CN107683469A - A kind of product classification method and device based on deep learning - Google Patents
A kind of product classification method and device based on deep learning Download PDFInfo
- Publication number
- CN107683469A CN107683469A CN201580001265.6A CN201580001265A CN107683469A CN 107683469 A CN107683469 A CN 107683469A CN 201580001265 A CN201580001265 A CN 201580001265A CN 107683469 A CN107683469 A CN 107683469A
- Authority
- CN
- China
- Prior art keywords
- product
- image
- characteristic
- words
- text
- 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
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000013135 deep learning Methods 0.000 title claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 31
- 238000013145 classification model Methods 0.000 claims abstract description 16
- 230000006870 function Effects 0.000 claims description 31
- 238000000605 extraction Methods 0.000 claims description 19
- 238000010586 diagram Methods 0.000 claims description 15
- 238000012216 screening Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 9
- 238000012706 support-vector machine Methods 0.000 claims description 9
- 238000005286 illumination Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims 1
- 239000000047 product Substances 0.000 description 215
- 238000004590 computer program Methods 0.000 description 7
- 239000013598 vector Substances 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of product classification method and device based on deep learning, wherein, this method comprises the following steps：The text feature of product is extracted from the description text of product；The convolutional neural networks model obtained based on pre-training, the characteristics of image of product is extracted from the image of product；By the text feature of product and the multi-features of product, the characteristic information of product is obtained；The product classification model obtained based on pre-training is handled the characteristic information of product, obtains the classification results of product.Because the program has considered the product text feature and product characteristics of image of product to be sorted, compared with only carrying out product classification according to the text message of product, classification accuracy is improved.
Description
The invention relates to the technical field of pattern recognition, in particular to a product classification method and device based on deep learning.
With the rapid development of electronic commerce, online shopping has gradually become a daily behavior of online citizens. The network products are various and large in quantity, and the E-commerce website needs to expend great energy in the aspect of article management to provide good shopping experience for users. The product classification problem is the primary problem of article management, however, at present, the product classification mainly depends on manual product classification. Although most of the existing intelligent classification methods use text information of products for classification, since characters cannot completely describe all contents of pictures, products are wrongly classified under the condition of deviation of description of the character information, and a lot of labor cost is needed to correct product categories, the existing product classification methods are poor in classification accuracy.
Disclosure of Invention
The embodiment of the invention provides a product classification method based on deep learning, which solves the technical problem of poor product classification accuracy according to the text information of a product in the prior art. The product classification method comprises the following steps:
extracting the text characteristics of the product from the description text of the product;
extracting image features of the product from the image of the product based on a convolutional neural network model obtained by pre-training;
fusing the text characteristic of the product and the image characteristic of the product to obtain characteristic information of the product;
and processing the characteristic information of the product based on the product classification model obtained by pre-training to obtain the classification result of the product.
The embodiment of the invention also provides a product classification device based on deep learning, which solves the technical problem of poor product classification accuracy according to the text information of the product in the prior art. This product classification device includes:
the text feature extraction module is used for extracting the text features of the product from the description text of the product;
the image feature extraction module is used for extracting the image features of the product from the image of the product based on the convolutional neural network model obtained by pre-training;
the characteristic information acquisition module is used for fusing the text characteristic of the product with the image characteristic of the product to obtain the characteristic information of the product;
and the classification module is used for processing the characteristic information of the product based on the product classification model obtained by pre-training to obtain the classification result of the product.
In the embodiment of the invention, the text characteristic and the image characteristic of the product are extracted, and then the text characteristic of the product is fused with the image characteristic of the product to obtain the characteristic information of the product, so that the characteristic information of the product is utilized for classification to obtain the classification result.
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a flowchart of a deep learning-based product classification method according to an embodiment of the present invention;
fig. 2 is a flowchart of a text feature extraction method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a pre-training network according to an embodiment of the present invention;
fig. 4 is a schematic flow chart of an image feature extraction method according to an embodiment of the present invention;
FIG. 5 is a flow chart of a training model and a predictive product according to an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a deep learning-based product classification device according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a text feature extraction module according to an embodiment of the present invention.
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
The existing method for classifying products only simply uses text information of the products for classification, if deviation occurs in the description of the text information, the products are wrongly classified, a lot of labor cost is needed for correcting the product types, and the classification accuracy is poor. If the text information and the image information of the product are used in a combined mode, the problem that the existing classification method is poor in classification accuracy can be solved. Based on the above, the invention provides a product classification method and device based on deep learning.
Fig. 1 is a flowchart of a deep learning-based product classification method according to an embodiment of the present invention, and as shown in fig. 1, the method includes:
this example classifies C-type products among internet products, such as knitwear, T-shirts, coats, pants, shirts, one-piece dresses, shoulder bags, shoes, business bags, boots, etc., with 500 products per category.
The description text of the product refers to a text for describing the product to be classified, and includes characters, symbols, numbers and the like. The description text of the product may be stored in a product text document, such that one product text corresponds to one product text document.
Step 101: the specific process of extracting the text features of the product from the description text of the product is shown in fig. 2, and includes:
inputting a product pjExtracting corresponding Text features by using a Text feature extraction method according to the given Text to obtain Tj。
The method comprises the following steps: segmenting words of a description text of a product to obtain candidate words;
each product information is taken as a document, and the document is firstly segmented into a series of word sequences. The Chinese word segmentation is carried out by adopting a Chinese academy of sciences computational Technology research Institute based on a Chinese Lexical Analysis System ICTCCLAS (Institute of Computing Technology, Chinese Lexical Analysis System) of a multilayer hidden Markov model, and the segmentation precision reaches 98.45%.
