497-2014 " cigarette Chinese-style cigarette style sensory evaluation method ", apply the method that this standard specifies and can carry out quantitative evaluation to 15 of cigarette kind of odor type, also radar map can be adopted to obtain corresponding evaluation result figure, but be difficult to the similarity analyzing different product odor type feature according to quantitative result and radar map intuitively.

For the problems referred to above, how adopting data visualization method, analyze intuitively the similarity of the odor type feature of cigarette product and dissimilarity, to meet the demand that cigarette product comparative analysis and category build, is the technical barrier that this area waits to solve.

Summary of the invention

The object of the invention is to for the deficiencies in the prior art part, a kind of flavor type of cigarette feature visualization method is provided, thus cigarette Analyses Methods for Sensory Evaluation Results is vivo represented in face graph mode, the odor type feature similarity of each cigarette sample can be judged intuitively whether according to the shape of the types of facial makeup in Beijing operas, which aspect similar place shows, difference table is which aspect now.

Object of the present invention is achieved by following technical proposals:

1. obtain sensory evaluation data

The valuation officer of cigarette sensory evaluation adopts YC/T497-2014 " cigarette Chinese-style cigarette style sensory evaluation method " to evaluate 15 odor type features of N number of cigarette sample;

2. the cigarette sensory evaluation data of 15 of N number of cigarette sample odor type features is read in Matlab software according to the sequencing of table 1, obtain data X;

$X=\left(\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1p}\\ x_{21}& x_{22}& ...& x_{2p}\\ \·& \·& & \·\\ \·& \·& & \·\\ \·& \·& & \·\\ x_{N1}& x_{N2}& ...& x_{Np}\end{array}\right),$

In formula 1: score value evaluated by the Flue-cured Tobacco perfume that X11 represents first cigarette sample, score value evaluated by the airing cigarette cigarette perfume that X12 represents first assess sample, all the other orders according to table (2) odor type index the like; Score value evaluated by the Flue-cured Tobacco perfume that XN1 represents N number of cigarette sample, all the other orders according to table (2) odor type index the like;

Following normalization transfer pair data X is adopted to carry out pre-service:

${x^{*}}_{ik}=\frac{x_{ik}-x_{\min ,k}}{x_{\max ,k}-x_{\min ,k}}(b-a)+a,i=1,...,N,k=1,...,p,$ X in formula _ikbe the raw data of the kth comfort characteristic index of the i-th sample, x ^* _ikrepresent the data after conversion, x _{max, k}=max _1≤i≤Nx _ik, x _{min, k}=min _1≤i≤Nx _ik, a=0.05, b=0.95;

3. couple pretreated data X ^*call Matlab function glyphplot and draw cigarette sample face graph, input parameter ' standardize ' value ' off ' of glyphplot;

4. according to the similarity between cigarette sample face graph to sample classification, and call Matlab function glyphplot and draw sub-category sample face graph;

5. comparative observation face graph, analyzes the prominent feature of all kinds of types of facial makeup in Beijing operas.

The invention has the beneficial effects as follows: face graph is lively, directly perceived, very visually flavor type of cigarette feature can be converted into face feature, utilize the mankind to the psychic trait of face feature slight change sensitivity, the similarity of the odor type feature of cigarette product and dissimilarity are identified, intuitively to meet the demand of cigarette product comparative analysis.

Accompanying drawing explanation

Fig. 1 is 8 cigarette sample face graphs of the face recognition method of a kind of flavor type of cigarette feature of the present invention;

Fig. 2 is 8 cigarette sample face graphs of the sub-category display of the face recognition method of a kind of flavor type of cigarette of the present invention feature;

Fig. 3 is a kind of face recognition method operating process schematic diagram of flavor type of cigarette feature.

Embodiment

In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with the drawings and specific embodiments, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

Consult shown in Fig. 3, the inventive method chooses 8 cigarette samples as embodiment, and concrete steps are consistent with step described in summary of the invention part.

