CN104504409B - A kind of ancient wall disease identification method based on Global Dictionary feature - Google Patents
A kind of ancient wall disease identification method based on Global Dictionary feature Download PDFInfo
- Publication number
- CN104504409B CN104504409B CN201410843650.2A CN201410843650A CN104504409B CN 104504409 B CN104504409 B CN 104504409B CN 201410843650 A CN201410843650 A CN 201410843650A CN 104504409 B CN104504409 B CN 104504409B
- Authority
- CN
- China
- Prior art keywords
- disease
- pixel
- dictionary
- super
- feature
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
Abstract
The invention discloses a kind of ancient wall disease identification method based on Global Dictionary feature, the described method comprises the following steps：Using dictionary training method to original mural painting image, according to the true value image identified in advance, online dictionary learning method is used by disease region and non-disease region, disease dictionary and non-disease dictionary is respectively trained out；Super-pixel is carried out to target image using super-pixel method；Bayesian model is established to identify each super-pixel block.It is proposed by the present invention more rapidly and effectively to carry out image segmentation using super-pixel and sparse coding.In ancient times in the disease segmentation of mural painting image, using identical Bayesian model, the sparse coding based on super-pixel can make the algorithm speed of service improve 103Times, near real-time segmentation.
Description
Technical field
The present invention relates to Computer Image Processing and area of pattern recognition, more particularly to it is a kind of based on Global Dictionary feature
Ancient wall disease identification method.
Background technology
Mural painting is the civilized witness of human history, is ancient and modern culture as one of form of painting earliest in human history
The carrier of succession.With the deduction in age, large-scale mural painting because various natures and human factor have different damages, by
This protects the field that always people explore to mural painting, and it is its pass to find and measure is targetedly made to different diseases
Key.How more preferably, the more efficient mural painting that finds traditional-handwork describes that disease obvious efficiency is low and be not easy to store and update,
Disease has great significance to mural painting protection.Meanwhile mural painting disease herein can be obtained to the continuous monitoring of every width mural painting image
Evil develops collection of illustrative plates, contacts the mural painting present position air ambient and geographic factor, can analyze varying environment factor to identical or
Influence caused by different diseases, by implementing in full image, physics, mechanics, chemical monitoring and evaluation to typical disease region, build
Vertical disease development prediction model, proposes, by the value disciplines of phenomenon to the origin cause of formation, to establish historical relic and disease scientific classification, determine wall
The risk class of picture, precipice body stability and rate of decay is horizontal, improves the monitoring efficiency and pre-alerting ability to key area disease,
Construction for World Heritage Site (mural painting) Risk Pre-control Demonstration Base provides data and basis.Technically, mural painting disease point
The effect for cutting technology is relevant with two principal elements：One is feature extraction side of the algorithm for pattern recognition under this application scenarios
Formula；, to differentiate the foundation of disease Bayesian model, model has decisive role to disease classification for another.
Feature extraction is a concept in computer vision and image procossing.It is referred to using computer extraction image
Information, determines whether the point of each image belongs to a characteristics of image.The result of feature extraction is that the point on image is divided into not
Same subset, these subsets tend to belong to isolated point, continuous curve or continuous region.The extracting mode of feature is often
Determined by problem or application type.It is characterized in the part of " interesting " in a digital picture, it is many computer pictures point
Analyse the starting point of algorithm.Therefore the whether successful feature for often being used and being defined by it of an algorithm determines.Therefore feature extraction is most
An important characteristic is " repeatability "：The feature that the different images of Same Scene are extracted should be identical.Conventional
Characteristics of image has color characteristic, textural characteristics, shape facility, spatial relation characteristics.Color histogram is common in color characteristic
Method, such as：RGB color, hsv color space, the advantage is that：The overall situation of color in piece image can be briefly described in it
Distribution, i.e. different color ratio shared in entire image, especially suitable for describe image that those are difficult to split automatically and
Without the concern for the image of object space position.Its shortcoming is：It can not describe in image the local distribution of color and every kind of
Locus residing for color, i.e., a certain specific object or object in image can not be described.