Step two: screening out product characteristic words from the candidate words according to a preset evaluation function;
the evaluation function for feature extraction of the present invention is: the system comprises five functions of a characteristic frequency function, a document frequency function, an information gain function, a mutual information function, an extraction function and a check function. The five evaluation functions can be selected from one or a combination of several, and most preferably, five evaluation functions are used to obtain five product feature words, and then the five product feature words are used comprehensively.
1) Characteristic Frequency function (TF):
and calculating the occurrence times of the candidate words in the sample document, and taking the candidate words with the occurrence times larger than or equal to a time threshold value as product characteristic words.
Specifically, firstly, traversing all candidate words (useful words), obtaining the occurrence frequency of each candidate word (useful word) in a sample document, setting a certain threshold (such as 10), deleting words with the occurrence frequency less than the threshold and having little contribution to classification, and selecting the participles more than the threshold as product feature words.
2) Document Frequency function (DF):
and calculating the proportion of the sample documents containing the candidate words in the total number of the sample documents, and taking the candidate words with the proportion within a preset range as product characteristic words.
Specifically, the document frequency P of each candidate word (useful word) t is calculated according to the formula (1)t：
Wherein n istFor the number of sample documents containing each candidate word (useful word) t, n is the total number of sample documents.
Setting a characteristic word frequency threshold (such as (0.005, 0.08)), and screening out candidate words (useful words) t in the threshold range as product characteristic words.
3) Information Gain function (IG):
and calculating the information gain weight of the candidate word, and taking the candidate word with the information gain weight larger than the information gain weight threshold as a product characteristic word.
Specifically, the information gain weight of each candidate word (useful word) t is calculated according to formula (2):
where t represents a candidate word (useful word), C represents a document category, m represents the number of categories, and P (C)i) Is represented by CiProbability of class documents appearing in the training sample set, P (t) represents probability of documents containing entry t in the training sample set, P (C)iI t) represents the conditional probability of belonging to the class C when the document contains the entry t, represents the probability of the document not containing the entry t in the training sample set, and represents the conditional probability of belonging to the class C when the document does not contain the entry t.
After the weight is calculated, a threshold (such as 0.006) is set, and useful words with the weight larger than the threshold are selected as product feature words.
4) Mutual Information function (MI):
and calculating the mutual information value of the candidate words, and taking the candidate words with the mutual information values larger than the threshold value of the mutual information value as the product characteristic words.
Specifically, each candidate word (useful word) t is calculated according to the formula (3) or (4)kWith each of the classes CiMutual information value of (a):
can also be expressed as
MI(tk，Ci)＝logP(tk|Ci)-logP(tk) (4)
Wherein, P (t)k，Ci) Is of class CiCharacteristic P (t)k) Probability of occurrence in the training sample set, P (t)k) Is tkProbability of occurrence in the entire training sample set, P (C)i) Is CiProbability of occurrence of class sample documents in the entire training sample set, P (t)k|Ci) Is tkAt CiConditional probabilities of occurrence in class sample documents.
And selecting useful words larger than a threshold value of 1.54 from the calculated mutual information values as the characteristic words.
5) Evolution and test function (Chi-square, CHI):
and calculating the relevance of the candidate words and the preset categories, and taking the candidate words with the relevance larger than a relevance threshold value as product characteristic words.
Specifically, each useful word candidate (useful word) t is calculated according to formula (5)kWith each of the classes CiThe value of which is defined as
Where n is the number of sample documents in the training sample set, P (t)k，Ci) For the occurrence of features t in a training sample setkAnd belong to class CiThe probability of occurrence of the sample document is that the feature t does not occur in the training sample setkAnd do not belong to class CiThe probability of occurrence of the sample document is the occurrence of the feature t in the training sample setkAnd do not belong to class CiThe probability of occurrence of the sample document of (2), the absence of the feature tkAnd belong to class CiThe probability of occurrence of the sample document.
And setting a correlation threshold (such as 10), and screening out useful words larger than the threshold as the characteristic words.
The above 1) to 5) can generate five groups of product feature words, correspond to five types of product feature texts, and can remarkably improve the capability of the product text feature description on the product to be classified, thereby improving the accuracy of classification.
In specific implementation, before the step two, the method further comprises the following steps: filtering out the candidate words contained in a preset stop word list.
There may be some words or words (stop words) in the candidate words that may interfere with the classification and have no value in the classification, such as the mood words, the help words, etc. Therefore, the stop word list is preset, and the characters or words which can cause classification interference are added into the stop word list, so that candidate words contained in the preset stop word list are filtered, unnecessary calculation can be avoided, and the time required by product classification is saved.
Step three: and determining the weight of the product characteristic words according to the frequency of the product characteristic words appearing in the sample documents, the total number of the sample documents and the number of the sample documents containing the product characteristic words.
Specifically, after the product feature words are selected by the five methods, the weight of each product feature word is calculated for each group of product feature words according to the formula (6):
Wi＝TFi(t，d)×n/DF(t) (6)
wherein, WiWeight of the ith product feature word, TFi(t, d) is the frequency of appearance of the product feature word t in the document d, n represents the number of documents, and DF (t) is the number of documents containing the product feature word t.
Step four: and generating the product text characteristics of the products to be classified according to the product characteristic word weight.
Specifically, after the weight of each product feature word calculated in each method is calculated according to the formula (6), the description text of each product can be converted into a vector using the product feature word as a dimension, and the attribute value of each dimension is the weight of the product feature word. Each method yields a vector, i.e., a product text feature. For a product text, five vectors, namely five product text features, can be obtained according to 1) to 5), so that the product text features of the product to be classified are obtained. By adopting the five product text characteristics, the accuracy of product classification can be improved.
Step 102: and extracting the image characteristics of the product from the image of the product based on the convolutional neural network model obtained by pre-training.
