CN109583564A - Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks - Google Patents
Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks Download PDFInfo
- Publication number
- CN109583564A CN109583564A CN201811284023.4A CN201811284023A CN109583564A CN 109583564 A CN109583564 A CN 109583564A CN 201811284023 A CN201811284023 A CN 201811284023A CN 109583564 A CN109583564 A CN 109583564A
- Authority
- CN
- China
- Prior art keywords
- neural networks
- convolutional neural
- extremely similar
- similar animal
- animal origin
- 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
- 241001465754 Metazoa Species 0.000 title claims abstract description 44
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 38
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 27
- 239000000835 fiber Substances 0.000 claims abstract description 15
- 238000002360 preparation method Methods 0.000 claims abstract description 7
- 210000004209 hair Anatomy 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 8
- 210000002268 wool Anatomy 0.000 claims description 8
- 241000283707 Capra Species 0.000 claims description 7
- 239000011521 glass Substances 0.000 claims description 7
- 238000002790 cross-validation Methods 0.000 claims description 5
- 238000000399 optical microscopy Methods 0.000 claims description 5
- 238000003756 stirring Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005464 sample preparation method Methods 0.000 claims description 2
- 238000013461 design Methods 0.000 abstract description 2
- 210000000085 cashmere Anatomy 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 238000001000 micrograph Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000000879 optical micrograph Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
The present invention provides a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks, sample sample preparation and Image Acquisition are carried out first, image is pre-processed again, then improved VGG convolutional neural networks are established, by the established network of training set training, and pass through the parameter of verifying set debugging network；The identifier, is used for the automatic identification of extremely similar animal origin by the identifier that the extremely similar animal origin automatic identification based on VGG convolutional neural networks can be obtained.Extremely similar animal origin automatic identifying method provided by the invention based on VGG convolutional neural networks can automatically extract and identify fiber surface morphology feature, thus it is objective, accurately and rapidly identify extremely similar animal origin.In addition, user does not need to provide the microphoto of up to a million extremely similar animal origins by the convolutional neural networks, does not need artificially to design complicated characteristic index yet, that is, can reach the higher classifying quality of precision.
Description
Technical field
The present invention relates to a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks, belongs to number
According to the depth learning technology field in science.
Background technique
Often mix the extremely similar animal origin of fraud of non-white goat wool, such as soil species in the cashmere product of white goat wool
Hair, merino hair, the green suede of Mongolia, domestic green suede, purple suede etc..How it is identified, is always the problem of this field.
Belong to kind by the configuration of surface of optical microphotograph sem observation animal origin and to it to differentiate, is current commercial cashmere
The main method that wool fiber identifies.This method depends on the subjective experience of technical staff, and error rate is high, working efficiency
It is low, and it is difficult to use in automatic identification process.
Therefore, how to automatically extract and identify fiber surface morphology feature, thus it is objective, accurately and rapidly identify white mountain
The extremely similar animal origins such as cashmere, local hair, merino hair, the green suede of Mongolia, domestic green suede, purple suede, are those skilled in the art
It is dedicated to the problem solved.
Summary of the invention
The technical problem to be solved by the present invention is how to automatically extract and identify fiber surface morphology feature, thus objective,
Accurately and rapidly identify extremely similar animal origin.
In order to solve the above-mentioned technical problem, the technical solution of the present invention is to provide a kind of based on VGG convolutional neural networks
Extremely similar animal origin automatic identifying method, which comprises the following steps:
Step 1: sample sample preparation；
All kinds of extremely similar animal origin individually sample preparations；All kinds of extremely similar animal origins to be identified are cut, load is placed in
On slide, atoleine is added, stirring is uniformly distributed it, covers glass slide, completes the sample preparation of all kinds of extremely similar animal origins；
Step 2: Image Acquisition；
Sample made of step 1 is placed under optical microscopy, after manual focus, shoots single fiber micrograph
Picture；Each type of fibers shoots 8000~10000 micro-images, and the micro-image of each fiber-like is divided into three groups, point
Set, cross validation set and test set Yong Zuo not trained；
Step 3: image preprocessing；
Original micro-image is reduced, then random cropping, then the pixel in each image after cutting is subtracted into training set
In all images 3 channel average values；
Step 4: establishing improved VGG convolutional neural networks；
Using the VGG convolutional neural networks of 13 layer architectures, concrete configuration is as follows:
Wherein, conv3-32 represents 32 3 × 3 convolution kernels, and conv3-64 represents 64 3 × 3 convolution kernels, conv3-
128 represent 128 3 × 3 convolution kernels, and conv3-256 represents 256 3 × 3 convolution kernels.FC-512 represents 512 neurons
The full articulamentum constituted, FC-6 represent the full articulamentum of 6 neurons composition.Softmax represent final activation primitive as
Softmax type activation primitive.
