CN110210567A - A kind of image of clothing classification and search method and system based on convolutional neural networks - Google Patents
A kind of image of clothing classification and search method and system based on convolutional neural networks Download PDFInfo
- Publication number
- CN110210567A CN110210567A CN201910492484.9A CN201910492484A CN110210567A CN 110210567 A CN110210567 A CN 110210567A CN 201910492484 A CN201910492484 A CN 201910492484A CN 110210567 A CN110210567 A CN 110210567A
- Authority
- CN
- China
- Prior art keywords
- neural networks
- convolutional neural
- image
- networks model
- feature vector
- 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
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a kind of image of clothing classification and search method based on convolutional neural networks, pass through the feature vector for the deep layer convolutional neural networks model extraction image to be retrieved for constructing and training, and the feature vector in the feature vector library of the feature vector of image to be retrieved and foundation is subjected to cosine similarity comparison, be capable of efficiently and accurately retrieves several images identical or most like with image to be retrieved, since convolutional neural networks are to geometric transformation, deformation, illumination has a degree of invariance, therefore this method can greatly improve the accuracy rate of image classification and retrieval, applied in image of clothing retrieval, convenient and efficient and robustness with higher.
Description
Technical field
The present embodiments relate to image identification technical fields, and in particular to a kind of clothes figure based on convolutional neural networks
As classification and search method and system.
Background technique
With the development of clothes e-commerce, the classification of image of clothing and retrieval technique are also in progressive updating to adapt to client
Continually changing demand.Traditional image search method is to retrieve image of clothing by keyword or text, its essence is with
Text searches figure, and with the growth of image of clothing quantity, the disadvantages of this method is more and more significant.Firstly, keyword can only describe to be easy to
Semantic feature extract, abstract, can not comprehensively reflect the visual signature of image of clothing, finer, difficult
With the feature of description；Secondly as amount of images is huge, need that a large amount of manpower and material resources is spent manually to be marked, Er Qieren
Work mark is easy to produce subjective bias；Finally, being difficult to retrieve if the search key of user's input describes not accurate enough
Desired commodity.
And with the development of image processing techniques, occur to scheme to search diagram technology, i.e., one image of input can return more
Open the same or similar image list.Current image of clothing classification and the obtained result of search method based on to scheme to search figure is simultaneously
Be not it is very ideal, especially when there is complicated background or not high picture quality in image, search result is tended not to
It is satisfactory.The existing image search method to scheme to search figure is mostly based on retrieving to image content features, as image is complete
The features such as the color of office, shape, Texture eigenvalue or the point of interest of part, angle point, have the following deficiencies: (1) for clothes
Color characteristic, shape feature or textural characteristics use which kind of lower-level vision feature, how selection sort device is all rule of thumb
Selection, it is difficult to ensure that optimum detection effect；(2) identification calculating ratio is relatively time-consuming, inefficient；(3) it is identified and is taken with specific method
Dress, can only generally apply, robustness is poor, and practical value is low under special scenes；(4) this method is existing for actual implementation
Can be comparatively laborious in, cause application cost to rise.
Summary of the invention
For this purpose, the embodiment of the present invention provides a kind of image of clothing classification based on convolutional neural networks and search method and is
System, to solve the problems, such as that existing image of clothing search method accuracy rate is low, detection efficiency is not high, application is poor.
To achieve the goals above, the embodiment of the present invention provides the following technical solutions:
According to a first aspect of the embodiments of the present invention, propose it is a kind of based on convolutional neural networks image of clothing classification with
Search method, which comprises
Garment image data collection with classification information is divided into training set and verifying collection；
Construct convolutional neural networks model；
The convolutional neural networks model is trained using the training set；
The convolutional neural networks model after training is verified using verifying collection；
The feature vector of image is concentrated using garment image data described in the convolutional neural networks model extraction, and is established
Feature vector library；
Use the feature vector of the convolutional neural networks model extraction image to be retrieved；
Feature vector in the feature vector of the image to be retrieved and described eigenvector library is used into cosine similarity
It is compared one by one, exports the garment image data and concentrate and multiple most like preceding images of the image to be retrieved.