1. by sensory evaluation expert to N=8 cigarette sample, the evaluating data of p=15 odor type feature, i.e. table 1, read in Matlab software, according to the odor type feature sequencing of flavor type of cigarette feature and facial feature map table (table 1), former characteristic sequence is adjusted, obtain data variable X.

$X=\left[\begin{array}{cccccccccc}7.0& 7.5& 6.0& 5.5& 6.0& 5.5& 7.0& 5.5& 8.5& 6.5\\ 8.5& 9.0& 8.5& 8.0& 7.5& 9.5& 9.0& 8.5& 9.0& 8.0\\ 6.0& 7.0& 6.5& 5.5& 6.5& 5.5& 6.5& 5.5& 5.5& 5.0\\ 9.5& 9.0& 8.0& 8.5& 8.5& 8.5& 9.0& 9.0& 8.5& 9.5\\ 9.5& 8.5& 9.0& 9.0& 9.0& 8.5& 9.5& 7.5& 9.5& 8.5\\ 8.5& 7.5& 8.5& 8.0& 8.0& 9.0& 8.5& 7.0& 8.5& 7.5\\ 8.0& 8.5& 7.5& 7.5& 7.5& 6.5& 8.5& 8.0& 6.5& 6.5\\ 9.0& 7.5& 7.5& 7.5& 8.5& 8.5& 7.0& 9.0& 9.0& 7.0\\ 6.5& 7.0& 6.5& 5.5& 6.0& 5.5& 6.4& 4.6& 7.9& 6.0\\ 8.3& 8.0& 8.0& 7.5& 7.5& 8.5& 8.8& 7.5& 7.8& 7.0\\ 6.4& 7.2& 6.5& 5.5& 6.0& 5.5& 6.7& 5.4& 7.5& 6.0\\ 9.5& 8.7& 8.5& 8.8& 8.8& 8.6& 9.3& 8.0& 8.7& 9.0\\ 6.9& 7.5& 6.0& 5.5& 6.2& 5.5& 6.9& 5.0& 8.3& 6.5\\ 8.0& 7.6& 7.8& 7.7& 7.7& 8.5& 8.8& 7.3& 7.5& 7.4\\ 9.2& 8.0& 8.3& 8.4& 8.7& 8.5& 8.5& 8.4& 9.2& 7.0\end{array}\right]$

Following normalization transfer pair data X is adopted to carry out pre-service:

${x^{*}}_{ik}=\frac{x_{ik}-x_{min,k}}{x_{max,k}-x_{min,k}}(b-a)+a,i=1,...,N,k=1,...,p,$

X in formula _ikbe the raw data of the kth comfort characteristic index of the i-th sample, x ^* _ikrepresent the data after conversion, x _{max, k}=max _1≤i≤Nx _ik, x _{min, k}=min _1≤i≤Nx _ik, a=0.05, b=0.95.

2. couple pretreated data X ^*perform following Matlab order:

N={' sample 1', ' sample 2' ..., ' sample N'};

glyphplot(X ^*,'standardize','off','ObsLabels',n,'glyph','face')

The input parameter of glyphplot ' standardize' value ' off', glyph' value that input parameter uses ' ObsLabels' can make to increase the concrete title n of sample below each face graph, input parameter ' ' face' draws face graph, can obtain Fig. 1.

Observation Fig. 1 is known: the face of sample 3 and sample 2, eyes, nose are all more similar with face, by contrast, the types of facial makeup in Beijing operas and this two samples of sample 5, sample 6, sample 8 are in fact comparatively similar, just each position ratio is smaller, so these 5 samples can be divided into a class (class 1); Because the face graph of each types of facial makeup in Beijing operas and sample 7 differs greatly, so sample 7 is separately a class (class 2); Sample 1, the shape of face of sample 4, eyes, nose, face are more similar, and these 2 samples can be divided into a class (class 3).

3. according to the similarity between cigarette sample face graph to sample classification, and call Matlab function glyphplot and draw sub-category sample face graph, i.e. Fig. 2.

Face graph can be divided into l=3 class as shown in Figure 1.Note X _j ^*for jth class sample data, n _jfor jth class sample ID;

Perform Matlab order:

forj＝1:l

figure(j)；

glyphplot(X _j ^*,'standardize','off','ObsLabels',n _j,'glyph','face')；

end

The input parameter of glyphplot ' standardize' value ' off'.

4. the face graph that comparative observation is all kinds of, analyzes the prominent feature of all kinds of types of facial makeup in Beijing operas, and according to flavor type of cigarette feature and facial feature map table (see table 1), determines the main body odor type feature of all kinds of sample.

Observation Fig. 2 is known: in general, in class 1 sample be bold little, the width of eyes and the upright position of face all comparatively outstanding relative to other position, according to odor type feature and facial feature map table (see table 1), these 5 samples known be, delicate fragrance fragrant with Flue-cured Tobacco and fragrant and sweet be main body odor type.Class 2 sample 7 face has a certain size, and shape of face is unusual, and first face is bold than second, and the distance between two is wider, according to table 1 known it be for main body odor type with Flue-cured Tobacco perfume, the fragrance of a flower, herbal.Two each positions of sample of class 3 have some features, but outstanding especially feature, and according to table 1 and associative list 1, known they are fragrant for main body odor type with Flue-cured Tobacco, are aided with certain pungent perfume (or spice), fruital, beans perfume, milk, cure perfume and fragrant and sweet.

Table 2: raw data

The invention has the beneficial effects as follows: face graph is lively, directly perceived, very visually flavor type of cigarette feature can be converted into face feature, utilize the mankind to the psychic trait of face feature slight change sensitivity, the similarity of the odor type feature of cigarette product and dissimilarity are identified, intuitively to meet the demand of cigarette product comparative analysis.

The above, be only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, and any change of expecting without creative work or replacement, all should be encompassed in the new protection domain of this practicality and be as the criterion.

Claims (3)

1. a face recognition method for flavor type of cigarette feature, is characterized in that, comprise the following steps:

Step one, acquisition sensory evaluation data: the valuation officer of cigarette sensory evaluation adopts YC/T497-2014 " cigarette Chinese-style cigarette style sensory evaluation method " to evaluate the odor type feature of N number of cigarette sample;

Step 2, cigarette sensory evaluation data is read in Matlab software, the comfort characteristic index sequencing according to flavor type of cigarette feature and facial feature map table adjusts former characteristic sequence, line number of going forward side by side Data preprocess;

Step 3, cigarette sample face graph is drawn to pretreated data call Matlab function glyphplot;

Step 4, according to the similarity between cigarette sample face graph to sample classification, and call Matlab function glyphplot and draw sub-category sample face graph;

Step 5, comparative observation face graph, analyze the prominent feature of all kinds of types of facial makeup in Beijing operas, and according to flavor type of cigarette feature and facial feature map table, determine the main body odor type feature of all kinds of sample;

The method, by 15 of Chinese-style cigarette odor type/note features and face features, comprises the size of face, the position of the length of nose and face, eyes, eyebrow etc. maps one by one.

2. the face recognition method of flavor type of cigarette feature as claimed in claim 1, it is characterized in that, the data preprocessing method of described step 2 is:

Following normalization is adopted to convert to the sensory evaluation data of N number of sample p odor type feature: ${x^{*}}_{ik}=\frac{x_{ik}-x_{\min ,k}}{x_{\max ,k}-x_{\min ,k}}(b-a)+a,i=1,\mathrm{...},N,k=1,\mathrm{...},p,$ X in formula _ikbe the raw data of the kth odor type feature of the i-th sample, x ^* _ikrepresent the data after conversion, x _{max, k}=max _1≤i≤Nx _ik, x _{min, k}=min _1≤i≤Nx _ik, a=0.05, b=0.95.

3. the face recognition method of flavor type of cigarette feature as claimed in claim 1 or 2, it is characterized in that, the flavor type of cigarette feature of described step 2 and step 5 and facial feature map table are in table 1:

Table 1: flavor type of cigarette feature and facial feature map table

Comfort characteristic index due to Chinese-style cigarette is 15, and the face feature of mapping in Matlab is 17, does not have 2 corresponding face features automatically to get default value when Matlab draws face graph.