CN101996328B - Wood identification method - Google Patents
Wood identification method Download PDFInfo
- Publication number
- CN101996328B CN101996328B CN201010506395A CN201010506395A CN101996328B CN 101996328 B CN101996328 B CN 101996328B CN 201010506395 A CN201010506395 A CN 201010506395A CN 201010506395 A CN201010506395 A CN 201010506395A CN 101996328 B CN101996328 B CN 101996328B
- Authority
- CN
- China
- Prior art keywords
- image
- timber
- cluster
- feature
- weight
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 239000002023 wood Substances 0.000 title claims abstract description 26
- 238000012216 screening Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 4
- 238000012545 processing Methods 0.000 claims description 12
- 239000011347 resin Substances 0.000 claims description 7
- 229920005989 resin Polymers 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 239000012141 concentrate Substances 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 claims 1
- 238000000638 solvent extraction Methods 0.000 abstract 1
- 240000005020 Acaciella glauca Species 0.000 description 5
- 235000003499 redwood Nutrition 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000002329 infrared spectrum Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 240000007711 Peperomia pellucida Species 0.000 description 1
- 235000012364 Peperomia pellucida Nutrition 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a wood identification method. The method comprises the following steps of: (1) generating characteristic quantities, namely partitioning a wood image, extracting local characteristics of each partitioned image, combining the local characteristics of a plurality of partitioned images to form the characteristics of the wood image, clustering and merging the characteristics of the wood image by clustering, extracting clustered information, and on the basis, screening the clustered information to form the characteristic quantities of the wood image; (2) repeating the step (1) for each image in an N original wood images-containing training sample set so as to generate an N characteristic quantities-containing characteristic set; (3) repeating the step (1) for the wood images to be identified so as to generate the characteristic quantities of the wood images to be identified; and (4) calculating the characteristic quantities of the wood images to be identified and the characteristic set by an earth mover's distance (EMD) algorithm to generate N EMDs, and classifying the N EMDs to obtain an identification result. The method has higher identification rate.
Description
Technical field
The present invention relates to a kind of Wood Identification Method, especially a kind of supplementary means that uses a computer, based on the Wood Identification Method of analyzing image texture.
Background technology
The early stage method of timber identification mainly is to rely on artificial experience and knowledge, and combines some aids, like key, verge-perforated card; And for example the day for announcing is on October 1st, 1986, and notification number is the disclosed number timber of the one Chinese patent application of a CN86200457U identification card, or the like; According to macrofeature and microscopic feature, through observing, relatively and analyze and progressively identify identification timber, this method receives that assessor's the influence of subjective factor is very big, discrimination is low; Cause some precious, timber that price is higher or that have specific use, often be taken as generally and handle, as for generally using timber with material; Also can occur not only waiting each link to make troubles, influence the raising of forest industrial enterprise economic interests to timber producing and selling and use because of discerning the unclear phenomenon that causes blindness to use, misuse; Also wasted valuable resource; In addition, because it is very long to rely on artificial recognition methods to expend time in, also caused waste of manpower resource to a certain extent.
In recent years; The identification of area of computer aided timber becomes a kind of Wood Identification Method of main flow gradually; As Granted publication day be on August 5th, 2009, notification number is the near infrared spectrum recognition methods of the redwood announced of the Chinese patent of 100523793C, adopts redwood and non-redwood wood sample exactly; Utilize the diverse location collection several near infrared spectrum of near infrared spectrum equipment on the wood sample surface; Gather spectrum for same sample respectively 3～10 positions, pass through the spectrum pre-service again, after level and smooth, baseline correction, first order derivative, second derivative, polynary scatter correction or the pre-service of data dimensionality reduction; Through multivariate data analysis methods such as soft independent modeling classification or least square discriminatory analyses; Set up the discrimination model of true and false redwood and timber varieties of trees thereof, thereby, utilize the model of setting up can realize why adding the differentiation of redwood and timber varieties of trees.This method has higher analysis efficiency and accuracy, but since its based on sample be that timber is in kind, for some timber that is difficult for the sample circuit carrying, aspect practical operation, exist certain difficulty.Based on lumber fibre or stereogram picture; Adopt the recognition methods of machine vision means to overcome the problems referred to above; As everyone knows, timber identification often determines by the zone of the key feature in the timber picture, and in the timber picture contain fracture, resin canal, the noise region of characteristic such as damage by worms, go mouldy; Then the identification to timber has certain inhibiting effect; Existing this type of recognition methods does not have effectively to give prominence to the effect in timber key feature zone, can not well eliminate the influence of noise region, thereby reduce discrimination greatly.