In particular, in recent years, deep learning is prominent in image classification, and in particular, a convolutional neural network can automatically learn image features, and the extracted features are stable and reliable. The embodiment of the invention collects the picture information of ten products in the Internet products for classification, such as knitwear, T-shirts, coats, trousers, shirts, one-piece dresses, single-shoulder bags, single shoes, business bags, boots and the like, wherein each product comprises 300 products. Each product will contain a descriptive text and a picture, and the embodiment will automatically learn the image characteristics of the product by using the pre-trained convolutional neural network.
The product image refers to an image including an image of a product to be classified. Color features (such as color histograms), texture features, or shape features of the product image may be extracted as the product image features.
Specifically, first, the convolutional neural network model is adopted in the embodiment of the present invention, and because the network parameters are huge, a large amount of training data is required, so that data enhancement (image enhancement) is necessary. The data enhancement (image enhancement) mode adopted by the embodiment of the invention comprises the steps of firstly scaling each product image and scaling the short edge to 256 pixels; then, turning over the image; and finally, randomly adding illumination noise, and randomly changing the contrast, brightness and the like of the image.
Then, the convolutional neural network is pre-trained:
in the embodiment of the invention, ImageNet 2012 data set is adopted to pre-train the convolutional neural network, the schematic diagram of the network is shown in FIG. 3, and specifically, five convolutional layers Ci{ N, S } i ═ 1.., 5, where N denotes the number of convolution kernels and S denotes the convolution kernel size, and each convolution layer employs a Rectified linear unit (ReLU) activation function. The parameter of each convolution layer adopted by the embodiment of the invention is C1{48, 5 × 5}, C2{128, 3 × 3}, C3{192, 3 × 3}, C4{128, 3 × 3}, C5{128, 3 × 3}, the first four convolutional layers are respectively connected with a max Pooling (max Pooling) layer, that is, the element with the maximum value is selected from a local range, and the fifth convolutional layer is connected with a multi-scale Spatial Pooling (SPP) layer. Specifically, each image with different sizes is divided into 6 × 6, 3 × 3 and 2 × 2 subblocks on average, subblock features are extracted in a max posing mode, and finally Feature vectors with dimensions of 6 × 6+3 × 3+2 × 2 ═ 49 × Feature are obtained, wherein Feature is the size of a Feature map output by the fifth convolutional layer.
And connecting three full-connection layers behind the convolutional layer, wherein the first two layers of FC1 and FC2 are 2048 nodes respectively, and the last layer is a 1000-output softmax classifier. The training network adopts a random gradient descent method, and in order to avoid overfitting, a drop (dropout) strategy with a random drop ratio of 0.5 is adopted in the first two fully-connected layers.
And then, fine tuning the pre-trained convolutional neural network:
because the ImageNet 2012 data set used for training is 1000 categories, the trained convolutional neural network outputs 1000-way, and the embodiment classifies the Internet products into C categories, so that the last layer of the fully-connected layer is changed into C nodes, and the last layer of the fully-connected layer of the network is finely adjusted by the Internet products. The fine tuning employs a random gradient descent method, with momentum set to 0.9, weight attenuation set to 0.0005, and an initial learning rate set to 0.01, with the learning rate gradually decreasing as the number of iterations increases.
And finally, extracting the image characteristics of the product from the product image based on the fine-tuned pre-trained convolutional neural network.
In the embodiment of the invention, a test image is input into a pre-trained convolutional neural network, the convolutional neural network is used for abstracting the characteristics of the image, the image characteristics of a higher level can be extracted through a five-layer convolutional neural network, the image characteristics are pulled into a one-dimensional vector through a full connection layer, and the output of a second full connection layer is selected as the image characteristics tjSee fig. 4.
Step 103: fusing the text characteristic of the product and the image characteristic of the product to obtain characteristic information of the product;
the text characteristic of the product and the image characteristic of the product are both one-dimensional vectors, and the text characteristic of the product and the image characteristic of the product are spliced together to serve as the characteristic P of the jth productj＝{xj，tj}。
Step 104: and processing the characteristic information of the product based on the product classification model obtained by pre-training to obtain the classification result of the product.
In specific implementation, among the existing numerous intelligent classification methods, the classification and training speed of a Support Vector Machine (SVM) technology is high, the generalization capability of a model is strong, and the SVM technology is taken as a hotspot and a key point in the relevant field of machine learning. The basic idea is to create a hyperplane or a series of hyperplanes in a high dimensional space such that the distance between the hyperplane and the nearest neighboring training sample is maximized. An important task in SVM technology is the selection of the kernel function. When the sample characteristics have heterogeneous information, the sample scale is large, the multidimensional data are irregular or the data are not flat in high-order characteristic space distribution, it is not reasonable to process all samples by adopting a single-core mapping mode, that is, a multi-core learning method and a combination of a plurality of kernel functions are required.
One of the most common and most common methods for constructing multi-kernel learning is to consider a convex combination of multiple kernel functions, as shown by the following equation:
in the formula KjIs the basic kernel function, M is the total number of basic sums, βjIs a weight coefficient
The method for synthesizing the kernel is multiple, the embodiment of the invention adopts a multi-kernel learning method based on sparse coding and proposed by France sco, and the improvement of sparsity can reduce redundancy and improve the operation efficiency under some conditions.
Specifically, the product classification model is obtained through pre-training by a multi-core learning algorithm according to the following mode:
extracting text characteristics of the product samples from the description texts of the product samples in the training sample set;
extracting image features of the product samples from the images of the product samples in the training sample set based on a convolutional neural network model obtained by pre-training;
fusing the text characteristics of the product sample and the image characteristics of the product sample to obtain the characteristic information of the product sample;
training the characteristic information of the product sample to obtain a product classification model based on a support vector machine;
the training sample set comprises a plurality of product samples of preset categories, and the product samples comprise description texts and images of the product samples.
According to the multi-core learning algorithm, a plurality of product feature information are input into a product classification Model for processing, and then the classification label of the product can be obtainedjThe flow chart is shown in figure 5.
