CN106528826A - Deep learning-based multi-view appearance patent image retrieval method - Google Patents
Deep learning-based multi-view appearance patent image retrieval method Download PDFInfo
- Publication number
- CN106528826A CN106528826A CN201611027856.3A CN201611027856A CN106528826A CN 106528826 A CN106528826 A CN 106528826A CN 201611027856 A CN201611027856 A CN 201611027856A CN 106528826 A CN106528826 A CN 106528826A
- Authority
- CN
- China
- Prior art keywords
- image
- network
- classification
- training
- appearance patent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Library & Information Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a deep learning-based multi-view appearance patent image retrieval method, which comprises the steps of carrying out appearance patent image preprocessing, carrying out normalization processing on the appearance patent image from the aspects of the dimension and the scale and distinguishing multi-view images according to views; constructing a multi-view depth convolutional neural network, constructing multi-view parallel processing convolutional neural network, carrying out feature fusion according to the spatial position relationship among the views and then adopting a full-connected convolutional network; optimizing and adjusting pre-trained network parameters during network training; carrying out image classification and feature extraction in an image library after training is completed and storing image features into corresponding classes; and feeding back similar images and the similarity according to successive classes and features during image retrieval. The multi-view depth convolutional neural network provided by the invention is merged with the correlation among the views, and the retrieval accuracy is improved.
Description
Technical field
The present invention relates to a kind of field of image search, especially a kind of multi views appearance patent image based on deep learning
Search method.
Technical background
Deeply implement innovation in China to drive under development strategy overall background, our province proposes that accelerating the upgrading of dominant tradition industry changes
Generation.Patented technology is the engine of bootstrap technique industry development, 90%-95% of the patent comprising world's whole scientific and technological information, and skill
Art information discloses early 1～2 year compared with other carriers.Design patent become protection Intellectual Property Right of Enterprises, safeguard itself rights and interests,
Protection innovation and creation important channel.
Being currently based on appearance patent image retrieval mainly has two big class, and the first kind is that, based on character search, this is the most normal
One class method, the major defect for existing are image cannot to be labeled with suitable word, that is to say so-called one
Width picture is to thousand speeches.The result error that the result of this retrieval leads to not retrieve is very big.
Second method is adopted to scheme to search the method for figure, and the method that tradition is used is by such as Gabor filter, SIFT
Extract the feature of image etc. so-called " optimal characteristics extraction algorithm ", such as shape, texture, color etc., further using feature it
Between distance carrying out similarity-rough set.Using each view of design patent image as separate in these methods
Carrying out characteristic processing, caused retrieval rate is low for image.
Design patent image represents the outward appearance of invention object generally using multi views (such as 4 views or 6 views).
The multi views of appearance patent image are organic wholes, therefore, design the inspection of organic appearance patent image using these views
Rope is a problem for needing in the art to solve.
The content of the invention
For deficiency of the prior art, the present invention provides a kind of multi views appearance patent image based on deep learning and examines
The method of rope, builds multi views convolutional neural networks deep learning framework, and the network parameter by the use of pre-training is used as learning network
Initial weight, while image is come subchannel feature extraction according to view, carry out pond according to the spatial relation of view
Fusion, and carry out follow-up feature extraction and classifying.The method substantially increases the accuracy of image searching result, solves outward appearance
Feature between patent image retrieval process multi views lacks the problem for organically blending.
According to design provided by the present invention, a kind of multi views appearance patent image retrieval side based on deep learning
Method, specifically comprises the steps of：
Step 1. appearance patent image pretreatment, by appearance patent image dimension normalization, image dimension normalization, while
Each view of appearance patent image is made a distinction into classification；Appearance patent image data set is divided into into test data set and training
Data set two parts.
Step 2. constructs multi views deep learning network, branches into each class view construction comprising 3 according to seven views, seven tunnel
The network of layer convolution, is merged using pond afterwards, is then 3 layers of full connection, is carried out classification finally by Softmax defeated
Go out, by the use of pre-training network parameter as network initial weight.
Multi views depth convolutional neural networks are instructed by step 3. image characteristics extraction and classification using training sample
Practice, network parameter weight is adjusted, the multi views deep learning network model after training network is obtained.By test set and instruction
Practice the image of level by network model, be calculated the character representation and its classification of image.