Summary of the invention
The objective of the invention is to overcome in the prior art based on the lower problem of the recognition methods discrimination of timber picture, proposed a kind of Wood Identification Method with higher discrimination.
For achieving the above object, the technical scheme that the present invention adopted is:
A kind of Wood Identification Method comprises the steps:
(1) characteristic quantity generates: a timber image block is handled; Extract the local feature of every block image, and the local feature of a plurality of piecemeals is combined to form the characteristic of timber image, the characteristic of timber image is carried out the cluster merger with clustering processing; Extract clustering information; On this basis, clustering information is carried out Screening Treatment, form the characteristic quantity of timber image;
(2) to including concentrated every the image repeating step (1) of training sample that N opens the timber original image, generate a feature set that comprises N characteristic quantity;
(3), generate the characteristic quantity of timber image to be identified to timber image repeating step to be identified (1);
(4) characteristic quantity and the feature set with timber image to be identified generates N EMD distance with the EMD algorithm computation, and N EMD distance carried out classification processing, draws recognition result.
Technique scheme can also be further perfect:
As preferably, further comprise a pre-service substep of handling to carry out prior to piecemeal in the step (1): choose the subregion of a w * h at the obvious position of characteristic of timber image, and gradation conversion is carried out in this zone, form pretreatment image.
As preferably, the obvious position of characteristic be meant between two annual ring lines of timber, do not contain fracture, resin canal, any one the position in four of damaging by worms, go mouldy at least.
As preferably, piecemeal handle with pretreatment image according to horizontal m piece, vertically the n piece is divided into mn identical block image, makes mn=m * n.
As preferably, use the Gabor wavelet algorithm to extract the characteristic of every block image, and form the characteristic of timber image, Gabor handles and comprises the steps: that (1) carry out the Gabor conversion of s yardstick, an o direction one by one respectively to every block image; (2) eigenwert of every block image of extraction; (3) make os=o * s, every block image is obtained the proper vector that os eigenwert is arranged in 1 * os dimension; (4) every pretreatment image is obtained the eigenmatrix that the capable proper vector of mn is arranged in a mn * os, be called the Gabor characteristic, called after G.
As preferably, the value of s, o is respectively 4,6, or 5,8.
As preferably, clustering processing comprises the steps: (1) with each row of proper vector G point as the os dimension space, and every pretreatment image forms mn point; (2) with clustering algorithm mn point gathered into k cluster, a point on the corresponding os dimension space of each cluster, i.e. k os dimensional vector; (3) make cluster cluster=[cluster
1, cluster
2..., cluster
k]; (4) write down each cluster and comprise number n um=[num a little
1, num
2..., num
k]; (5) claim cluster, num is a clustering information.
As preferably, Screening Treatment comprises the steps: that (1) makes weight
i=num
i/ mn; (2) weight is once arranged from big to small, i weight makes their ∑ that satisfies condition just before filtering out
1 iWeight
i>=M (i≤k); (3) make feature=[feature
1, feature
2..., feature
i], feature wherein
i=cluster
i(4) make weight=[weight
1, weight
2..., weight
i], claim weight
iBe feature
iWeight; (5) make signature=[s
1, s
2..., s
i], s wherein
i=(feature
i, weight
i), signature is the characteristic quantity of timber image.