Further, the convolutional neural networks model include sequentially connected first convolutional layer, the first ReLU active coating,
Second convolutional layer, the 2nd ReLU active coating, third convolutional layer, the 3rd ReLU active coating, Volume Four lamination, the 4th ReLU activation
Layer, the 5th convolutional layer, the 5th ReLU active coating, average pond layer and full articulamentum.
Further, the input image size of the convolutional neural networks model is 256 × 256, first convolutional layer
The convolution for being 3 × 3 including a packet size, the output characteristic pattern of first convolutional layer is having a size of 112 × 112, the volume Two
Lamination, third convolutional layer, Volume Four product, layer and the 5th convolutional layer respectively include the convolution that two packet sizes are 3 × 3, and described second
The output characteristic pattern of convolutional layer is having a size of 56 × 56, and the output characteristic pattern of the third convolutional layer is having a size of 28 × 28, described
For the output characteristic pattern of two convolutional layers having a size of 14 × 14, the output characteristic pattern of second convolutional layer is described flat having a size of 7 × 7
Equal pond layer includes the pond that a packet size is 7 × 7, and the output characteristic pattern of the average pond layer is described to connect entirely having a size of 1 × 1
Layer is connect using Softmax function as its classification function.
Further, described that the convolutional neural networks model after training is verified using verifying collection, it wraps
It includes:
If there is error result when verifying, according to classification belonging to the verifying collection image for error result occur, described
The amount of images for increasing respective classes in training set carries out continuing to train to the convolutional neural networks model.
It is further, described that the convolutional neural networks model is trained using the training set, comprising:
The convolutional neural networks model is trained using Softmax loss function.
It is further, described that the convolutional neural networks model is trained using the training set, comprising:
The convolutional neural networks model is trained using stochastic gradient descent algorithm SGD.
Further, before the use training set is trained the convolutional neural networks model further include:
The training set and verifying are integrated into image preprocessing as fixed dimension 256 × 256.
According to a second aspect of the embodiments of the present invention, a kind of image classification and retrieval based on convolutional neural networks is proposed
System, the system comprises:
Data set processing module, for the garment image data collection for having classification information to be divided into training set and verifying
Collection；
Model construction module, for constructing convolutional neural networks model；
Model training module, for being trained using the training set to the convolutional neural networks model；
Model authentication module, for being tested using verifying collection the convolutional neural networks model after training
Card；
Feature vector library constructs module, for using garment image data collection described in the convolutional neural networks model extraction
The feature vector of middle image, and establish feature vector library；
Test module, for using the feature vector of the convolutional neural networks model extraction image to be retrieved；
Feature vector in the feature vector of the image to be retrieved and described eigenvector library is used into cosine similarity
It is compared one by one, exports the garment image data and concentrate and multiple most like preceding images of the image to be retrieved.
The embodiment of the present invention has the advantages that
The embodiment of the present invention proposes a kind of image of clothing classification and search method and system based on convolutional neural networks,
By the feature vector for the deep layer convolutional neural networks model extraction image to be retrieved for constructing and training, and by image to be retrieved
Feature vector in the feature vector library of feature vector and foundation carries out cosine similarity comparison, is capable of retrieving for efficiently and accurately
Several images identical or most like with image to be retrieved, since convolutional neural networks have one to geometric transformation, deformation, illumination
Determine the invariance of degree, therefore this method can greatly improve the accuracy rate of image classification and retrieval, is applied to image of clothing
In retrieval, convenient and efficient and robustness with higher.