In specific implementation, because each product in the internet product has one text description and a plurality of product images, the embodiment of the invention automatically learns the image characteristics of the product by using a convolutional neural network which is not limited by the size of the input image, fuses the image characteristics and the text characteristics, finally performs cross-sample maximized pooling max Pooling on different image sample prediction results of each product, and selects the prediction result with the strongest response to the category in each product as the prediction category of each product so as to automatically eliminate noise information, thereby improving the accuracy of automatic classification in internet product classification.
The method is tested on the AMAX server platform, and higher accuracy can be obtained in product classification than the method of combining and classifying manually-made image features and text information.
Based on the same inventive concept, the embodiment of the present invention further provides a product classification device based on deep learning, as described in the following embodiments. Because the principle of the product classification device based on deep learning for solving the problem is similar to the product classification method based on deep learning, the implementation of the product classification device based on deep learning can refer to the implementation of the product classification method based on deep learning, and repeated parts are not repeated. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 6 is a block diagram of a product classification device based on deep learning according to an embodiment of the present invention, and as shown in fig. 6, the product classification device includes:
a text feature extraction module 601, configured to extract text features of a product from a description text of the product;
an image feature extraction module 602, configured to extract an image feature of a product from an image of the product based on a convolutional neural network model obtained through pre-training;
a feature information obtaining module 603, configured to fuse a text feature of the product with an image feature of the product to obtain feature information of the product;
the classification module 604 is configured to process feature information of the product based on the product classification model obtained through pre-training to obtain a classification result of the product.
This structure will be explained below.
In specific implementation, as shown in fig. 7, the text feature extraction module 601 specifically includes:
the word segmentation module 701 is used for segmenting words of a description text of a product to obtain candidate words;
a feature word screening module 702, configured to screen a product feature word from the candidate words according to a preset evaluation function;
a feature word weight determining module 703, configured to determine a product feature word weight according to the frequency of occurrence of the product feature word in a sample document, the total number of sample documents, and the number of sample documents containing the product feature word;
the text feature generation module 704 is configured to generate a product text feature of the product to be classified according to the product feature word weight;
wherein the description text of the product is stored in a sample document.
In specific implementation, the text feature extraction module 601 further includes:
and the candidate word filtering module is used for filtering out the candidate words contained in a preset stop word list.
In specific implementation, the feature word screening module 702 is specifically configured to:
determining the occurrence times of the candidate words in the sample document, and taking the candidate words with the occurrence times larger than or equal to a time threshold value as product characteristic words; and/or the presence of a gas in the gas,
determining the proportion of the sample documents containing the candidate words in the total number of the sample documents, and taking the candidate words with the proportion within a preset range as product characteristic words; and/or the presence of a gas in the gas,
determining an information gain weight of the candidate word, and taking the candidate word with the information gain weight larger than an information gain weight threshold as a product characteristic word; and/or the presence of a gas in the gas,
determining a mutual information value of the candidate words, and taking the candidate words with the mutual information values larger than a mutual information value threshold as product characteristic words; and/or the presence of a gas in the gas,
and determining the relevance of the candidate words and the preset categories, and taking the candidate words with the relevance larger than a relevance threshold value as product characteristic words.
In specific implementation, the feature word screening module 702 specifically determines the relevance between the candidate word and the preset category as follows:
and determining the correlation degree of the candidate words and the preset category according to the probability of whether the candidate words appear in the training sample set and whether the candidate words belong to the preset category.
In specific implementation, the classification module 604 is specifically configured to obtain a product classification model as follows:
extracting text characteristics of the product samples from the description texts of the product samples in the training sample set;
extracting image features of the product samples from the images of the product samples in the training sample set based on a convolutional neural network model obtained by pre-training;
fusing the text characteristics of the product sample and the image characteristics of the product sample to obtain the characteristic information of the product sample;
training the characteristic information of the product sample to obtain a product classification model based on a support vector machine;
the training sample set comprises a plurality of product samples of preset categories, and the product samples comprise description texts and images of the product samples.
In specific implementation, the product classification device further comprises:
the image enhancement module is used for carrying out image enhancement on the product image of the product to be classified;
the image feature extraction module 602 is further configured to extract an image feature of the product from the image of the product after image enhancement based on a convolutional neural network model obtained through pre-training.
In specific implementation, the image enhancement module 605 is specifically configured to:
scaling the image of the product according to a preset proportion;
turning over the image of the scaled product;
adding illumination noise into the image of the turned product;
the contrast and/or brightness of the image of the product incorporating the illumination noise is varied.
In summary, the invention provides a (multi-feature) product classification method and device based on deep learning, which breaks through the traditional method of extracting image features by using an artificially formulated image descriptor, directly inputs the original data of a product image into a convolutional neural network to automatically learn the image features, finally fuses the image features and the text features, and predicts the product category through an SVM classifier so as to realize automatic product classification and improve the accuracy of intelligent classification.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (18)
- A product classification method based on deep learning is characterized by comprising the following steps:extracting the text characteristics of the product from the description text of the product;extracting image features of the product from the image of the product based on a convolutional neural network model obtained by pre-training;fusing the text characteristic of the product and the image characteristic of the product to obtain characteristic information of the product;and processing the characteristic information of the product based on the product classification model obtained by pre-training to obtain the classification result of the product.
- The deep learning-based product classification method of claim 1, further comprising: pre-training to obtain a product classification model as follows:extracting text characteristics of the product samples from the description texts of the product samples in the training sample set;extracting image features of the product samples from the images of the product samples in the training sample set based on a convolutional neural network model obtained by pre-training;fusing the text characteristics of the product sample and the image characteristics of the product sample to obtain the characteristic information of the product sample;training the characteristic information of the product sample to obtain a product classification model based on a support vector machine;the training sample set comprises a plurality of product samples of preset categories, and the product samples comprise description texts and images of the product samples.