Step 4. retrieval result sequencing of similarity is exported, by image to be retrieved after Image semantic classification, Jing Guoshen
Degree learning network, extracts the feature and classification of image, and the row distance that enters between similar characteristics of image compares, according to distance
Numerical value sort from small to large feedback output, and by corresponding image export.
Beneficial effects of the present invention：The present invention retrieves the multidimensional view lacked to patent figure for existing appearance patent image
Organic application, using depth convolutional neural networks construct multi views convolutional neural networks, by corresponding view according to view
Path carry out process of convolution, it is considered to pond fusion treatment is carried out on the basis of the spatial relation of view, view has been excavated
Between internal relation, substantially increase the accuracy of image retrieval.
Description of the drawings
Fig. 1. the schematic flow sheet of the present invention
Fig. 2. flow chart provided in an embodiment of the present invention.
Specific embodiment
In order that the purpose of the present invention, technical scheme are advantage become more apparent, it is below in conjunction with drawings and Examples, right
The present invention is further described.It should be appreciated that specific embodiment described herein is only to explain the present invention, and without
It is of the invention in limiting.
Embodiment one, with reference to shown in Fig. 1, a kind of multi views appearance patent image search method based on deep learning, its
It is characterised by, including：
In a step 101, to appearance patent Image semantic classification, appearance patent image dimension normalization, image dimension are returned
One changes, while each view of appearance patent image is made a distinction classification；Appearance patent image data set is divided into into test number
According to collection and training dataset two parts.
It is above-mentioned, in step 101, graphical rule normalization be by Image Adjusting be identical yardstick.
Preferably, yardstick is 128*128.
Above-mentioned, in step 101, image dimension normalization is that the gray level image of two dimension is changed into three-dimensional similar rgb format
Image.
Preferably, R, G, the value of channel B respective pixel and the gray level image respective pixel value of new image will be increased
It is identical.
In a step 102, multi views deep learning network is constructed, and each class view construction is branched into according to seven views, seven tunnel
Network comprising 3 layers of convolution, is merged using pond afterwards, is then 3 layers of full connection, is carried out finally by Softmax
Classification output, by the use of pre-training network parameter as network initial weight.
Above-mentioned, in step 102, seven views, seven tunnel branch is respectively front view, left view, right view, top view, looks up
Figure, rearview and three-dimensional view branch.
Above-mentioned, in step 102, the network parameter of pre-training uses the network ginseng for arriving trained based on ImageNet
Number.
Preferably, VGG-M models are selected as network parameter.
Above-mentioned, in step 102, the initial parameter of seven road network branches is identical with the network architecture.
It is above-mentioned, in step 102, carry out merging adopting using pond and merged based on the maximum mode of pad.
Preferably, the rule of fusion is merged according to the adjacency of view spaces position.
Preferably, the size of pad is 2*2.
In step 103, image characteristics extraction and classification, are instructed to multi views deep learning network using training sample
Practice, network parameter weight is adjusted, the multi views deep learning network model after training network is obtained.By test set and instruction
Practice the image of level by network model, be calculated the character representation and its classification of image.
Above-mentioned, in step 103, image is characterized in that the multidimensional image feature exported after the ReLU before Softmax
Xi, obtains classification Ck of image after Softmax.
Further, for multidimensional image feature xi is compressed coding.
At step 104, retrieval result sequencing of similarity output, by image to be retrieved after Image semantic classification,
Through deep learning network, the feature and classification of image are extracted, the row distance that enters between similar characteristics of image compares, and presses
Range from numerical value sort from small to large feedback output, and by corresponding image export.
Above-mentioned, in step 104, image to be retrieved is Cn by the classification that deep learning network is obtained, then follow-up phase
Only consider when comparing like property that Cn classification images are previously calculated the multidimensional image feature of storage.Using distance can adopt European
Distance, mahalanobis distance etc..
Embodiment two：With reference to shown in Fig. 2, a kind of multi views appearance patent image search method based on deep learning, its
It is characterised by, including：
It is input in initial pictures, for network parameter adjusting training process, the image of input is several appearance patents every time
Multi views.In retrieval, at least one image to be retrieved can be input into.
In step 201, picture size normalization, by input picture size unification, facilitate follow-up feature analysiss with
Extract.