Textural characteristics are also a kind of global characteristics, and it also illustrates the superficiality of scenery corresponding to image or image-region
Matter.Different from color characteristic, textural characteristics are not based on the feature of pixel, and it is needed in the region comprising multiple pixels
Carry out statistics calculating.In pattern match, this zonal feature has larger superiority, will not be due to the deviation of part
And can not the match is successful.Common mode identification method is to image characteristic extracting method, and one kind is pixel scale, i.e., in image
Every pixel extraction feature, advantage is matching accuracy height, but speed is slow；Another kind is to divide an image into regular shape
Subgraph (Patch), feature is extracted to each subgraph, subgraph is bigger, and speed is faster, but at the same time the degree of accuracy is lower.Obviously this
Two methods can not ensure the degree of accuracy and speed simultaneously, it is therefore desirable to ensure the degree of accuracy and speed while proposing a kind of more reasonable
The feature extracting method of degree, to help the efficiency and precision of segmentation is improved.
Image recognition technology is mainly for the object that concentration is distributed in picture at this stage, such as：Animal, face, plant etc., but
In some fields, detection object is not that picture centre region is distributed in continuum, but is uniformly distributed in whole picture.
Now, existing technology such as deformable part model (Deformable Parts Model, DPM), significance analysis
(Saliency)[6], the in general detection technique such as region merging technique (Region Merge) be not suitable for this field.Meanwhile
In existing detection-segmentation problem, show good detection algorithm such as sparse coding etc. and have the problem of speed is extremely slow, because
This, the present invention proposes a kind of improved sparse coding method (Sparse Coding)[3,4], and mural painting disease is examined in ancient times for application
Survey in the particular problem of segmentation, greatly promote testing result.
The content of the invention
The invention provides a kind of ancient wall disease identification method based on Global Dictionary feature, the present invention is ensureing to calculate
On the basis of the method degree of accuracy, run time is improved, realizes the Fast Segmentation of mural painting disease, it is described below：
A kind of ancient wall disease identification method based on Global Dictionary feature, the described method comprises the following steps：
Using dictionary training method to original mural painting image, according to the true value image identified in advance, by disease region and non-
Disease region uses online dictionary learning method, and disease dictionary and non-disease dictionary is respectively trained out；
Super-pixel is carried out to target image using super-pixel method；
Bayesian model is established to identify each super-pixel.
It is described to use dictionary training method to original mural painting image, according to the true value image identified in advance, by disease region
Online dictionary learning method is used with non-disease region, disease dictionary is respectively trained out is specially with the step of non-disease dictionary：
Artificial spotting region and nontarget area in training set, by stochastical sampling obtain disease sample set and
Non- disease sample set, i.e. target sample set and non-targeted samples set；
Sliding window extraction feature is used to target sample set and non-targeted samples set.Feature extracting method is to every
One pixel takes the pixel of p × p around, just obtains the feature of the pixel after vectorization, and feature is normalized one by one
Operation；
Dictionary training is carried out using online dictionary learning method to the feature after being normalized in previous step, obtains disease dictionary
With non-disease dictionary.
It is described to establish the step of Bayesian model identifies to each super-pixel and be specially：
Extract the feature of each super-pixel block；
To each feature calculation sparse coefficient；For the feature of each super-pixel block, using sparse coding to this feature
Carry out rarefaction representation；
Establish Bayesian model.
The beneficial effect of technical scheme provided by the invention is：This method produces mural painting dictionary using supervised study, leads to
Study different diseases feature is crossed, can effectively identify the different types of disease of mural painting, while original is replaced by using super-pixel block
Beginning pixel, greatly speed up algorithm identification disease speed.Test result indicates that super-pixel block size has negative to associate with experimental result
System, i.e., super-pixel block is bigger, and Riming time of algorithm is shorter, differs bigger with Pixel-level reconstruction result；Super-pixel block is smaller, algorithm
Run time is longer, differs smaller with Pixel-level reconstruction result.By the rapid image special section based on super-pixel and sparse coding
Domain identification is applied among mural painting disease recognition, on the premise of result is ensured, speed lifting 103Times.Therefore the present invention can
It is favorably improved the efficiency of mural painting different diseases identification, near real-time segmentation.The Bayesian model of foundation, and using wall in ancient times
In the particular problem for drawing disease mark.It is proposed by the present invention more rapidly and effectively to be carried out using super-pixel and sparse coding
Image is split.In ancient times in the disease segmentation of mural painting image, using identical Bayesian model, the sparse coding based on super-pixel
The algorithm speed of service can be made to improve 103Times, near real-time segmentation.