- The deep learning based product classification method according to claim 1, wherein the extracting the text features of the product from the description text of the product comprises:segmenting words of a description text of a product to obtain candidate words;screening out product characteristic words from the candidate words according to a preset evaluation function;determining a product characteristic word weight according to the frequency of the product characteristic words appearing in the sample documents, the total number of the sample documents and the number of the sample documents containing the product characteristic words;generating product text characteristics of the products to be classified according to the product characteristic word weight;wherein the description text of the product is stored in a sample document.
- The deep learning-based product classification method according to claim 3, wherein before the step of screening the candidate words for product feature words according to a preset evaluation function, the method further comprises:filtering out the candidate words contained in a preset stop word list.
- The deep learning-based product classification method according to claim 3, wherein the screening of the candidate words for product feature words according to a preset evaluation function comprises:determining the occurrence times of the candidate words in the sample document, and taking the candidate words with the occurrence times larger than or equal to a time threshold value as product characteristic words; and/or the presence of a gas in the gas,determining the proportion of the sample documents containing the candidate words in the total number of the sample documents, and taking the candidate words with the proportion within a preset range as product characteristic words; and/or the presence of a gas in the gas,determining an information gain weight of the candidate word, and taking the candidate word with the information gain weight larger than an information gain weight threshold as a product characteristic word; and/or the presence of a gas in the gas,determining a mutual information value of the candidate words, and taking the candidate words with the mutual information values larger than a mutual information value threshold as product characteristic words; and/or the presence of a gas in the gas,and determining the relevance of the candidate words and the preset categories, and taking the candidate words with the relevance larger than a relevance threshold value as product characteristic words.
- The deep learning-based product classification method of claim 5, wherein the determining the relevance of the candidate words to the preset category comprises:and determining the correlation degree of the candidate words and the preset category according to the probability of whether the candidate words appear in the training sample set and whether the candidate words belong to the preset category.
- The deep learning-based product classification method of claim 1, further comprising:performing image enhancement on an image of a product;based on the convolutional neural network model obtained by pre-training, extracting the image characteristics of the product from the image of the product, and further comprising:and extracting the image characteristics of the product from the image of the product after image enhancement based on the pre-trained convolutional neural network model.
- The deep learning-based product classification method of claim 7, wherein the image enhancing the image of the product comprises:scaling the image of the product according to a preset proportion;turning over the image of the scaled product;adding illumination noise into the image of the turned product;the contrast and/or brightness of the image of the product incorporating the illumination noise is varied.
- The deep learning-based product classification method according to claim 1, wherein the extracting image features of the product from the image of the product based on the pre-trained convolutional neural network model specifically comprises:extracting a product characteristic diagram from an image of a product through a five-layer convolutional neural network;dividing the product feature map into product feature map sub-blocks with different sizes;and extracting the product image features from the product feature map sub-blocks by adopting a maximization pooling method.
- A product classification device based on deep learning, comprising:the text feature extraction module is used for extracting the text features of the product from the description text of the product;the image feature extraction module is used for extracting the image features of the product from the image of the product based on the convolutional neural network model obtained by pre-training;the characteristic information acquisition module is used for fusing the text characteristic of the product with the image characteristic of the product to obtain the characteristic information of the product;and the classification module is used for processing the characteristic information of the product based on the product classification model obtained by pre-training to obtain the classification result of the product.
- The deep learning-based product classification device of claim 10, wherein the classification module is specifically configured to obtain a product classification model as follows:extracting text characteristics of the product samples from the description texts of the product samples in the training sample set;extracting image features of the product samples from the images of the product samples in the training sample set based on a convolutional neural network model obtained by pre-training;fusing the text characteristics of the product sample and the image characteristics of the product sample to obtain the characteristic information of the product sample;training the characteristic information of the product sample to obtain a product classification model based on a support vector machine;the training sample set comprises a plurality of product samples of preset categories, and the product samples comprise description texts and images of the product samples.
- The deep learning-based product classification device according to claim 10, wherein the text feature extraction module specifically comprises:the word segmentation module is used for segmenting the description text of the product to obtain candidate words;the characteristic word screening module is used for screening out product characteristic words from the candidate words according to a preset evaluation function;the characteristic word weight determining module is used for determining a product characteristic word weight according to the frequency of the product characteristic words appearing in the sample documents, the total number of the sample documents and the number of the sample documents containing the product characteristic words;the text feature generation module is used for generating the product text features of the products to be classified according to the product feature word weights;wherein the description text of the product is stored in a sample document.
- The deep learning-based product classification device of claim 12, wherein the text feature extraction module further comprises:and the candidate word filtering module is used for filtering out the candidate words contained in a preset stop word list.
- The deep learning-based product classification device of claim 12, wherein the feature word screening module is specifically configured to:determining the occurrence times of the candidate words in the sample document, and taking the candidate words with the occurrence times larger than or equal to a time threshold value as product characteristic words; and/or the presence of a gas in the gas,determining the proportion of the sample documents containing the candidate words in the total number of the sample documents, and taking the candidate words with the proportion within a preset range as product characteristic words; and/or the presence of a gas in the gas,determining an information gain weight of the candidate word, and taking the candidate word with the information gain weight larger than an information gain weight threshold as a product characteristic word; and/or the presence of a gas in the gas,determining a mutual information value of the candidate words, and taking the candidate words with the mutual information values larger than a mutual information value threshold as product characteristic words; and/or the presence of a gas in the gas,and determining the relevance of the candidate words and the preset categories, and taking the candidate words with the relevance larger than a relevance threshold value as product characteristic words.