In step 202., unified is tri- channels of RGB, and the image to being input into is not processed for the image of RGB；To gray-scale maps
Picture, builds width new images, and the R of new images, G, the value of channel B respective pixel are identical with gray level image respective pixel value.
In step 203, view classification, the image that will be input into are separately input to corresponding passage according to its view label
In.
In step 204, CNN1, image is extracted to the feature of image using 3 layers of neural convolutional network.Initial
That the CNN1 network parameters of each paths are identical, according to training after, the network parameter of each paths is likely to occur inconsistent.
In step 205, multi views pondization is processed, and each road view passage is adopted the Pad poles of 2*2 according to its locus
Big value mode carries out pondization fusion.
In step 205, CNN2, extracts characteristics of image using the depth convolution modes of 3 layers of full connection, CNN2 last
The activation primitive of layer adopts ReLU, and the high-level characteristic of output image, this feature is associated with image name and is deposited after which
Storage.
In step 207, Softmax graders, the feature of step 205 is classified, and obtains the classification of image.
In a step 208, the further associated storage of characteristic that will be obtained in the classification of image and step 205.Follow-up
During image retrieval, the classification of image is first determined whether, then, the distance of movement images feature under the category, according to being calculated
Distance, export the picture number specified from small to large.
Finally it should be noted that：Above example only to illustrate technical scheme, rather than a limitation；Although
With reference to the foregoing embodiments the present invention has been described in detail, it will be understood by those within the art that：Which still may be used
To modify to the technical scheme described in foregoing embodiments, or equivalent is carried out to which part technical characteristic；
And these modification or replace, do not make appropriate technical solution essence depart from various embodiments of the present invention technical scheme spirit and
Scope.
Claims (6)
1. a kind of multi views appearance patent image search method based on deep learning, it is characterised in that include：Appearance patent figure
As pretreatment, multi views depth convolutional neural networks, network training and image characteristics extraction and classification are built, retrieval result is similar
Degree sequence output；
Described appearance patent image pretreatment, the appearance patent image to being input into are normalized and standardization processing；
Described construction multi views depth convolutional neural networks, build multichannel depth according to the maximum view classification of appearance patent
Convolutional neural networks, and using pre-training model parameter as initiation parameter；
Described network training and image characteristics extraction and classification, adjust pre-training using using the appearance patent image of labelling
Deep learning network parameter.The feature and classification information of appearance patent image collection are extracted, and is stored；
Described retrieval result sequencing of similarity output, image to be retrieved is compared with the characteristics of image in picture library, defeated
Go out the high image of similarity and its similarity.
2. the multi views appearance patent image search method based on deep learning according to claim 1, it is characterised in that
Described appearance patent image pretreatment is included appearance patent image dimension normalization, image dimension normalization, while will be outer
Each view for seeing patent image makes a distinction classification；Appearance patent image data set is divided into into test data set and training data
Collection two parts.
3. the multi views appearance patent image search method based on deep learning according to claim 1, it is characterised in that
The construction multi views depth convolutional neural networks include branching into each class view construction comprising 3 floor convolution according to seven road views
Network, merged using pond afterwards, be then 3 layers it is complete connect, carry out classification output finally by Softmax, profit
With the network parameter of pre-training as network initial weight.
4. the multi views appearance patent image search method based on deep learning according to claim 1, it is characterised in that
Described network training, image characteristics extraction and classification include carrying out multi views depth convolutional neural networks using training sample
Training, is adjusted to network parameter weight, obtains the multi views deep learning network model after training network.By test set and
The image of training level is calculated the character representation and its classification of image by network model.
5. the multi views appearance patent image search method based on deep learning according to claim 1, it is characterised in that
Retrieval result sequencing of similarity output, including by image to be retrieved after Image semantic classification, through deep learning
Network, extracts the feature and classification of image, and the row distance that enters between similar characteristics of image compares, according to the numerical value of distance
Sequence feedback output from small to large, and corresponding image is exported.
6. the multi views appearance patent image search method based on deep learning according to claim 1, it is characterised in that
Comprise the following steps：
In step 201, picture size normalization, by the picture size unification of input, facilitates follow-up analysis and extraction of features；
In step 202., unified is tri- channels of RGB, and the image to being input into is not processed for the image of RGB；To gray level image, structure
Width new images are built, the R of new images, G, the value of channel B respective pixel are identical with gray level image respective pixel value；
In step 203, view classification, the image that will be input into are separately input in corresponding passage according to its view label；
In step 204, CNN1, image is extracted to the feature of image using 3 layers of neural convolutional network.Being initially,
The CNN1 network parameters of each paths are identical, according to training after, the network parameter of each paths is likely to occur inconsistent；
In step 205, multi views pondization is processed, and each road view passage is adopted the Pad maximum of 2*2 according to its locus
Mode carries out pondization fusion；
In step 205, CNN2, extracts characteristics of image using the depth convolution mode of 3 layers of full connection, last layer of CNN2
Activation primitive adopts ReLU, after which the high-level characteristic of output image, and this feature and image name are associated storage；
In step 207, Softmax graders, the feature of step 205 is classified, and obtains the classification of image；
In a step 208, the further associated storage of characteristic that will be obtained in the classification of image and step 205.In follow-up image
During retrieval, first determine whether the classification of image, then, the distance of movement images feature under the category, according to it is calculated away from
From exporting the picture number specified from small to large.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611027856.3A CN106528826A (en) | 2016-11-18 | 2016-11-18 | Deep learning-based multi-view appearance patent image retrieval method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611027856.3A CN106528826A (en) | 2016-11-18 | 2016-11-18 | Deep learning-based multi-view appearance patent image retrieval method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106528826A true CN106528826A (en) | 2017-03-22 |
Family
ID=58353108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611027856.3A Pending CN106528826A (en) | 2016-11-18 | 2016-11-18 | Deep learning-based multi-view appearance patent image retrieval method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106528826A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991451A (en) * | 2017-04-14 | 2017-07-28 | 武汉神目信息技术有限公司 | A kind of identifying system and method for certificate picture |
| CN107610004A (en) * | 2017-10-25 | 2018-01-19 | 深圳市宝盒科技有限公司 | A kind of commodity are really weighed and acquisition methods, computer |
| CN107609587A (en) * | 2017-09-11 | 2018-01-19 | 浙江工业大学 | A kind of multi-class multi views data creation method that confrontation network is generated based on depth convolution |
| CN108389227A (en) * | 2018-03-01 | 2018-08-10 | 深圳市唯特视科技有限公司 | A kind of dimensional posture method of estimation based on multiple view depth perceptron frame |
| CN108984576A (en) * | 2018-02-07 | 2018-12-11 | 广州集创佳禾知识产权运营有限公司 | A kind of patent document acquisition methods and system based on convolutional neural networks |
| CN109241314A (en) * | 2018-08-27 | 2019-01-18 | 维沃移动通信有限公司 | A kind of selection method and device of similar image |
| CN109445688A (en) * | 2018-09-29 | 2019-03-08 | 上海百功半导体有限公司 | A kind of storage controlling method, storage control, storage equipment and storage system |
| CN109509228A (en) * | 2017-09-15 | 2019-03-22 | 安讯士有限公司 | Method for positioning one or more candidate digital images |
| CN109857895A (en) * | 2019-01-25 | 2019-06-07 | 清华大学 | Stereoscopic vision search method and system based on polycyclic road view convolutional neural networks |
| CN109919208A (en) * | 2019-02-25 | 2019-06-21 | 中电海康集团有限公司 | A kind of appearance images similarity comparison method and system |
| CN110543581A (en) * | 2019-09-09 | 2019-12-06 | 山东省计算中心（国家超级计算济南中心） | Multi-view three-dimensional model retrieval method based on non-local graph convolution network |
| CN110704653A (en) * | 2019-09-09 | 2020-01-17 | 上海慧之建建设顾问有限公司 | Method for searching component by graph in BIM (building information modeling) model and graph-text searching system |