Brief description of the drawings
Fig. 1 is a kind of flow chart of the ancient wall disease identification method based on Global Dictionary feature；
Fig. 2 is bleb disease Comparative result schematic diagram；
First is classified as original image, and each column is followed successively by true value image afterwards, the mural painting disease mark side based on Global Dictionary
Method, spectral residuals conspicuousness, inequality conspicuousness, result significantly caused by filtering.White is disease region, is below every width picture
F1-measure between true value.
Fig. 3 is that crisp alkali disease does harm to Comparative result schematic diagram.
First is classified as original image, and each column is followed successively by true value image afterwards, the mural painting disease mark side based on Global Dictionary
Method, spectral residuals conspicuousness, inequality conspicuousness, result significantly caused by filtering.White is disease region, is below every width picture
F1-measure between true value.
Embodiment
In the present invention, a kind of method of super-pixel (Superpixel) will be used to carry out over-segmentation to picture, for every
One super-pixel block extracts feature.Super-pixel carries out over-segmentation to image so that pixel has in each super-pixel block
High similarity, usual this super-pixel method speed are exceedingly fast.It can be used in various computer vision problems
Super-pixel method[1], efficiency of algorithm is improved while algorithm reliability is ensured.Using a kind of based on the efficient of figure in the present invention
Image partition method (Efficient Graph-Based Image Segmentation, EGS)[2], use the mode based on figure
The frontier distance of two super-pixel block is expressed, formulates a kind of dividing method close to linear speed.On the basis of splitting herein, this hair
The feature of the bright each super-pixel block of extraction, completes disease recognition process.
The present invention proposes a kind of quickly image-recognizing method based on Global Dictionary.Super picture is carried out to input picture first
Plain over-segmentation, the feature of each super-pixel block is then extracted, known afterwards by a kind of reconstructed error of improved sparse coding
Classify belonging to other super-pixel block, produce segmentation result.Referring to Fig. 1, concrete technical scheme includes herein below：
101：Using dictionary training method to original mural painting image, according to the true value image identified in advance, by disease region
Online dictionary learning method is used with non-disease region[5], disease dictionary and non-disease dictionary is respectively trained out；
1) artificial spotting region and nontarget area in training set, disease sample set is obtained by stochastical sampling
With non-disease sample set, i.e. target sample set and non-targeted samples set；
2) sliding window extraction feature is used to target sample set and non-targeted samples set.Feature extracting method is pair
Each pixel takes the pixel of p × p around, just obtains the feature of the pixel after vectorization, and carry out normalizing one by one to feature
Change operation because the input feature vector of online dictionary learning method usually require to meet it is claimed below：
A) average of feature substantially 0；
B) variance of different characteristic is similar each other；
C) two norms of feature are 1.
Because this samples sources is in natural image, so the feature extracted using sliding window has smooth performance
(stationarity), even if being operated without normalized square mean, condition b) also meets naturally, so only needing to carry out herein
A) and c) two operations.
3) dictionary training is carried out using online dictionary learning method to the feature after being normalized in previous step, obtains disease word
Allusion quotation and non-disease dictionary.For the feature of disease extracted region, disease dictionary D is obtained using following object function1。
Wherein, J1Represent the number of disease extracted region feature, xjRepresent j-th of feature, D1For disease dictionary, αjFor jth
Sparse coefficient corresponding to individual feature, λ are one and are manually set weighting parameter.Disease dictionary D can so be obtained1.Similarly, to non-
Disease extracted region feature, non-disease dictionary D can be obtained0。
Disease dictionary only needs to be trained once in advance with non-disease dictionary, without instructing again during mark
Practise handwriting allusion quotation.