- The deep learning-based product classification device of claim 14, wherein the feature word screening module determines the degree of correlation between the candidate word and the preset category specifically as follows:and determining the correlation degree of the candidate words and the preset category according to the probability of whether the candidate words appear in the training sample set and whether the candidate words belong to the preset category.
- The deep learning based product classification apparatus of claim 10, further comprising:the image enhancement module is used for carrying out image enhancement on the product image of the product to be classified;the image feature extraction module is further used for extracting the image features of the product from the image of the product after image enhancement based on the convolutional neural network model obtained through pre-training.
- The deep learning-based product classification apparatus of claim 16, wherein the image enhancement module is specifically configured to:scaling the image of the product according to a preset proportion;turning over the image of the scaled product;adding illumination noise into the image of the turned product;the contrast and/or brightness of the image of the product incorporating the illumination noise is varied.
- The deep learning-based product classification device of claim 16, wherein the image feature extraction module is specifically configured to:extracting a product characteristic diagram from an image of a product through a five-layer convolutional neural network;dividing the product feature map into product feature map sub-blocks with different sizes;and extracting the product image features from the product feature map sub-blocks by adopting a maximization pooling method.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2015/099965 WO2017113232A1 (en) | 2015-12-30 | 2015-12-30 | Product classification method and apparatus based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107683469A true CN107683469A (en) | 2018-02-09 |
Family
ID=59224011
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201580001265.6A Pending CN107683469A (en) | 2015-12-30 | 2015-12-30 | A kind of product classification method and device based on deep learning |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN107683469A (en) |
| WO (1) | WO2017113232A1 (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108960412A (en) * | 2018-06-29 | 2018-12-07 | 北京京东尚科信息技术有限公司 | Image-recognizing method, device and computer readable storage medium |
| CN109002852A (en) * | 2018-07-11 | 2018-12-14 | 腾讯科技（深圳）有限公司 | Image processing method, device, computer readable storage medium and computer equipment |
| CN109830271A (en) * | 2019-01-15 | 2019-05-31 | 安徽理工大学 | A kind of health data management system and analysis method based on edge calculations and cloud computing |
| CN109947526A (en) * | 2019-03-29 | 2019-06-28 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
| CN110020435A (en) * | 2019-04-03 | 2019-07-16 | 湖北工业大学 | A method of using parallel binary bat algorithm optimization text feature selection |
| CN110033008A (en) * | 2019-04-29 | 2019-07-19 | 同济大学 | A kind of iamge description generation method concluded based on modal transformation and text |
| CN110413997A (en) * | 2019-07-16 | 2019-11-05 | 深圳供电局有限公司 | New word discovery method, system and readable storage medium for power industry |
| CN110413769A (en) * | 2018-04-25 | 2019-11-05 | 北京京东尚科信息技术有限公司 | Scene classification method, device, storage medium and its electronic equipment |
| CN110533018A (en) * | 2018-05-23 | 2019-12-03 | 北京国双科技有限公司 | A kind of classification method and device of image |
| CN111191025A (en) * | 2018-11-15 | 2020-05-22 | 腾讯科技（北京）有限公司 | Method and device for determining article relevance, readable medium and electronic equipment |
| CN112329835A (en) * | 2020-10-30 | 2021-02-05 | 天河超级计算淮海分中心 | Image processing method, electronic device, and storage medium |
| CN112580329A (en) * | 2019-09-30 | 2021-03-30 | 北京国双科技有限公司 | Text noise data identification method and device, computer equipment and storage medium |
| CN113344121A (en) * | 2021-06-29 | 2021-09-03 | 北京百度网讯科技有限公司 | Method for training signboard classification model and signboard classification |
| WO2023178798A1 (en) * | 2022-03-25 | 2023-09-28 | 平安科技（深圳）有限公司 | Image classification method and apparatus, and device and medium |
Families Citing this family (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532571B (en) * | 2017-09-12 | 2022-11-18 | 腾讯科技（深圳）有限公司 | Text processing method and related device |
| CN109598517B (en) * | 2017-09-29 | 2023-09-12 | 阿里巴巴集团控股有限公司 | Commodity clearance processing, object processing and category prediction method and device thereof |
| CN109598275A (en) * | 2017-09-30 | 2019-04-09 | 富士通株式会社 | Feature selecting device, method and electronic equipment |
| CN107844257B (en) * | 2017-11-06 | 2020-04-21 | Oppo广东移动通信有限公司 | Information display method, device, terminal and storage medium |
| CN108470184A (en) * | 2018-02-11 | 2018-08-31 | 青岛海尔智能技术研发有限公司 | Food materials recognition methods, identification device and household appliance |
| CN108491867A (en) * | 2018-03-12 | 2018-09-04 | 苏州卓融新能源科技有限公司 | Image Matching based on artificial intelligence and recognition methods |
| US11055557B2 (en) | 2018-04-05 | 2021-07-06 | Walmart Apollo, Llc | Automated extraction of product attributes from images |
| TWI705414B (en) * | 2018-05-29 | 2020-09-21 | 長庚醫療財團法人林口長庚紀念醫院 | Classification system and classification method of autoantibody immunofluorescence image |
| CN110704880B (en) * | 2018-06-19 | 2022-12-02 | 中国石化工程建设有限公司 | Correlation method of engineering drawings |