| CN110992217A (en) * | 2019-11-15 | 2020-04-10 | 广东工业大学 | Method and device for expressing and searching multi-view features of design patent |
| CN111310821A (en) * | 2020-02-11 | 2020-06-19 | 佛山科学技术学院 | Multi-view feature fusion method, system, computer device and storage medium |
| CN111598130A (en) * | 2020-04-08 | 2020-08-28 | 天津大学 | Traditional Chinese medicine identification method based on multi-view convolutional neural network |
| CN112836010A (en) * | 2020-10-22 | 2021-05-25 | 长城计算机软件与系统有限公司 | Patent retrieval method, storage medium and device |
| CN114020953A (en) * | 2021-10-27 | 2022-02-08 | 北京中知智慧科技有限公司 | Multi-image retrieval method and device for appearance design product |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101847163A (en) * | 2010-05-28 | 2010-09-29 | 广东工业大学 | Design patent image retrieval method with multi-characteristics fusion |
| CN101866352A (en) * | 2010-05-28 | 2010-10-20 | 广东工业大学 | Design patent retrieval method based on analysis of image content |
| CN105183737A (en) * | 2014-11-17 | 2015-12-23 | 广东工业大学 | Industrial Design instant searching system and searching method thereof |
| CN106062774A (en) * | 2014-11-15 | 2016-10-26 | 北京旷视科技有限公司 | Face detection using machine learning |
| CN106095089A (en) * | 2016-06-06 | 2016-11-09 | 郑黎光 | A kind of method obtaining interesting target information |
-
2016
- 2016-11-18 CN CN201611027856.3A patent/CN106528826A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101847163A (en) * | 2010-05-28 | 2010-09-29 | 广东工业大学 | Design patent image retrieval method with multi-characteristics fusion |
| CN101866352A (en) * | 2010-05-28 | 2010-10-20 | 广东工业大学 | Design patent retrieval method based on analysis of image content |
| CN106062774A (en) * | 2014-11-15 | 2016-10-26 | 北京旷视科技有限公司 | Face detection using machine learning |
| CN105183737A (en) * | 2014-11-17 | 2015-12-23 | 广东工业大学 | Industrial Design instant searching system and searching method thereof |
| CN106095089A (en) * | 2016-06-06 | 2016-11-09 | 郑黎光 | A kind of method obtaining interesting target information |
Non-Patent Citations (1)
| Title |
|---|
| HANG SU 等: "Multi-view Convolutional Neural Networks for 3D Shape Recognition", 《ICCV "15 PROCEEDINGS OF THE 2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION》 * |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106991451A (en) * | 2017-04-14 | 2017-07-28 | 武汉神目信息技术有限公司 | A kind of identifying system and method for certificate picture |
| CN107609587A (en) * | 2017-09-11 | 2018-01-19 | 浙江工业大学 | A kind of multi-class multi views data creation method that confrontation network is generated based on depth convolution |
| CN107609587B (en) * | 2017-09-11 | 2020-08-18 | 浙江工业大学 | Multi-class multi-view data generation method for generating countermeasure network based on deep convolution |
| CN109509228A (en) * | 2017-09-15 | 2019-03-22 | 安讯士有限公司 | Method for positioning one or more candidate digital images |
| JP2019079494A (en) * | 2017-09-15 | 2019-05-23 | アクシス アーベー | Method for locating one or more candidate digital images being likely candidates for depicting object |
| CN107610004A (en) * | 2017-10-25 | 2018-01-19 | 深圳市宝盒科技有限公司 | A kind of commodity are really weighed and acquisition methods, computer |
| CN108984576A (en) * | 2018-02-07 | 2018-12-11 | 广州集创佳禾知识产权运营有限公司 | A kind of patent document acquisition methods and system based on convolutional neural networks |
| CN108389227A (en) * | 2018-03-01 | 2018-08-10 | 深圳市唯特视科技有限公司 | A kind of dimensional posture method of estimation based on multiple view depth perceptron frame |
| CN109241314A (en) * | 2018-08-27 | 2019-01-18 | 维沃移动通信有限公司 | A kind of selection method and device of similar image |
| CN109445688A (en) * | 2018-09-29 | 2019-03-08 | 上海百功半导体有限公司 | A kind of storage controlling method, storage control, storage equipment and storage system |
| CN109857895B (en) * | 2019-01-25 | 2020-10-13 | 清华大学 | Stereo vision retrieval method and system based on multi-loop view convolutional neural network |
| CN109857895A (en) * | 2019-01-25 | 2019-06-07 | 清华大学 | Stereoscopic vision search method and system based on polycyclic road view convolutional neural networks |
| CN109919208A (en) * | 2019-02-25 | 2019-06-21 | 中电海康集团有限公司 | A kind of appearance images similarity comparison method and system |
| CN110704653A (en) * | 2019-09-09 | 2020-01-17 | 上海慧之建建设顾问有限公司 | Method for searching component by graph in BIM (building information modeling) model and graph-text searching system |