102：Super-pixel is carried out to target image using super-pixel method；
This method carries out super-pixel segmentation using a kind of efficient image dividing method based on figure to image.Split in figure
Cheng Zhong, a segmentation S include many image-region C, and pixel has identical label in each region C.The target of this method is just
It is to find a segmentation so that the pixel in each region C has a larger similitude, and the pixel in different zones has larger difference
The opposite sex.Key step is as follows：
1) initialization figure.Start in image segmentation process, initialization figure G=<V, E>, wherein summit V expression image pixels,
When between E being two pixels；V={ v1, v2...vn, E={ e1, e2...em, it is assumed that a total of n pixel of image, then
v1Represent the 1st pixel, v2Represent the 2nd pixel, vnRepresent nth pixel.e1Represent what a two neighboring pixel was established
Side, there is weight w1(such as the two pixel luminosity equations), similarly e2Represent the side that another two neighboring pixel is established, there is weight w2
(such as the two pixel luminosity equations).
2) side E is pressed into weight ascending sort.
After E presses weight ascending sort, it is possible to prevente effectively from figure segmentation result is excessively coarse or preciosity.Figure segmentation knot
Fruit is coarse, shows as pixel in a cut zone and substantially has any different, i.e., two regions that should not merge are merged；Figure point
It is meticulous to cut, and shows as two regions and is not substantially separated but respectively；
π=EAscending order arranges{o1, o2...om}
New sequence π is obtained after side E is sorted in ascending order, element therein is still side, only the order of elements in E
Change.o1Represent the side that a two neighboring pixel is established, there are weight w (o1), while o2Represent that a two neighboring pixel is built
Vertical side, there are weight w (o2), there are w (o1)≤w(o2)≤…≤w(om)。
3) segmentation result is initialized.It is a region to make each pixel, i.e. S0={ v1, v2...vn, S0Represent to follow for the 0th time
The segmentation result that ring obtains.
4) two regions are constantly merged.After side E is sorted from small to large, whether two summits for detecting side every time belong to
In two different adjacent areas, if adjacent, the minimum weights side connected between the two regions is calculated, this weights is referred to as
The distance in two regions.
The segmentation result S of known the q-1 times circulationq-1.If oq=<vi, vj>, i.e. the q articles side oqI-th and j-th of connection
Summit, if viWith vjIn Sq-1In belong to two different regions, if the side minimum weights for having connection between the two regions are w,
If w is less than the global disparity of both regions, merge the two summit regions.The step is repeated m times until last
One side.Then we can obtain final segmentation result S=Sm。
103：Bayesian model is established to identify each super-pixel.
The rarefaction representation of signal is not new thing, for example, simplest jpeg image compression algorithm.Compressed sensing is just
It is this openness hypothesis of the signal utilized.It is inherently to have openness signal for the signal of processing, in time domain very much
Few.But certain conversion can be found so that signal has openness after some transform domain.This conversion is a lot
, most common is exactly dct transform, wavelet transformation, gabor conversion etc..Currently used is generally not orthogonal transformation, but
Based on specimen sample.Learn to obtain by great amount of images data, referred to as dictionary, each element in dictionary is referred to as atom.
The object function of study is that to find all samples in the case where the linear combination of these atoms represents be sparse, i.e., while estimates dictionary
With the two targets of the coefficient of rarefaction representation.It is assumed here that disease dictionary and background (non-disease) dictionary are by learning
Arrive, disease is single disease.In ancient times in mural painting disease mark, after obtaining the super-pixel segmentation of image, it is necessary to establish pattra leaves
This model is discriminated whether as disease to each cut zone.
1) feature of each super-pixel block is extracted；
Assuming that the super-pixel result of image, S={ C are obtained from previous step1, C2... Cr, S is segmentation
As a result, it is made up of r segmentation block, C1, C2... CrEach segmentation block is represented, is some irregular areas in image, Mei Gequ
Pixel in domain has similitude.Then feature is extracted to each super-pixel block of target image, is herein the one of the present invention
Individual innovative point, using area feature replace pixel characteristic, reduce the size of characteristic set, accelerate algorithm speed.
The feature of selected point represents the feature in region in image each segmentation block.In any one region of image
CiIn centered on any choosing a bit, take withIt is x after vectorization and normalization for the square neighborhood pixels of lengthi, this to
Amount is the feature of i-th of super-pixel block.