| CN109271523A (en) * | 2018-11-23 | 2019-01-25 | 中电科大数据研究院有限公司 | A kind of government document subject classification method based on information retrieval |
| CN109858027A (en) * | 2019-01-22 | 2019-06-07 | 北京万诚信用评价有限公司 | One tool method for identifying and classifying of internet four product of electric business merchandise news |
| CN111581368A (en) * | 2019-02-19 | 2020-08-25 | 中国科学院信息工程研究所 | Intelligent expert recommendation-oriented user image drawing method based on convolutional neural network |
| CN111768214A (en) * | 2019-03-14 | 2020-10-13 | 北京京东尚科信息技术有限公司 | Product attribute prediction method, system, device and storage medium |
| CN110288085B (en) * | 2019-06-20 | 2022-06-03 | 厦门市美亚柏科信息股份有限公司 | Data processing method, device and system and storage medium |
| CN110502969A (en) * | 2019-07-03 | 2019-11-26 | 国网江西省电力有限公司检修分公司 | A kind of paper material key message extraction method |
| CN112348043B (en) * | 2019-08-09 | 2024-04-02 | 杭州海康机器人股份有限公司 | Feature screening method and device in machine learning |
| CN110472002B (en) * | 2019-08-14 | 2022-11-29 | 腾讯科技（深圳）有限公司 | Text similarity obtaining method and device |
| CN112489769B (en) * | 2019-08-22 | 2024-05-17 | 浙江远图技术股份有限公司 | Chronic disease intelligent Chinese medicine diagnosis and medicine recommendation system based on deep neural network |
| CN111125366B (en) * | 2019-12-25 | 2023-05-23 | 腾讯科技（深圳）有限公司 | Text classification method and device |
| CN111159200B (en) * | 2019-12-31 | 2023-10-17 | 华中科技大学鄂州工业技术研究院 | Data storage method and device based on deep learning |
| CN111401428A (en) * | 2020-03-12 | 2020-07-10 | Oppo广东移动通信有限公司 | Image classification method and device, electronic equipment and storage medium |
| CN111307805B (en) * | 2020-03-20 | 2023-07-18 | 湖南烟叶复烤有限公司郴州复烤厂 | Threshing quality detection device and detection method based on visual feature fusion |
| CN111931664B (en) * | 2020-08-12 | 2024-01-12 | 腾讯科技（深圳）有限公司 | Mixed-pasting bill image processing method and device, computer equipment and storage medium |
| CN112000807A (en) * | 2020-09-07 | 2020-11-27 | 辽宁国诺科技有限公司 | Method for accurately classifying proposal |
| CN112396111A (en) * | 2020-11-20 | 2021-02-23 | 平安普惠企业管理有限公司 | Text intention classification method and device, computer equipment and storage medium |
| CN112382388A (en) * | 2020-12-14 | 2021-02-19 | 中南大学 | Early warning method for adverse pressure sore event |
| CN112733549B (en) * | 2020-12-31 | 2024-03-01 | 厦门智融合科技有限公司 | Patent value information analysis method and device based on multiple semantic fusion |
| CN113032581B (en) * | 2021-04-09 | 2024-02-06 | 北京百度网讯科技有限公司 | Method and device for updating product list |
| CN114324361B (en) * | 2021-12-31 | 2024-03-15 | 大连理工大学 | Metallographic structure degradation rating method and device |
| CN115409549B (en) * | 2022-08-23 | 2024-05-14 | 中国民航信息网络股份有限公司 | Data processing method, system, electronic equipment and computer storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5867593A (en) * | 1993-10-20 | 1999-02-02 | Olympus Optical Co., Ltd. | Image region dividing apparatus |
| CN103366180A (en) * | 2013-06-14 | 2013-10-23 | 山东大学 | Cell image segmentation method based on automatic feature learning |
| CN103699523A (en) * | 2013-12-16 | 2014-04-02 | 深圳先进技术研究院 | Product classification method and device |
| CN104778464A (en) * | 2015-05-04 | 2015-07-15 | 中国科学院重庆绿色智能技术研究院 | Garment positioning and detecting method based on depth convolution nerve network |
| CN104866524A (en) * | 2015-04-10 | 2015-08-26 | 大连交通大学 | Fine classification method for commodity images |
| CN104992191A (en) * | 2015-07-23 | 2015-10-21 | 厦门大学 | Image classification method based on deep learning feature and maximum confidence path |
| CN105117739A (en) * | 2015-07-29 | 2015-12-02 | 南京信息工程大学 | Clothes classifying method based on convolutional neural network |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080089591A1 (en) * | 2006-10-11 | 2008-04-17 | Hui Zhou | Method And Apparatus For Automatic Image Categorization |
| CN101937445B (en) * | 2010-05-24 | 2011-12-07 | 中国科学技术信息研究所 | Automatic file classification system |
| CN104376105B (en) * | 2014-11-26 | 2017-08-25 | 北京航空航天大学 | The Fusion Features system and method for image low-level visual feature and text description information in a kind of Social Media |
-
2015
- 2015-12-30 CN CN201580001265.6A patent/CN107683469A/en active Pending
- 2015-12-30 WO PCT/CN2015/099965 patent/WO2017113232A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5867593A (en) * | 1993-10-20 | 1999-02-02 | Olympus Optical Co., Ltd. | Image region dividing apparatus |
| CN103366180A (en) * | 2013-06-14 | 2013-10-23 | 山东大学 | Cell image segmentation method based on automatic feature learning |
| CN103699523A (en) * | 2013-12-16 | 2014-04-02 | 深圳先进技术研究院 | Product classification method and device |
| CN104866524A (en) * | 2015-04-10 | 2015-08-26 | 大连交通大学 | Fine classification method for commodity images |
| CN104778464A (en) * | 2015-05-04 | 2015-07-15 | 中国科学院重庆绿色智能技术研究院 | Garment positioning and detecting method based on depth convolution nerve network |
| CN104992191A (en) * | 2015-07-23 | 2015-10-21 | 厦门大学 | Image classification method based on deep learning feature and maximum confidence path |
| CN105117739A (en) * | 2015-07-29 | 2015-12-02 | 南京信息工程大学 | Clothes classifying method based on convolutional neural network |
Non-Patent Citations (2)
| Title |
|---|
| CHRISTIAN S.PERON: "基于Python的卷积神经网络和特征提取", 《HTTP://WWW.UML.ORG.CN/OTHER/2015083110.ASP》 * |
| HJIMCE: "深度学习（十九）基于空间金字塔池化的卷积神经网络物体检测方法", 《HTTPS://BLOG.CSDN.NET/HJIMCE/ARTICLE/DETAILS/50187655》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110413769A (en) * | 2018-04-25 | 2019-11-05 | 北京京东尚科信息技术有限公司 | Scene classification method, device, storage medium and its electronic equipment |
| CN110533018A (en) * | 2018-05-23 | 2019-12-03 | 北京国双科技有限公司 | A kind of classification method and device of image |
| CN108960412A (en) * | 2018-06-29 | 2018-12-07 | 北京京东尚科信息技术有限公司 | Image-recognizing method, device and computer readable storage medium |
| CN109002852A (en) * | 2018-07-11 | 2018-12-14 | 腾讯科技（深圳）有限公司 | Image processing method, device, computer readable storage medium and computer equipment |
| CN111191025B (en) * | 2018-11-15 | 2023-12-12 | 深圳市雅阅科技有限公司 | Method and device for determining article relevance, readable medium and electronic equipment |
| CN111191025A (en) * | 2018-11-15 | 2020-05-22 | 腾讯科技（北京）有限公司 | Method and device for determining article relevance, readable medium and electronic equipment |
| CN109830271A (en) * | 2019-01-15 | 2019-05-31 | 安徽理工大学 | A kind of health data management system and analysis method based on edge calculations and cloud computing |
| CN109947526A (en) * | 2019-03-29 | 2019-06-28 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
| CN109947526B (en) * | 2019-03-29 | 2023-04-11 | 北京百度网讯科技有限公司 | Method and apparatus for outputting information |
| CN110020435B (en) * | 2019-04-03 | 2023-04-07 | 湖北工业大学 | Method for optimizing text feature selection by adopting parallel binary bat algorithm |
| CN110020435A (en) * | 2019-04-03 | 2019-07-16 | 湖北工业大学 | A method of using parallel binary bat algorithm optimization text feature selection |
| CN110033008A (en) * | 2019-04-29 | 2019-07-19 | 同济大学 | A kind of iamge description generation method concluded based on modal transformation and text |
| CN110033008B (en) * | 2019-04-29 | 2023-08-04 | 同济大学 | Image description generation method based on modal transformation and text induction |
| CN110413997B (en) * | 2019-07-16 | 2023-04-07 | 深圳供电局有限公司 | New word discovery method, system and readable storage medium for power industry |
| CN110413997A (en) * | 2019-07-16 | 2019-11-05 | 深圳供电局有限公司 | New word discovery method, system and readable storage medium for power industry |
| CN112580329A (en) * | 2019-09-30 | 2021-03-30 | 北京国双科技有限公司 | Text noise data identification method and device, computer equipment and storage medium |
| CN112580329B (en) * | 2019-09-30 | 2024-02-20 | 北京国双科技有限公司 | Text noise data identification method, device, computer equipment and storage medium |
| CN112329835A (en) * | 2020-10-30 | 2021-02-05 | 天河超级计算淮海分中心 | Image processing method, electronic device, and storage medium |
| CN113344121A (en) * | 2021-06-29 | 2021-09-03 | 北京百度网讯科技有限公司 | Method for training signboard classification model and signboard classification |
| CN113344121B (en) * | 2021-06-29 | 2023-10-27 | 北京百度网讯科技有限公司 | Method for training a sign classification model and sign classification |
| WO2023178798A1 (en) * | 2022-03-25 | 2023-09-28 | 平安科技（深圳）有限公司 | Image classification method and apparatus, and device and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2017113232A1 (en) | 2017-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107683469A (en) | A kind of product classification method and device based on deep learning | |
| CN110866140B (en) | Image feature extraction model training method, image searching method and computer equipment | |
| Yang et al. | Visual sentiment prediction based on automatic discovery of affective regions | |
| CN111079639B (en) | Method, device, equipment and storage medium for constructing garbage image classification model | |
| CN103345645B (en) | Commodity image class prediction method towards net purchase platform | |
| CN109471944B (en) | Training method and device of text classification model and readable storage medium | |
| US10867169B2 (en) | Character recognition using hierarchical classification | |
| CN107766323B (en) | Text feature extraction method based on mutual information and association rule | |
| JP6897749B2 (en) | Learning methods, learning systems, and learning programs | |
| CN110751027B (en) | Pedestrian re-identification method based on deep multi-instance learning | |
| CN112163114B (en) | Image retrieval method based on feature fusion | |
| CN111080551B (en) | Multi-label image complement method based on depth convolution feature and semantic neighbor | |
| CN111460200B (en) | Image retrieval method and model based on multitask deep learning and construction method thereof | |
| CN110008365B (en) | Image processing method, device and equipment and readable storage medium | |
| CN106250909A (en) | A kind of based on the image classification method improving visual word bag model | |
| CN108470025A (en) | Partial-Topic probability generates regularization own coding text and is embedded in representation method | |
| CN111259823A (en) | Pornographic image identification method based on convolutional neural network | |
| CN114461890A (en) | Hierarchical multi-modal intellectual property search engine method and system | |
| CN108154156A (en) | Image Ensemble classifier method and device based on neural topic model | |
| Chen et al. | Deep net architectures for visual-based clothing image recognition on large database | |
| Chen et al. | Visual-based deep learning for clothing from large database | |
| CN111144453A (en) | Method and equipment for constructing multi-model fusion calculation model and method and equipment for identifying website data | |
| Meng et al. | Concept-concept association information integration and multi-model collaboration for multimedia semantic concept detection | |
| CN117636032A (en) | Multi-label image classification method based on multi-scale local features and difficult class mining | |
| Bahrami et al. | Automatic image annotation using an evolutionary algorithm (IAGA) |
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: 20180209 |