| CN110543581A (en) * | 2019-09-09 | 2019-12-06 | 山东省计算中心（国家超级计算济南中心） | Multi-view three-dimensional model retrieval method based on non-local graph convolution network |
| CN110543581B (en) * | 2019-09-09 | 2023-04-04 | 山东省计算中心（国家超级计算济南中心） | Multi-view three-dimensional model retrieval method based on non-local graph convolution network |
| CN110992217A (en) * | 2019-11-15 | 2020-04-10 | 广东工业大学 | Method and device for expressing and searching multi-view features of design patent |
| CN111310821A (en) * | 2020-02-11 | 2020-06-19 | 佛山科学技术学院 | Multi-view feature fusion method, system, computer device and storage medium |
| CN111310821B (en) * | 2020-02-11 | 2023-11-21 | 佛山科学技术学院 | Multi-view feature fusion method, system, computer equipment and storage medium |
| CN111598130A (en) * | 2020-04-08 | 2020-08-28 | 天津大学 | Traditional Chinese medicine identification method based on multi-view convolutional neural network |
| CN112836010A (en) * | 2020-10-22 | 2021-05-25 | 长城计算机软件与系统有限公司 | Patent retrieval method, storage medium and device |
| CN112836010B (en) * | 2020-10-22 | 2024-04-05 | 新长城科技有限公司 | Retrieval method, storage medium and device for patent |
| CN114020953A (en) * | 2021-10-27 | 2022-02-08 | 北京中知智慧科技有限公司 | Multi-image retrieval method and device for appearance design product |
| CN114020953B (en) * | 2021-10-27 | 2022-12-13 | 北京中知智慧科技有限公司 | Multi-image retrieval method and device for appearance design product |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106528826A (en) | Deep learning-based multi-view appearance patent image retrieval method | |
| CN107169956B (en) | Color woven fabric defect detection method based on convolutional neural network | |
| CN109212617B (en) | Automatic identification method and device for electric imaging logging phase | |
| CN106650806B (en) | A kind of cooperating type depth net model methodology for pedestrian detection | |
| CN102054178B (en) | A kind of image of Chinese Painting recognition methods based on local semantic concept | |
| CN102314614B (en) | Image semantics classification method based on class-shared multiple kernel learning (MKL) | |
| CN106529605B (en) | The image-recognizing method of convolutional neural networks model based on theory of immunity | |
| CN109272500B (en) | Fabric classification method based on adaptive convolutional neural network | |
| CN107679250A (en) | A kind of multitask layered image search method based on depth own coding convolutional neural networks | |
| CN103106265B (en) | Similar image sorting technique and system | |
| CN107622272A (en) | A kind of image classification method and device | |
| CN110852316A (en) | Image tampering detection and positioning method adopting convolution network with dense structure | |
| CN111339935B (en) | Optical remote sensing picture classification method based on interpretable CNN image classification model | |
| CN109002755B (en) | Age estimation model construction method and estimation method based on face image | |
| CN114998220B (en) | Tongue image detection and positioning method based on improved Tiny-YOLO v4 natural environment | |
| CN106408030A (en) | SAR image classification method based on middle lamella semantic attribute and convolution neural network | |
| CN113326390B (en) | Image retrieval method based on depth feature consistent Hash algorithm | |
| CN111627080A (en) | Gray level image coloring method based on convolution nerve and condition generation antagonistic network | |
| CN113159067A (en) | Fine-grained image identification method and device based on multi-grained local feature soft association aggregation | |
| CN110929746A (en) | Electronic file title positioning, extracting and classifying method based on deep neural network | |
| CN116665054A (en) | Remote sensing image small target detection method based on improved YOLOv3 | |
| CN110659680B (en) | Image patch matching method based on multi-scale convolution | |
| CN111310820A (en) | Foundation meteorological cloud chart classification method based on cross validation depth CNN feature integration | |
| CN118154550A (en) | Yarn production process detection method based on lightweight target detection network | |
| CN112364193A (en) | Image retrieval-oriented method for fusing multilayer characteristic deep neural network model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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: 20170322 |
|
| RJ01 | Rejection of invention patent application after publication |