2) to each feature calculation sparse coefficient.It is special to this using sparse coding for the feature of each super-pixel block
Sign carries out rarefaction representation；
According to the disease dictionary obtained early stageWith background (non-disease) dictionaryCarry out sparse coding, sparse coefficient L.Different from common use associative mode word
Allusion quotation, (as used face dictionary detection face), is used herein D={ D1, D2Dictionary is used as, it is sparse using being obtained after OMP algorithms
Coefficient vector α={ α1, α2..., α2n, by this method, following Bayesian model can be established, preferably to target
Territorial classification.
3) Bayesian model is established.
Each segmentation block of the result of traversal image superpixel, compares the probability that segmentation block is target disease, with not being
The probability of target disease, probability it is big for final result.If P (y=1 | x) x corresponding regions are characterized labeled as target disease
Probability, and P (y=0 | x) it is characterized the probability that x corresponding regions are labeled as target disease.
By α={ α1, α2..., α2nFront and rear part score value is set to 0, there is δ1(α)={ α1, α2..., αn, 0,0 ..., 0 }, this
When can to obtain reconstructed error be r1=‖ x-D δ1(α)‖.Similarly, by α={ α1, α2..., α2nFirst half score value is set to 0, have
δ2(α)={ 0,0 ..., 0, αn+1, αn+2..., α2n, it is r that can now obtain reconstructed error2=‖ x-D δ2(α)‖.Classification letter
Number is defined as follows
By Bayesian formula, can write outWhereinAccording to
The reconstructed error that sparse coding process obtains, definition
Again due to finding that pixel count is smaller in super-pixel block during expression, then the super-pixel block is more possible to as disease
Region, then we can obtain finally identifying result.
In order to verify the validity of this method, picture shooting ancient wall disease, shooting are shot using high definition slr camera
Address is Dunhuang, Gansu Province city Mogao Grottoes.According to national standard, four kinds of Major Diseases regions are manually marked.On this basis, for
Every kind of disease, choose in labeled data and gather generally as training, disease dictionary and non-disease dictionary are obtained by study.In addition
Half is gathered as test.
Table 1
| Bleb disease | Crisp alkali disease evil | |
| Global Dictionary identifies | 0.467202 | 0.277887 |
| Inequality conspicuousness[7] | 0.171816 | 0.33225 |
| Spectral residuals conspicuousness[8] | 0.199615 | 0.343987 |
| Significantly filtering[6] | 0.188851 | 0.32952 |
Table 1 contrasts for Mo kao grotto at Dunhuang data acquisition system experimental result mean absolute error, and numerical value is the smaller the better.
Table 2
| Bleb disease | Crisp alkali disease evil |
| Global Dictionary identifies | 0.313172 | 0.596966 |
| Inequality conspicuousness[7] | 0.004491 | 0.013186 |
| Spectral residuals conspicuousness[8] | 0.099195 | 0.243102 |
| Significantly filtering[6] | 0.035981 | 0.131995 |
Table 2 contrasts for Mo kao grotto at Dunhuang data acquisition system experimental result F1-measure, and numerical value is the bigger the better.
By experimental result it is observed that compared with traditional sparse coding, the method for super-pixel is used in the present invention
Nothing is remarkably decreased in disease mark result, but calculating speed improves and is more than 200 times.
Bibliography
[1]Liang Li,Wei Feng,Liang Wan,Jiawan Zhang,Maximum Cohesive Grid of
Superpixels for Fast Object Localization,IEEE Computer Society Conference on
Computer Vision and Pattern Recognition(CVPR’2013).
[2]Pedro Felzenszwalb and Daniel Huttenlocher,Efficient graph-based
image segmentation.IJCV,59(2):167–181,2004.
[3]John Wright,Yi Ma,Julien Mairal,Guillermo Sapiro,Thomas Huang,and
Shuicheng Yan,Sparse Representation for Computer Vision and Pattern
Recognition,the Proceedings of the IEEE,June 2010.
[4]John Wright,Allen Yang,Arvind Ganesh,Shankar Sastry,and Yi Ma,
Robust Face Recognition via Sparse Representation,IEEE Transactions on
Pattern Analysis and Machine Intelligence(TPAMI),vol.31.no.2,February 2009.
[5]Julien Mairal,Francis Bach,Jean Ponce,Guillermo Sapiro,Online
dictionary learning for sparse coding.In Proceedings of the 26th Annual
International Conference on Machine Learning(pp.689-696).ACM.