Step 5: the improved VGG convolutional neural networks established are trained；
The improved VGG convolutional neural networks established by training set training, and pass through verifying set debugging net
The parameter of network；
Step 6: the generation of identifier；
By the training of step 5, the extremely similar animal origin automatic identification based on VGG convolutional neural networks can be obtained
The identifier is used for the automatic identification of extremely similar animal origin by identifier.
Preferably, in the step 1, all kinds of extremely similar animal origins to be identified are cut using Kazakhstan food slicer
0.5mm is placed on glass slide.
Preferably, in the step 1, all kinds of extremely similar animal origins to be identified include white goat wool, local hair, Mei Li
Slave's hair, the green suede of Mongolia, domestic green suede, six class of purple suede.
Preferably, it in the step 1, is carried out using analysis needle when stirring.
Preferably, in the step 2, sample made of step 1 is placed under optical microscopy, amplifies 500 times.
Preferably, in the step 2, the resolution ratio of single fiber micro-image is at least 640 × 480.
Preferably, it in the step 2, by the micro-image of each fiber-like according to the ratio of 8:1:1, selects therein
80% gathers as training, 10% is used as cross validation set, 10% as test set.
Preferably, in the step 3, original micro-image is contracted to 256 × 256 sizes, then by the figure after diminution
As random cropping to size is 224 × 224.
Preferably, in the step 5, under TensorFlow frame, specific training parameter is provided that
Training batch size is set as 32 every time；
The optimization method of gradient decline uses Adam algorithm, and the value of constant ρ is 10-8, moments estimation exponential decay rate ρ1And ρ2
It is respectively set to 0.9 and 0.999；
Learning rate is provided when optimizing using exponential damping method for each round training, initial learning rate is set as 0.001, declines
Subtracting coefficient is 0.93, and the rate of decay is the decaying that every 2 wheel carries out a learning rate.
Preferably, in the step 6, obtained identifier have white goat wool, local hair, merino hair, the green suede of Mongolia,
Domestic blueness suede, six Classification and Identification ability of purple suede.
The present invention provides a kind of best VGG framework convolutional neural networks for the extremely similar animal origin of automatic identification,
The convolutional neural networks can automatically extract and identify fiber surface morphology feature, thus it is objective, accurately and rapidly identify pole phase
Apparent movement fibres.In addition, user does not need to provide the aobvious of up to a million extremely similar animal origins by the convolutional neural networks
Micro- photo does not need artificially to design complicated characteristic index yet, that is, can reach the higher classifying quality of precision.
Detailed description of the invention
Fig. 1 is existing VGG convolutional neural networks structural schematic diagram；
Fig. 2 is the improved VGG convolutional neural networks structural schematic diagram of the present invention；
Fig. 3 is discrimination of the different number of plies VGG networks in extremely similar animal origin is classified automatically；
Discrimination when Fig. 4 is different convolutional network configurations under 13 layers of VGG framework.
Specific embodiment
Present invention will be further explained below with reference to specific examples.
A kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks is present embodiments provided,
Specific implementation step is as follows:
Step 1, sample sample preparation.
Cashmere and Woolens fiber is cut into 0.5mm or so using Kazakhstan food slicer, is placed on glass slide, appropriate amount of fluid stone is added
Wax is uniformly distributed it, covers glass slide, completed sample preparation with analysis needle stirring.
Step 2, Image Acquisition.
Sample is placed under optical microscopy, 500 times of amplification, after manual focus, shoots single fiber micro-image,
Resolution ratio is at least 640 × 480.To reach preferable accuracy of identification, each type of fibers needs shooting at least 10000 micro-
Image.Then according to the ratio of 8:1:1, therein 80% is selected to gather as trained, 10% conduct cross validation set, 10%
Gather as test.
Step 3, image preprocessing.
Original image is contracted to 256 × 256 sizes, is then 224 × 224 by image random cropping to size, then will
Pixel in each image subtracts 3 channel average values of all images in training set.
Step 4 establishes improved VGG convolutional neural networks.
In conjunction with Fig. 1, neural network under conventional VGG framework, the setting side of convolutional layer, pond layer and full articulamentum
Formula is as follows:
Convolutional layer: the convolution kernel for the use of size being 3 × 3, the step-length (Stride) of convolution are 1, and filling (Padding) is also set
It is set to 1.
Pond layer: pond layer window size is 2 × 2, and step value 2 uses the side of maximum pond (Max-pooling)
Method.