Detailed description of the invention
It, below will be to embodiment party in order to illustrate more clearly of embodiments of the present invention or technical solution in the prior art
Formula or attached drawing needed to be used in the description of the prior art are briefly described.It should be evident that the accompanying drawings in the following description is only
It is merely exemplary, it for those of ordinary skill in the art, without creative efforts, can also basis
The attached drawing of offer, which is extended, obtains other implementation attached drawings.
Fig. 1 is a kind of image of clothing classification and search method based on convolutional neural networks that the embodiment of the present invention 1 provides
Flow diagram.
Specific embodiment
Embodiments of the present invention are illustrated by particular specific embodiment below, those skilled in the art can be by this explanation
Content disclosed by book is understood other advantages and efficacy of the present invention easily, it is clear that described embodiment is the present invention one
Section Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
Embodiment 1
As shown in Figure 1, the present embodiment proposes a kind of image of clothing classification and search method based on convolutional neural networks,
It can be used for the classification and retrieval of image of clothing, this method comprises:
S100, the garment image data collection with classification information is divided into training set and verifying collection.
The ration of division that training set and verifying integrate in the present embodiment is incited somebody to action as 4:1, and under the premise of guaranteeing that image is distortionless
Training set and verifying integrate image preprocessing as fixed dimension 256 × 256.
S200, building convolutional neural networks model.
Convolutional neural networks model Primary Reference resnet network structure thought in the present embodiment, major design 3 × 3
Convolution kernel.Further, convolutional neural networks model includes sequentially connected first convolutional layer, the first ReLU active coating,
Two convolutional layers, the 2nd ReLU active coating, third convolutional layer, the 3rd ReLU active coating, Volume Four lamination, the 4th ReLU active coating,
5th convolutional layer, the 5th ReLU active coating, average pond layer and full articulamentum.Line rectification function (Rectified Linear
Unit, ReLU) it is a kind of linearity rectification function, when training data is bigger, ReLU is as activation primitive than traditional
The function of Sigmoid has better adaptability.
The input image size of convolutional neural networks model is 256 × 256, and the first convolutional layer includes that a packet size is 3 × 3
Convolution, the output characteristic pattern of the first convolutional layer is having a size of 112 × 112, second convolutional layers, third convolutional layer, Volume Four product, layer
Respectively include the convolution that two packet sizes are 3 × 3 with the 5th convolutional layer, the output characteristic pattern of the second convolutional layer having a size of 56 × 56,
Output characteristic pattern of the output characteristic pattern of third convolutional layer having a size of 28 × 28, second convolutional layers is having a size of 14 × 14, volume Two
For the output characteristic pattern of lamination having a size of 7 × 7, average pond layer includes the pond that a packet size is 7 × 7, the output of average pond layer
Characteristic pattern is having a size of 1 × 1, and full articulamentum is using Softmax function as its classification function.
S300, convolutional neural networks model is trained using training set.
In the training process, convolutional neural networks model is trained using Softmax loss function, and using random
Gradient descent algorithm SGD is as optimization algorithm.
S400, the convolutional neural networks model after training is verified using verifying collection.
When the convergence of model loss function, the convolutional neural networks model that training obtains is verified on verifying collection,
Further, there is error result when if verifying, the picture and input picture such as exported is simultaneously dissimilar, then according to appearance mistake knot
Fruit verifying collection image belonging to classification, in training set increase respective classes amount of images to convolutional neural networks model into
Row continues to train, to enhance the classifying quality to the type.It repeats the above process, until model loss function is restrained or verified
Testing result on collection is stablized, and network model parameter at this time is the trained deep layer volume with image of clothing classification feature
The parameter of product neural network model.
S500, the feature vector that image is concentrated using convolutional neural networks model extraction garment image data, and establish spy
Levy vector library.
S600, the feature vector of convolutional neural networks model extraction image to be retrieved is used.
In the present embodiment, the input image size of convolutional neural networks model is 256 × 256 when due to training, is surveyed
Input image size when trying unknown picture also should be 256 × 256, roll up fixed-size picture as the present embodiment mid-deep strata
The input of product neural network extracts feature vector using trained network model.