[6]Federico Perazzi,PhilippYael Pritch,Alexander Hornung,
Saliency Filters:Contrast Based Filtering for Salient Region Detection,
Computer Vision and Pattern Recognition(CVPR),2012IEEE Conference
[7]Achanta R,Hemami S,Estrada F,et al.Frequency-tuned salient region
detection.In Computer Vision and Pattern Recognition,2009.CVPR
2009.IEEEConference on,June 2009:1597–1604.
[8]Hou X,Zhang L.Saliency Detection:A Spectral Residual Approach.In
Computer Vision and Pattern Recognition(CVPR),2007IEEE Conference on,June
2007:1–8.
Claims (1)
1. a kind of ancient wall disease identification method based on Global Dictionary feature, it is characterised in that methods described is used for blister
Exanthema is harmful and/or the mark of crisp alkali disease evil, comprises the following steps：
Using dictionary training method to original mural painting image, according to the true value image identified in advance, by disease region and non-disease
Region uses online dictionary learning method, and disease dictionary and non-disease dictionary is respectively trained out；
Super-pixel is carried out to target image using super-pixel method；
Bayesian model is established to identify each super-pixel；
It is described using dictionary training method to original mural painting image, according to the true value image identified in advance, by disease region and non-
Disease region uses online dictionary learning method, and disease dictionary is respectively trained out and is specially with the step of non-disease dictionary：
Artificial spotting region and nontarget area, disease sample set and non-disease are obtained by stochastical sampling in training set
Evil sample set, i.e. target sample set and non-targeted samples set；
Sliding window extraction feature is used to target sample set and non-targeted samples set；Feature extracting method is to each
Pixel takes the pixel of p × p around, just obtains the feature of the pixel after vectorization, and feature is normalized one by one operation；
Dictionary training is carried out using online dictionary learning method to the feature after being normalized in previous step, obtains disease dictionary and non-
Disease dictionary；
It is described to establish the step of Bayesian model identifies to each super-pixel and be specially：
Extract the feature of each super-pixel block；
To each feature calculation sparse coefficient；For the feature of each super-pixel block, this feature is carried out using sparse coding
Rarefaction representation；
Establish Bayesian model；
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<mi>y</mi>
<mo>=</mo>
<mover>
<mi>y</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>1</mn>
<mo>-</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mi>x</mi>
<mo>-</mo>
<mover>
<mi>y</mi>
<mo>^</mo>
</mover>
<msub>
<mi>r</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mover>
<mi>y</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<msub>
<mi>r</mi>
<mn>2</mn>
</msub>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mn>2</mn>
<mn>2</mn>
</msubsup>
</mrow>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mi>x</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mn>2</mn>
<mn>2</mn>
</msubsup>
</mrow>
</mfrac>
</mrow>
P is probability,r1、r2For reconstructed error, x and y are super-pixel block feature；
Wherein, it is described to be specially to target image progress super-pixel using super-pixel method：
Initialization figure G=<V, E>, summit V represents image pixel, between E is two pixels while；
Side E is pressed into weight ascending sort；Whether two summits on detection side belong to two different adjacent areas every time, if phase
Neighbour, the side minimum weights for having connection between two regions are w, if w is less than the global disparity of both regions, merge two tops
Point region；Repeat until last side, obtains final segmentation result；
Methods described is when identifying bleb disease, mean absolute error 0.467202；When identifying crisp alkali disease evil, average absolute
Error is 0.277887.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410843650.2A CN104504409B (en) | 2014-12-30 | 2014-12-30 | A kind of ancient wall disease identification method based on Global Dictionary feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410843650.2A CN104504409B (en) | 2014-12-30 | 2014-12-30 | A kind of ancient wall disease identification method based on Global Dictionary feature |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104504409A CN104504409A (en) | 2015-04-08 |
| CN104504409B true CN104504409B (en) | 2018-01-05 |
Family
ID=52945805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410843650.2A Active CN104504409B (en) | 2014-12-30 | 2014-12-30 | A kind of ancient wall disease identification method based on Global Dictionary feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104504409B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957035A (en) * | 2016-04-29 | 2016-09-21 | 广东小天才科技有限公司 | Removing method and device of picture marks, and intelligent device |