Full articulamentum (FC): being 3 full articulamentums after a series of superposition of convolutional layers and pond layer.
1000 classifications in conventional VGG network training ImageNet data set, about 1,200,000 width images, and can obtain
Good recognition effect.And during a variety of extremely similar animal origins automatic Classification and Identification, it is unable to reach coordinates data rule
Mould, used data set are 6 classifications totally 6 ten thousand width fibre image, are less than ImageNet data set in data volume and classification
Under the premise of, the number of plies is excessive or convolution nuclear volume is excessive, and will cause the convolutional neural networks under VGG framework, over-fitting occur existing
As.
The present invention improves conventional VGG network.It is fine in extremely similar animal by studying different number of plies VGG networks
Discrimination in the automatic classification of dimension, such as Fig. 3, discovery use the VGG network test collection discrimination highest of 13 layer architectures.Therefore, originally
Invention uses the VGG network of 13 layer architectures, as shown in Figure 2.
On the basis of 13 layers of VGG framework, convolution nuclear volume of the present invention is reasonably selected.For the configuration 1 in table 1
~configuration 5 is tested respectively, finds the discrimination highest in configuration 4, using the network of configuration 4, has obtained test set identification
The Fiber-Net framework that rate is 92.74%.
Different convolutional network configurations under the 13 layers of VGG framework of the invention of table 1.
Step 5 is trained the improved VGG convolutional neural networks established.
Gather training net network model by training, debugging model hyper parameter is gathered by verifying, in TensorFlow frame
Under, specific parameter setting is as follows:
Training batch size (batch_size) is set as 32 every time；
The optimization method of gradient decline uses Adam algorithm, and the value of constant ρ is 10-8, moments estimation exponential decay rate ρ1And ρ2
It is respectively set to 0.9 and 0.999.
Learning rate is provided when optimizing using exponential damping method for each round training, initial learning rate is set as 0.001, declines
Subtracting coefficient is 0.93, and the rate of decay is the decaying that every 2 wheel (epoch) carries out a learning rate.
The generation of step 6, identifier.
By the training of step 5, the discrimination that network can achieve training set is 95.70%, and verifying collection discrimination is
93.55%, test set discrimination is 92.74%.That is the Generalization Capability of network is good.Save weight matrix at this time and amount of bias
Matrix.It produces with six Classification and Identifications such as white goat wool, local hair, merino hair, the green suede of Mongolia, domestic green suede, purple suedes
The identifier of ability.
The above, only presently preferred embodiments of the present invention, not to the present invention in any form with substantial limitation,
It should be pointed out that under the premise of not departing from the method for the present invention, can also be made for those skilled in the art
Several improvement and supplement, these are improved and supplement also should be regarded as protection scope of the present invention.All those skilled in the art,
Without departing from the spirit and scope of the present invention, when made using disclosed above technology contents it is a little more
Dynamic, modification and the equivalent variations developed, are equivalent embodiment of the invention；Meanwhile all substantial technologicals pair according to the present invention
The variation, modification and evolution of any equivalent variations made by above-described embodiment, still fall within the range of technical solution of the present invention
It is interior.
Claims (10)
1. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks, which is characterized in that including following
Step:
Step 1: sample sample preparation；
All kinds of extremely similar animal origin individually sample preparations；All kinds of extremely similar animal origins to be identified are cut, glass slide is placed in
On, atoleine is added, stirring is uniformly distributed it, covers glass slide, completes the sample preparation of all kinds of extremely similar animal origins；
Step 2: Image Acquisition；
Sample made of step 1 is placed under optical microscopy, after manual focus, shoots single fiber micro-image；Often
A type of fibers shoots 8000~10000 micro-images, and the micro-image of each fiber-like is divided into three groups, is used separately as
Training set, cross validation set and test set；
Step 3: image preprocessing；
Original micro-image is reduced, then random cropping, then the pixel in each image after cutting is subtracted into institute in training set
There are 3 channel average values of image；
Step 4: establishing improved VGG convolutional neural networks；
Using the VGG convolutional neural networks of 13 layer architectures, concrete configuration is as follows:
Wherein, conv3-32 represents 32 3 × 3 convolution kernels, and conv3-64 represents 64 3 × 3 convolution kernels, conv3-128
128 3 × 3 convolution kernels are represented, conv3-256 represents 256 3 × 3 convolution kernels.FC-512 represents 512 neuron structures
At full articulamentum, FC-6 represent 6 neurons composition full articulamentum.Softmax represent final activation primitive as
Softmax type activation primitive；
Step 5: the improved VGG convolutional neural networks established are trained；
The improved VGG convolutional neural networks established by training set training, and pass through verifying set debugging network
Parameter；
Step 6: the generation of identifier；
By the training of step 5, the identification of the extremely similar animal origin automatic identification based on VGG convolutional neural networks can be obtained
The identifier is used for the automatic identification of extremely similar animal origin by device.
2. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
It is characterized in that: in the step 1, all kinds of extremely similar animal origins to be identified being cut into 0.5mm using Kazakhstan food slicer and are placed in
On glass slide.
3. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
Be characterized in that: in the step 1, all kinds of extremely similar animal origins to be identified include white goat wool, local hair, merino hair,
Mongolian blueness suede, domestic green suede, six class of purple suede.
4. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
It is characterized in that: in the step 1, being carried out using analysis needle when stirring.
5. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
It is characterized in that: in the step 2, sample made of step 1 being placed under optical microscopy, amplify 500 times.
6. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
Be characterized in that: in the step 2, the resolution ratio of single fiber micro-image is at least 640 × 480.
7. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
It is characterized in that: in the step 2, by the micro-image of each fiber-like according to the ratio of 8:1:1, selecting therein 80% to make
For training set, 10% as cross validation set, 10% as test set.
8. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
It is characterized in that: in the step 3, original micro-image being contracted to 256 × 256 sizes, it is then that the image after diminution is random
Being cut to size is 224 × 224.
9. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as described in claim 1,
Be characterized in that: in the step 5, under TensorFlow frame, specific training parameter is provided that
Training batch size is set as 32 every time；
The optimization method of gradient decline uses Adam algorithm, and the value of constant ρ is 10-8, moments estimation exponential decay rate ρ1And ρ2It sets respectively
It is set to 0.9 and 0.999；
Using exponential damping method for each round training optimize when learning rate is provided, set initial learning rate be 0.001, decaying be
Number is 0.93, and the rate of decay is the decaying that every 2 wheel carries out a learning rate.
10. a kind of extremely similar animal origin automatic identifying method based on VGG convolutional neural networks as claimed in claim 3,
It is characterized by: in the step 6, obtained identifier has white goat wool, local hair, merino hair, the green suede of Mongolia, domestic
Green suede, six Classification and Identification ability of purple suede.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811284023.4A CN109583564A (en) | 2018-10-31 | 2018-10-31 | Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811284023.4A CN109583564A (en) | 2018-10-31 | 2018-10-31 | Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109583564A true CN109583564A (en) | 2019-04-05 |
Family
ID=65921353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811284023.4A Pending CN109583564A (en) | 2018-10-31 | 2018-10-31 | Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109583564A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378303A (en) * | 2019-07-25 | 2019-10-25 | 杭州睿琪软件有限公司 | Method and system for Object identifying |
| CN110490086A (en) * | 2019-07-25 | 2019-11-22 | 杭州睿琪软件有限公司 | A kind of method and system for Object identifying result secondary-confirmation |
| CN111414993A (en) * | 2020-03-03 | 2020-07-14 | 三星（中国）半导体有限公司 | Cutting and convolution calculating method and device of convolution neural network |
| CN112927180A (en) * | 2020-10-15 | 2021-06-08 | 内蒙古鄂尔多斯资源股份有限公司 | Cashmere and wool optical microscope image identification method based on generation countermeasure network |
| CN113077420A (en) * | 2021-03-19 | 2021-07-06 | 江南大学 | Fish fiber evaluation method and system based on convolutional neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103149210A (en) * | 2013-02-25 | 2013-06-12 | 东华大学 | System and method for detecting fabric cashmere content based on scale graphic features |
| CN107463965A (en) * | 2017-08-16 | 2017-12-12 | 湖州易有科技有限公司 | Fabric attribute picture collection and recognition methods and identifying system based on deep learning |
| CN107909107A (en) * | 2017-11-14 | 2018-04-13 | 深圳码隆科技有限公司 | Fiber check and measure method, apparatus and electronic equipment |
| CN108090498A (en) * | 2017-12-28 | 2018-05-29 | 广东工业大学 | A kind of fiber recognition method and device based on deep learning |
| CN108334938A (en) * | 2018-02-09 | 2018-07-27 | 广东省公共卫生研究院 | A kind of mosquito matchmaker's automatic monitoring system based on image recognition |
-
2018
- 2018-10-31 CN CN201811284023.4A patent/CN109583564A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103149210A (en) * | 2013-02-25 | 2013-06-12 | 东华大学 | System and method for detecting fabric cashmere content based on scale graphic features |
| CN107463965A (en) * | 2017-08-16 | 2017-12-12 | 湖州易有科技有限公司 | Fabric attribute picture collection and recognition methods and identifying system based on deep learning |
| CN107909107A (en) * | 2017-11-14 | 2018-04-13 | 深圳码隆科技有限公司 | Fiber check and measure method, apparatus and electronic equipment |
| CN108090498A (en) * | 2017-12-28 | 2018-05-29 | 广东工业大学 | A kind of fiber recognition method and device based on deep learning |
| CN108334938A (en) * | 2018-02-09 | 2018-07-27 | 广东省公共卫生研究院 | A kind of mosquito matchmaker's automatic monitoring system based on image recognition |
Non-Patent Citations (4)
| Title |
|---|
| 柴新玉 等: "基于SURF特征的羊绒羊毛鉴别算法", 《上海纺织科技》 * |
| 王飞 等: "应用卷积网络及深度学习理论的羊绒与羊毛鉴别", 《纺织学报》 * |
| 赵寿刚 等, 黄河水利出版社 * |
| 路凯 等: "基于视觉词袋模型的羊绒与羊毛快速鉴别方法", 《纺织学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378303A (en) * | 2019-07-25 | 2019-10-25 | 杭州睿琪软件有限公司 | Method and system for Object identifying |
| CN110490086A (en) * | 2019-07-25 | 2019-11-22 | 杭州睿琪软件有限公司 | A kind of method and system for Object identifying result secondary-confirmation |
| CN111414993A (en) * | 2020-03-03 | 2020-07-14 | 三星（中国）半导体有限公司 | Cutting and convolution calculating method and device of convolution neural network |
| CN111414993B (en) * | 2020-03-03 | 2024-03-01 | 三星（中国）半导体有限公司 | Convolutional neural network clipping and convolutional calculation method and device |
| CN112927180A (en) * | 2020-10-15 | 2021-06-08 | 内蒙古鄂尔多斯资源股份有限公司 | Cashmere and wool optical microscope image identification method based on generation countermeasure network |
| CN113077420A (en) * | 2021-03-19 | 2021-07-06 | 江南大学 | Fish fiber evaluation method and system based on convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109583564A (en) | Extremely similar animal origin automatic identifying method based on VGG convolutional neural networks | |
| CN107016405B (en) | A kind of pest image classification method based on classification prediction convolutional neural networks | |
| CN109508650A (en) | A kind of wood recognition method based on transfer learning | |
| Revathi et al. | Advance computing enrichment evaluation of cotton leaf spot disease detection using Image Edge detection | |
| CN108717554A (en) | A kind of thyroid tumors histopathologic slide image classification method and its device | |
| CN107067043A (en) | A kind of diseases and pests of agronomic crop detection method | |
| CN107464035A (en) | Chinese medicine performance rating method and system | |
| CN111652326A (en) | Improved fruit maturity identification method and identification system based on MobileNet v2 network | |
| CN108363969B (en) | Newborn pain assessment method based on mobile terminal | |
| CN109583507A (en) | A kind of pig body recognition methods based on depth convolutional neural networks | |
| CN105095907B (en) | A kind of heterogeneous fibre identification method of cotton based on RBF neural | |
| CN106997475A (en) | A kind of insect image-recognizing method based on parallel-convolution neutral net | |
| CN109034269A (en) | A kind of bollworm female male imago method of discrimination based on computer vision technique | |
| CN110689039A (en) | Trunk texture identification method based on four-channel convolutional neural network | |
| CN111462058A (en) | Method for quickly detecting effective ears of rice | |
| CN109087305A (en) | A kind of crack image partition method based on depth convolutional neural networks | |
| CN105678818A (en) | Method for extracting estuary intertidal zone classification area by using object-oriented classification technology | |
| CN109447097B (en) | Fabric principal component detection method based on convolutional neural network | |
| TWI782557B (en) | Cell counting and culture interpretation method, system and computer readable medium thereof | |
| CN110852398B (en) | Aphis gossypii glover recognition method based on convolutional neural network | |
| CN111240004A (en) | System and method for automatically identifying two insects by microscope | |
| CN110533102A (en) | Single class classification method and classifier based on fuzzy reasoning | |
| CN110334773A (en) | Model based on machine learning enters the screening technique of modular character | |
| CN107526440A (en) | The intelligent electric appliance control method and system of gesture identification based on decision tree classification | |
| CN110163855A (en) | A kind of color image quality evaluation method based on multichannel depth convolutional neural networks |
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 |
Application publication date: 20190405 |
|
| RJ01 | Rejection of invention patent application after publication |