S700, the feature vector of image to be retrieved and the feature vector in feature vector library are carried out using cosine similarity
It compares one by one, output garment image data is concentrated and multiple most like preceding images of image to be retrieved.
Cosine similarity refers to the included angle cosine value of two feature vectors of calculating to assess the similarity of the two, the model of cosine value
It is trapped among between [- 1,1], value more levels off to 1, shows vector angle closer to 0 degree, and it is more similar to represent two feature vectors, more becomes
It is bordering on -1, it is more dissimilar to represent two feature vectors.
By the cosine value being calculated according to being ranked up from big to small, the biggish preceding several feature vectors of cosine value are obtained
Feature vector similar with the feature vector of image to be retrieved in library, and concentrate and obtain and image to be retrieved from garment image data
Several identical or most like images in the present embodiment, export five images identical or most like with image to be retrieved.
The present embodiment proposes a kind of image of clothing classification and search method based on convolutional neural networks, by constructing simultaneously
The feature vector of trained deep layer convolutional neural networks model extraction image to be retrieved, and by the feature vector of image to be retrieved with
Feature vector in the feature vector library of foundation carries out cosine similarity comparison, is capable of retrieving and figure to be retrieved for efficiently and accurately
As several identical or most like images, due to convolutional neural networks to geometric transformation, deformation, illumination have it is a degree of not
Denaturation, therefore this method can greatly improve the accuracy rate of image classification and retrieval, be applied in image of clothing retrieval, it is convenient
Robustness efficient and with higher.
Embodiment 2
It is corresponding with above-described embodiment 1, a kind of image classification and searching system based on convolutional neural networks is proposed,
The system includes:
Data set processing module, for the garment image data collection for having classification information to be divided into training set and verifying
Collection；
Model construction module, for constructing convolutional neural networks model；
Model training module, for being trained using training set to convolutional neural networks model；
Model authentication module, for being verified using verifying collection to the convolutional neural networks model after training；
Feature vector library constructs module, for using convolutional neural networks model extraction garment image data to concentrate image
Feature vector, and establish feature vector library；
Test module, for using the feature vector of convolutional neural networks model extraction image to be retrieved；
Feature vector in the feature vector of image to be retrieved and feature vector library is carried out one by one using cosine similarity
It compares, output garment image data is concentrated and multiple most like preceding images of image to be retrieved.
In a kind of image classification and searching system based on convolutional neural networks provided in this embodiment performed by each component
Function be discussed in detail in above-described embodiment 1, therefore do not do excessively repeat here.
Although above having used general explanation and specific embodiment, the present invention is described in detail, at this
On the basis of invention, it can be made some modifications or improvements, this will be apparent to those skilled in the art.Therefore,
These modifications or improvements without departing from theon the basis of the spirit of the present invention are fallen within the scope of the claimed invention.
Claims (8)
1. a kind of image of clothing classification and search method based on convolutional neural networks, which is characterized in that the described method includes:
Garment image data collection with classification information is divided into training set and verifying collection；
Construct convolutional neural networks model；
The convolutional neural networks model is trained using the training set；
The convolutional neural networks model after training is verified using verifying collection；
The feature vector of image is concentrated using garment image data described in the convolutional neural networks model extraction, and establishes feature
Vector library；
Use the feature vector of the convolutional neural networks model extraction image to be retrieved；
The feature vector of the image to be retrieved and the feature vector in described eigenvector library are carried out using cosine similarity
It compares one by one, exports the garment image data and concentrate and multiple most like preceding images of the image to be retrieved.
2. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 1, feature
Be, the convolutional neural networks model include sequentially connected first convolutional layer, the first ReLU active coating, the second convolutional layer,
2nd ReLU active coating, third convolutional layer, the 3rd ReLU active coating, Volume Four lamination, the 4th ReLU active coating, the 5th convolution
Layer, the 5th ReLU active coating, average pond layer and full articulamentum.
3. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 2, feature
It is, the input image size of the convolutional neural networks model is 256 × 256, and first convolutional layer includes a packet size
For 3 × 3 convolution, the output characteristic pattern of first convolutional layer is having a size of 112 × 112, second convolutional layer, third convolution
Layer, Volume Four lamination and the 5th convolutional layer respectively include the convolution that two packet sizes are 3 × 3, and the output of second convolutional layer is special
Figure is levied having a size of 56 × 56, the output characteristic pattern of the third convolutional layer is having a size of 28 × 28, the output of second convolutional layer
Characteristic pattern is having a size of 14 × 14, and for the output characteristic pattern of second convolutional layer having a size of 7 × 7, the average pond layer includes one
The pond that packet size is 7 × 7, for the output characteristic pattern of the average pond layer having a size of 1 × 1, the full articulamentum uses Softmax
Function is as its classification function.
4. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 1, feature
It is, it is described that the convolutional neural networks model after training is verified using verifying collection, comprising:
If there is error result when verifying, according to classification belonging to the verifying collection image for error result occur, in the training
The amount of images for increasing respective classes is concentrated to carry out continuing to train to the convolutional neural networks model.
5. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 1, feature
It is, it is described that the convolutional neural networks model is trained using the training set, comprising:
The convolutional neural networks model is trained using Softmax loss function.
6. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 1, feature
It is, it is described that the convolutional neural networks model is trained using the training set, comprising:
The convolutional neural networks model is trained using stochastic gradient descent algorithm SGD.
7. a kind of image of clothing classification and search method based on convolutional neural networks according to claim 1, feature
It is, before the use training set is trained the convolutional neural networks model further include:
The training set and verifying are integrated into image preprocessing as fixed dimension 256 × 256.
8. a kind of image classification and searching system based on convolutional neural networks, which is characterized in that the system comprises:
Data set processing module, for the garment image data collection for having classification information to be divided into training set and verifying collection；
Model construction module, for constructing convolutional neural networks model；
Model training module, for being trained using the training set to the convolutional neural networks model；
Model authentication module, for being verified using verifying collection to the convolutional neural networks model after training；
Feature vector library constructs module, for concentrating figure using garment image data described in the convolutional neural networks model extraction
The feature vector of picture, and establish feature vector library；
Test module, for using the feature vector of the convolutional neural networks model extraction image to be retrieved；
The feature vector of the image to be retrieved and the feature vector in described eigenvector library are carried out using cosine similarity
It compares one by one, exports the garment image data and concentrate and multiple most like preceding images of the image to be retrieved.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910492484.9A CN110210567A (en) | 2019-06-06 | 2019-06-06 | A kind of image of clothing classification and search method and system based on convolutional neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910492484.9A CN110210567A (en) | 2019-06-06 | 2019-06-06 | A kind of image of clothing classification and search method and system based on convolutional neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110210567A true CN110210567A (en) | 2019-09-06 |
Family
ID=67791356
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910492484.9A Pending CN110210567A (en) | 2019-06-06 | 2019-06-06 | A kind of image of clothing classification and search method and system based on convolutional neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110210567A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110751209A (en) * | 2019-10-18 | 2020-02-04 | 北京邮电大学 | Intelligent typhoon intensity determination method integrating depth image classification and retrieval |
| CN111415221A (en) * | 2020-03-19 | 2020-07-14 | 重庆邮电大学 | Interpretable convolutional neural network-based garment recommendation method and device and terminal |
| CN111444974A (en) * | 2020-04-02 | 2020-07-24 | 成都三零凯天通信实业有限公司 | Clothing classification method based on zero sample recognition |
| CN111782837A (en) * | 2019-10-10 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Image retrieval method and device |
| CN113420173A (en) * | 2021-06-22 | 2021-09-21 | 桂林电子科技大学 | Minority dress image retrieval method based on quadruple deep learning |
| CN113435485A (en) * | 2021-06-15 | 2021-09-24 | 北京百度网讯科技有限公司 | Picture detection method and device, electronic equipment and storage medium |
| CN113743420A (en) * | 2021-08-26 | 2021-12-03 | 北京邮电大学 | Web AR image recognition method and system based on cloud edge-side cooperation |
| EP4195135A4 (en) * | 2021-10-11 | 2023-06-14 | Rakuten Group, Inc. | Information processing device, information processing method, information processing system, and program |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955718A (en) * | 2014-05-15 | 2014-07-30 | 厦门美图之家科技有限公司 | Image subject recognition method |
| CN103971342A (en) * | 2014-05-21 | 2014-08-06 | 厦门美图之家科技有限公司 | Image noisy point detection method based on convolution neural network |
| CN104504376A (en) * | 2014-12-22 | 2015-04-08 | 厦门美图之家科技有限公司 | Age classification method and system for face images |
| CN107123033A (en) * | 2017-05-04 | 2017-09-01 | 北京科技大学 | A kind of garment coordination method based on depth convolutional neural networks |
| CN107256246A (en) * | 2017-06-06 | 2017-10-17 | 西安工程大学 | PRINTED FABRIC image search method based on convolutional neural networks |
| CN107622071A (en) * | 2016-07-15 | 2018-01-23 | 上海交通大学 | By indirect correlation feedback without clothes image searching system and the method looked under source |
| CN108614884A (en) * | 2018-05-03 | 2018-10-02 | 桂林电子科技大学 | A kind of image of clothing search method based on convolutional neural networks |
| CN108984642A (en) * | 2018-06-22 | 2018-12-11 | 西安工程大学 | A kind of PRINTED FABRIC image search method based on Hash coding |
| CN109657719A (en) * | 2018-12-19 | 2019-04-19 | 浙江大学常州工业技术研究院 | A kind of image-recognizing method based on convolutional neural networks |
-
2019
- 2019-06-06 CN CN201910492484.9A patent/CN110210567A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955718A (en) * | 2014-05-15 | 2014-07-30 | 厦门美图之家科技有限公司 | Image subject recognition method |
| CN103971342A (en) * | 2014-05-21 | 2014-08-06 | 厦门美图之家科技有限公司 | Image noisy point detection method based on convolution neural network |
| CN104504376A (en) * | 2014-12-22 | 2015-04-08 | 厦门美图之家科技有限公司 | Age classification method and system for face images |
| CN107622071A (en) * | 2016-07-15 | 2018-01-23 | 上海交通大学 | By indirect correlation feedback without clothes image searching system and the method looked under source |
| CN107123033A (en) * | 2017-05-04 | 2017-09-01 | 北京科技大学 | A kind of garment coordination method based on depth convolutional neural networks |
| CN107256246A (en) * | 2017-06-06 | 2017-10-17 | 西安工程大学 | PRINTED FABRIC image search method based on convolutional neural networks |
| CN108614884A (en) * | 2018-05-03 | 2018-10-02 | 桂林电子科技大学 | A kind of image of clothing search method based on convolutional neural networks |
| CN108984642A (en) * | 2018-06-22 | 2018-12-11 | 西安工程大学 | A kind of PRINTED FABRIC image search method based on Hash coding |
| CN109657719A (en) * | 2018-12-19 | 2019-04-19 | 浙江大学常州工业技术研究院 | A kind of image-recognizing method based on convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| 孙晓峰等: "一种基于全卷积网络的目标检索方法", 《河南工程学院学报（自然科学版）》 * |
| 陈瑷瑷等: "基于关键点的服装检索", 《计算机应用》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111782837A (en) * | 2019-10-10 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Image retrieval method and device |
| CN111782837B (en) * | 2019-10-10 | 2024-04-09 | 北京沃东天骏信息技术有限公司 | Image retrieval method and device |
| CN110751209A (en) * | 2019-10-18 | 2020-02-04 | 北京邮电大学 | Intelligent typhoon intensity determination method integrating depth image classification and retrieval |
| CN111415221A (en) * | 2020-03-19 | 2020-07-14 | 重庆邮电大学 | Interpretable convolutional neural network-based garment recommendation method and device and terminal |
| CN111415221B (en) * | 2020-03-19 | 2023-04-07 | 重庆邮电大学 | Clothing recommendation method and device based on interpretable convolutional neural network and terminal |
| CN111444974A (en) * | 2020-04-02 | 2020-07-24 | 成都三零凯天通信实业有限公司 | Clothing classification method based on zero sample recognition |
| CN113435485A (en) * | 2021-06-15 | 2021-09-24 | 北京百度网讯科技有限公司 | Picture detection method and device, electronic equipment and storage medium |
| CN113420173A (en) * | 2021-06-22 | 2021-09-21 | 桂林电子科技大学 | Minority dress image retrieval method based on quadruple deep learning |
| CN113743420A (en) * | 2021-08-26 | 2021-12-03 | 北京邮电大学 | Web AR image recognition method and system based on cloud edge-side cooperation |
| CN113743420B (en) * | 2021-08-26 | 2023-12-05 | 北京邮电大学 | Web AR image recognition method and system based on cloud edge end cooperation |
| EP4195135A4 (en) * | 2021-10-11 | 2023-06-14 | Rakuten Group, Inc. | Information processing device, information processing method, information processing system, and program |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110210567A (en) | A kind of image of clothing classification and search method and system based on convolutional neural networks | |
| CN108564029B (en) | Face attribute recognition method based on cascade multitask learning deep neural network | |
| Carrière et al. | Stable topological signatures for points on 3d shapes | |
| Forsyth et al. | Finding pictures of objects in large collections of images | |
| CN106021575A (en) | Retrieval method and device for same commodities in video | |
| CN109241317A (en) | Based on the pedestrian's Hash search method for measuring loss in deep learning network | |
| CN107871100A (en) | The training method and device of faceform, face authentication method and device | |
| CN109543602A (en) | A kind of recognition methods again of the pedestrian based on multi-view image feature decomposition | |
| CN110825899A (en) | Clothing image retrieval method integrating color features and residual network depth features | |
| CN104866524A (en) | Fine classification method for commodity images | |
| CN106503729A (en) | A kind of generation method of the image convolution feature based on top layer weights | |
| CN109544501A (en) | A kind of transmission facility defect inspection method based on unmanned plane multi-source image characteristic matching | |
| CN104281679A (en) | Goods classification method and goods classification device both based on image features | |
| US20180232400A1 (en) | Sketch-based image searching system using cell-orientation histograms and outline extraction based on medium-level features | |
| CN108682007A (en) | Jpeg image resampling automatic testing method based on depth random forest | |
| Xiong et al. | An asian face dataset and how race influences face recognition | |
| CN110222587A (en) | A kind of commodity attribute detection recognition methods again based on characteristic pattern | |
| CN107977948A (en) | A kind of notable figure fusion method towards sociogram's picture | |
| CN104281572A (en) | Target matching method and system based on mutual information | |
| CN110516533A (en) | A kind of pedestrian based on depth measure discrimination method again | |
| CN116052218B (en) | Pedestrian re-identification method | |
| CN110263855A (en) | A method of it is projected using cobasis capsule and carries out image classification | |
| CN109670423A (en) | A kind of image identification system based on deep learning, method and medium | |
| CN116524583A (en) | Cross-domain pedestrian re-identification method based on point-by-point feature matching | |
| Xu et al. | Robust joint representation of intrinsic mean and kernel function of lie group for remote sensing scene classification |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190906 |