As preferably, the M value is 90%.
As preferably, classification processing adopts the SVM algorithm.
Because the employing of technique scheme, the present invention compared with prior art has the following advantages:
The present invention is through the reasonable piecemeal refinement to the timber picture; Independent each characteristic area (comprising the critical area that has key feature, common texture region, the region of fracture, resin canal zone, the zone of damaging by worms, the zone etc. of going mouldy); Adopt key feature zone and common texture characteristic area content as block image; And every piecemeal pattern extracted characteristic with the Gabor small echo, this way has been strengthened the effect in key feature zone, has solved because of whole image being carried out the Gabor small echo to handle and lowers the regional problem that acts on of key feature; And the noise region of can contain fracture, resin canal through removal, damaging by worms, going mouldy improves discrimination.
In addition, the present invention has introduced clustering algorithm, and the characteristic that the Gabor small echo is extracted concentrates, and has reduced the operand that the piecemeal refinement brings for follow-up identifying.
Moreover; Because the picture in zone such as contain fracture, resin canal, damage by worms, go mouldy has been given up by part in picture pre-treatment step process; So these zones shared proportion in whole Zhang Mucai wood chip is very little; Again the characteristic after the clustering processing is screened through suitable threshold is set, can remove these zones effectively, further improve discrimination.
In addition, another advantage of Screening Treatment is when having kept important crucial recognition feature, to have kept original discrimination, and through removing non-key character, improved the execution efficient of identifying.
At last; With existing find the solution the method for characteristic similarity between image different be; Classic method is to discharge characteristics of image in certain sequence earlier; Then a plurality of characteristics of different images are carried out the man-to-man similarity (or distance) of simply asking one by one; At last these similarities (or distance) are added up summation as the final foundation of judging the different images similarity degree, its shortcoming is the influence that the similarity of different images characteristic is put in order by characteristics of image, can not make rational judgement to the relative degree of characteristics of image.And the present invention adopts EMD (Earth Mover ' s Distance; This measure derives from famous cargo transfer problem; It is a kind of characteristic similarity that can calculate different picture effectively and comprised; And carry out the part Matching Algorithm) come the similarity of computed image characteristic, can solve the aforementioned problems in the prior effectively, eliminated characteristics of image and put in order and calculate the adverse effect of bringing to similarity.
Description of drawings
Fig. 1 is a kind of process flow diagram of the present invention.
Embodiment
Below in conjunction with accompanying drawing, the present invention is further specified:
Wood Identification Method as shown in Figure 1 mainly comprises the steps:
(1) characteristic quantity generates: a timber image block is handled; Extract the local feature of every block image, and the local feature of a plurality of piecemeals is combined to form the characteristic of timber image, the characteristic of timber image is carried out the cluster merger with clustering processing; Extract clustering information; On this basis, clustering information is carried out Screening Treatment, form the characteristic quantity of timber image;
(2) to including concentrated every the image repeating step (1) of training sample that N opens the timber original image, generate a feature set that comprises N characteristic quantity;
(3), generate the characteristic quantity of timber image to be identified to timber image repeating step to be identified (1);
(4) characteristic quantity and the feature set with timber image to be identified generates N EMD distance with the EMD algorithm computation, and N EMD apart from adopting the SVM algorithm to carry out classification processing, drawn recognition result.
EMD (earth mover ' s distance) is based on a kind of of famous cargo transfer problem (Hitchcock, 1941) and can be used for effectively calculating the characteristic similarity that different picture comprised, and can carry out a kind of algorithm of part coupling.
Its formula is following:
Wherein P, Q represent two signature, d
IjBe the ground distance between characteristic i, the j, f
IjBe the flow of i to j.
In addition, in step (1), also need carry out a pre-service substep of handling to carry out prior to piecemeal: choose the subregion of a w * h at the obvious position of characteristic of timber image, and gradation conversion is carried out in this zone, form pretreatment image.The obvious position of characteristic be meant between two annual ring lines of timber, do not contain fracture, resin canal, any one the position in four of damaging by worms, go mouldy at least, in the present embodiment, w, h value are 100, pretreatment image is 256 grades of gray level images.
Above-mentioned piecemeal is handled pretreatment image is divided into mn identical block image according to horizontal m piece, vertical n piece, makes mn=m * n, and in the present embodiment, the span of m, n is 4～7.
Use the Gabor wavelet algorithm to extract the characteristic of every block image, and form the characteristic of timber image, Gabor handles and comprises the steps:
(1) respectively every block image is carried out the Gabor conversion of s yardstick, an o direction one by one;
(2) eigenwert of every block image of extraction;
(3) make os=o * s, every block image is obtained the proper vector that os eigenwert is arranged in 1 * os dimension; In the present embodiment, the value of s, o is respectively 4,6, or 5,8;
(4) every pretreatment image is obtained the eigenmatrix that the capable proper vector of mn is arranged in a mn * os, be called the Gabor characteristic, called after G.
The Gabor small echo is the unique Gabor function that can obtain spatial domain and frequency-domain combined uncertainty relation lower limit as the strong instrument of multiple dimensioned expression of image and analysis, often is used as wavelet basis function, and image is carried out various analyses.
The kernel function of the two-dimensional Gabor Wavelets that the present invention adopted is following:
Wherein, u, v are respectively direction and scale factor, are row vector, and x, y are two-dimensional coordinate.
is the centre frequency of wave filter, and
embodied the directional selectivity of wave filter.
Clustering processing comprises the steps:
(1) with each capable point as the os dimension space of proper vector G, every pretreatment image forms mn point;
(2) with clustering algorithm mn point gathered into k cluster, a point on the corresponding os dimension space of each cluster, i.e. k os dimensional vector;
(3) make cluster cluster=[cluster
1, cluster
2..., cluster
k];
(4) write down each cluster and comprise number n um=[num a little
1, num
2..., num
k]
(5) claim cluster, num is a clustering information.
As preferably, Screening Treatment comprises the steps:
(1) makes weight
i=num
i/ mn;
(2) weight is once arranged from big to small, i weight makes their ∑ that satisfies condition just before filtering out
1 iWeight
i(i≤k), in the present embodiment, the M value is 90% to>=M;
(3) make feature=[feature
1, feature
2..., feature
i], feature wherein
i=cluster
i
(4) make weight=[weight
1, weight
2..., weight
i], claim weight
iBe feature
iWeight;
(5) make signature=[s
1, s
2..., s
i], s wherein
i=(feature
i, weight
i), signature is the characteristic quantity of timber image.
Should be understood that this embodiment only to be used to the present invention is described and be not used in the restriction scope of the present invention.Should be understood that in addition those skilled in the art can do various changes or modification to the present invention after the content of having read the present invention's instruction, these equivalent form of values fall within the application's appended claims institute restricted portion equally.
Claims (6)
1. a Wood Identification Method is characterized in that, comprises the steps:
(1) chooses the subregion of a w * h at the obvious position of characteristic of timber image; And this zone carried out gradation conversion; Form pretreatment image; The obvious position of said characteristic be meant between two annual ring lines of timber, do not contain fracture, resin canal, any one the position in four of damaging by worms, go mouldy at least, with described pretreatment image according to horizontal m piece, vertically the n piece is divided into mn identical block image, makes mn=m * n;
(2) use the Gabor wavelet algorithm to extract the local feature of every block image; And the local feature of a plurality of block images is combined to form the characteristic of said timber image; With clustering processing the characteristic of described timber image is carried out the cluster merger; Extract clustering information, clustering information is carried out Screening Treatment, form the characteristic quantity of said timber image; Wherein,
The local feature of every block image of said extraction specifically comprises the steps: one by one every block image to be carried out the Gabor conversion of s yardstick, an o direction, obtains os eigenwert of every block image, makes os=o * s; Every said block image is obtained the proper vector that os eigenwert is arranged in 1 * os dimension; Every described pretreatment image is obtained the eigenmatrix G that the capable proper vector of mn is arranged in a mn * os;
Described clustering processing comprises the steps: that specifically every described pretreatment image forms mn point with each capable point as the os dimension space of eigenmatrix G; With clustering algorithm mn point gathered into k cluster, a point on the corresponding os dimension space of each cluster, i.e. k os dimensional vector; Make cluster cluster=[cluster
1, cluster
2..., cluster
k], write down each cluster and comprise number n um=[num a little
1, num
2..., num
k], cluster, num are described clustering information;
Described Screening Treatment specifically comprises the steps: to make weight
i=num
i/ mn; Weight is once arranged from big to small, and i weight makes their ∑s that satisfies condition before filtering out
1 iWeight
i>=M (i≤k); Make feature=[feature
1, feature
2..., feature
i], feature wherein
i=cluster
iMake weight=[weight
1, weight
2..., weight
i], claim weight
iBe feature
iWeight; Make the characteristic quantity signature=[s of said timber image
1, s
2..., s
i], s wherein
i=(feature
i, weight
i);
(3) carry out above-mentioned steps (1) and step (2) to including every image that training sample that N opens the timber original image concentrates, generate a feature set that comprises N characteristic quantity;
(4) timber image to be identified is carried out above-mentioned steps (1) and step (2), generate the characteristic quantity of timber image to be identified;
(5) characteristic quantity of timber image to be identified and described feature set being used Earth Mover ' s Distance is that the EMD algorithm computation generates N EMD distance, and said N EMD apart from carrying out classification processing, drawn recognition result.
2. Wood Identification Method according to claim 1 is characterized in that: described w, h value are 100, and described pretreatment image is 256 grades of gray level images.
3. Wood Identification Method according to claim 1 and 2 is characterized in that: the span of said m, n is 4～7.
4. Wood Identification Method according to claim 1 is characterized in that the value of s, o is respectively 4,6 or 5,8.
5. Wood Identification Method according to claim 1 is characterized in that, the M value is 90%.
6. according to claim 1 or 2 or 3 described Wood Identification Method, it is characterized in that: described classification processing adopts SVMs SVM algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010506395A CN101996328B (en) | 2010-10-14 | 2010-10-14 | Wood identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010506395A CN101996328B (en) | 2010-10-14 | 2010-10-14 | Wood identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101996328A CN101996328A (en) | 2011-03-30 |
| CN101996328B true CN101996328B (en) | 2012-10-10 |
Family
ID=43786466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010506395A Expired - Fee Related CN101996328B (en) | 2010-10-14 | 2010-10-14 | Wood identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101996328B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102214305A (en) * | 2011-04-08 | 2011-10-12 | 大连理工大学 | Method for taking evidence for source of printing paper sheet by using grain characteristic |
| CN104794486B (en) * | 2015-04-10 | 2018-10-16 | 电子科技大学 | Video smoke detection method based on multi-feature fusion |
| CN105354823A (en) * | 2015-09-28 | 2016-02-24 | 佛山市朗达信息科技有限公司 | Tree-ring image edge extraction and segmentation system |
| CN108229314B (en) * | 2017-11-28 | 2021-05-04 | 深圳市商汤科技有限公司 | Target person searching method and device and electronic equipment |
| CN108262809B (en) * | 2017-12-14 | 2021-06-29 | 北京木业邦科技有限公司 | Wood board processing method and device based on artificial intelligence, electronic equipment and medium |
| CN108256550A (en) * | 2017-12-14 | 2018-07-06 | 北京木业邦科技有限公司 | A kind of timber classification update method and device |
| CN108663360A (en) * | 2018-04-23 | 2018-10-16 | 安徽农业大学 | The chemical identification method of the unapparent tree age of growth ring |
| CN112723076A (en) * | 2021-01-07 | 2021-04-30 | 昆明理工大学 | Fault diagnosis method for guide shoe of elevator |
| CN112767387B (en) * | 2021-01-29 | 2024-04-30 | 中华人民共和国张家港海关 | Automatic wood image recognition method based on block gradient weighting |
| TWI786555B (en) * | 2021-02-26 | 2022-12-11 | 寶元數控股份有限公司 | Pattern identification and classification method and system |
| CN114397250B (en) * | 2021-12-27 | 2024-02-13 | 中国林业科学研究院木材工业研究所 | Wood identification method based on spectral features and image features |
-
2010
- 2010-10-14 CN CN201010506395A patent/CN101996328B/en not_active Expired - Fee Related
Non-Patent Citations (5)
| Title |
|---|
| 何涛等.基于两方向二维主成分分析木材识别的研究.《林业机械与木工设备》.2009,第37卷(第10期),12-14. * |
| 吴思远等.基于2D Gabor小波和HSV空间的木材缺陷检测.《郑州大学学报(理学版)》.2010,第42卷(第1期),93-97. * |
| 张召等.基于纹理提取和SVM技术的自动木材缺陷识别.《计算机工程与应用》.2009,第45卷(第23期),219-223. * |
| 徐一清.木材纹理识别算法研究进展.《大众科技》.2010,(第8期),145-146. * |
| 汪杭军等.木材识别方法研究综述.《浙江林学院学报》.2009,第26卷(第6期),896-902. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101996328A (en) | 2011-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101996328B (en) | Wood identification method | |
| CN104834922B (en) | Gesture identification method based on hybrid neural networks | |
| Parisi-Baradad et al. | Otolith shape contour analysis using affine transformation invariant wavelet transforms and curvature scale space representation | |
| CN110837768B (en) | Online detection and identification method for rare animal protection | |
| Cope et al. | Plant texture classification using gabor co-occurrences | |
| CN102254188B (en) | Palmprint recognizing method and device | |
| CN104217221A (en) | Method for detecting calligraphy and paintings based on textural features | |
| CN105574063A (en) | Image retrieval method based on visual saliency | |
| Rathod et al. | Leaf disease detection using image processing and neural network | |
| CN105654107A (en) | Visible component classification method based on SVM | |
| CN110348289B (en) | Finger vein identification method based on binary image | |
| CN103177266A (en) | Intelligent stock pest identification system | |
| CN103390170A (en) | Surface feature type texture classification method based on multispectral remote sensing image texture elements | |
| CN102136074B (en) | Man-machine interface (MMI) based wood image texture analyzing and identifying method | |
| Farouk et al. | Microarray spot segmentation algorithm based on integro-differential operator | |
| CN113112498A (en) | Grape leaf scab identification method based on fine-grained countermeasure generation network | |
| CN104268587B (en) | False fingerprint detection method based on finger wave conversion and SVM | |
| Cope et al. | The extraction of venation from leaf images by evolved vein classifiers and ant colony algorithms | |
| Panigrahi et al. | Evaluations of fractal geometry and invariant moments for shape classification of corn germplasm | |
| CN111127407B (en) | Fourier transform-based style migration forged image detection device and method | |
| Jasani et al. | Review of shape and texture feature extraction techniques for fruits | |
| CN108960246A (en) | A kind of binary conversion treatment device and method for image recognition | |
| CN110136112B (en) | Computer-aided detection algorithm based on mammary gland X-ray photography calcification | |
| Bailey et al. | Image analysis of weaverbird nests reveals signature weave textures | |
| CN108846403B (en) | Method for extracting facet gem image features |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121010 Termination date: 20141014 |
|
| EXPY | Termination of patent right or utility model |