| CN108446017B (en) * | 2018-02-12 | 2021-04-23 | 天津大学 | MR (magnetic resonance) glasses-based ancient mural disease visual analysis method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831614A (en) * | 2012-09-10 | 2012-12-19 | 西安电子科技大学 | Sequential medical image quick segmentation method based on interactive dictionary migration |
| CN102879099A (en) * | 2012-08-08 | 2013-01-16 | 北京建筑工程学院 | Wall painting information extraction method based on hyperspectral imaging |
| CN103295199A (en) * | 2013-05-29 | 2013-09-11 | 西安建筑科技大学 | Intelligent repair assistance system for cracks of ancient wall murals |
-
2014
- 2014-12-30 CN CN201410843650.2A patent/CN104504409B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102879099A (en) * | 2012-08-08 | 2013-01-16 | 北京建筑工程学院 | Wall painting information extraction method based on hyperspectral imaging |
| CN102831614A (en) * | 2012-09-10 | 2012-12-19 | 西安电子科技大学 | Sequential medical image quick segmentation method based on interactive dictionary migration |
| CN103295199A (en) * | 2013-05-29 | 2013-09-11 | 西安建筑科技大学 | Intelligent repair assistance system for cracks of ancient wall murals |
Non-Patent Citations (2)
| Title |
|---|
| "Knowledge Based Lacunas Detection and Segmentation for Ancient Paintings";jianming Liu etc，;《Proceeding Vsmm 2007 Proceedings of the 13th Virtual Systems and Multimedia, International Conference》;20090923;参见摘要、第2.1-2.2节、图2、图6 * |
| "Online dictionary learning for sparse coding";Mairal J etc,;《International Conference on Machine Learning, ICML 2009》;20090614;第689-696页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104504409A (en) | 2015-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111126360B (en) | Cross-domain pedestrian re-identification method based on unsupervised combined multi-loss model | |
| CN105528794A (en) | Moving object detection method based on Gaussian mixture model and superpixel segmentation | |
| CN104239902B (en) | Hyperspectral image classification method based on non local similitude and sparse coding | |
| CN109657610A (en) | A kind of land use change survey detection method of high-resolution multi-source Remote Sensing Images | |
| CN106408030A (en) | SAR image classification method based on middle lamella semantic attribute and convolution neural network | |
| CN105513066B (en) | It is a kind of that the generic object detection method merged with super-pixel is chosen based on seed point | |
| CN106991370A (en) | Pedestrian retrieval method based on color and depth | |
| CN109344842A (en) | A kind of pedestrian's recognition methods again based on semantic region expression | |
| CN111507296A (en) | Intelligent illegal building extraction method based on unmanned aerial vehicle remote sensing and deep learning | |
| CN104463248A (en) | High-resolution remote sensing image airplane detecting method based on high-level feature extraction of depth boltzmann machine | |
| Yao et al. | Sensing urban land-use patterns by integrating Google Tensorflow and scene-classification models | |
| Chen et al. | Agricultural remote sensing image cultivated land extraction technology based on deep learning | |
| Jayakumar et al. | Automatic prediction and classification of diseases in melons using stacked RNN based deep learning model | |
| Alnihoud | Content-based image retrieval system based on self organizing map, fuzzy color histogram and subtractive fuzzy clustering. | |
| CN106778897B (en) | Plant species recognition methods twice based on COS distance and center profile distance | |
| Li et al. | Lightweight automatic identification and location detection model of farmland pests | |
| CN104504409B (en) | A kind of ancient wall disease identification method based on Global Dictionary feature | |
| CN110008828A (en) | Pairs of constraint ingredient assay measures optimization method based on difference regularization | |
| CN109741351A (en) | A kind of classification responsive type edge detection method based on deep learning | |
| CN113723558A (en) | Remote sensing image small sample ship detection method based on attention mechanism | |
| CN107944340A (en) | A kind of combination is directly measured and the pedestrian of indirect measurement recognition methods again | |
| CN107452001A (en) | A kind of remote sensing images sequences segmentation method based on improved FCM algorithm | |
| CN105844299B (en) | A kind of image classification method based on bag of words | |
| Yang et al. | Automatically adjustable multi-scale feature extraction framework for hyperspectral image classification | |
| CN104484676B (en) | A kind of interactive mode ancient wall disease identification method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |