CN111191723A - Few-sample commodity classification system and method based on cascade classifier - Google Patents
Few-sample commodity classification system and method based on cascade classifier Download PDFInfo
- Publication number
- CN111191723A CN111191723A CN201911398741.9A CN201911398741A CN111191723A CN 111191723 A CN111191723 A CN 111191723A CN 201911398741 A CN201911398741 A CN 201911398741A CN 111191723 A CN111191723 A CN 111191723A
- Authority
- CN
- China
- Prior art keywords
- commodity
- few
- sample
- classification
- commodities
- 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.)
- Granted
Links
Images
Classifications
-
- 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
- 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
- G06F18/2148—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the process organisation or structure, e.g. boosting cascade
Abstract
The invention discloses a few-sample commodity classification system based on a cascade classifier, which extracts image features corresponding to few-sample commodities through multiple layers, correspondingly inputs the image features extracted by each layer into each level of the cascade classifier, calculates classification weights of the few-sample commodities as corresponding commodity classes through the cascade classifier, outputs classification results of the few-sample commodities according to the image features extracted by the last layer and the classification weights finally obtained by the cascade classifier by analyzing the inter-class relevance between the few-sample commodities and basic commodities, calculates the classification weights of the hidden features extracted by each hidden layer of a feature extractor through the cascade classifier, then cascades and updates the classification weights of the few-sample commodities based on the inter-class distinction degree between the few-sample commodities and the basic commodities, the accuracy and the classification efficiency of classifying the commodities with few samples are improved.
Description
Technical Field
The invention relates to the technical field of commodity classification, in particular to a cascade classifier-based few-sample commodity classification system and a classification method.
Background
Currently, article classification methods based on visual recognition technology generally rely on a large amount of training data to train a classification model and output a classification of an article through classification model recognition. However, in some specific application scenarios, some types of data are difficult to obtain, for example, some cold commodities, new commodities, and the like, the commodity data of the commodities are very limited, and therefore, an effective commodity classifier cannot be trained by using the limited commodity data as a training sample to classify and identify the cold commodities.
In the prior art, for the classification and identification of the few-sample commodities, a feature extractor and a classifier are mostly adopted, the feature extractor extracts the commodity image features of the few-sample commodities, and then the classifier obtains the classification result of the few-sample features according to the commodity image features. The classifiers only classify the final output of the feature extractor, but do not classify the hidden layer features of the feature extractor, so that the inter-class discrimination of the hidden layer features of the feature extractor is not high, the inter-class discrimination of the image features finally output by the feature extractor is influenced, and the accuracy of classification judgment of the few-sample commodities is finally influenced.
Disclosure of Invention
The invention aims to provide a few-sample commodity classification system based on a cascade classifier so as to solve the technical problem.
In order to achieve the purpose, the invention adopts the following technical scheme:
the provided few-sample commodity classification system based on the cascade classifier is used for carrying out commodity classification on commodities lacking sample characteristics and comprises the following steps:
the characteristic extractor training module is used for training and forming a characteristic extractor according to a plurality of input basic commodity characteristic samples and storing the characteristic extractor;
the classifier training module is used for training a cascade classifier according to a plurality of input basic commodity class samples and storing the cascade classifier;
the few-sample commodity image acquisition module is used for acquiring commodity images of few-sample commodities;
the characteristic extraction module is respectively connected with the few-sample commodity image acquisition module and the characteristic extractor training module and is used for inputting the commodity images of the few-sample commodities into the characteristic extractor, and then the characteristic extractor extracts multilayer image characteristics related to the few-sample commodities in a multilayer image characteristic extraction mode;
a classification module, respectively connected to the feature extraction module and the classifier training module, for correspondingly inputting the extracted image features of each layer into the classifier of each stage of the cascade classifier,
the cascade classifier calculates and obtains classification weights of the few-sample commodities as corresponding commodity classes in each classification layer according to input image features, updates the classification weights in a cascade mode by analyzing the relevance between the few-sample commodities and basic commodities, classifies the few-sample commodities based on the image features input at the last layer and the classification weights finally obtained through the cascade updating, and finally outputs classification results of the few-sample commodities.
As a preferable aspect of the present invention, the image feature extracted by the feature extractor includes a hidden feature of the commodity image of the few-sample commodity after image feature extraction.
As a preferred embodiment of the present invention, each stage of the cascade of classifiers specifically includes:
an image feature input unit for inputting the image features extracted by the feature extractor correspondingly;
a classification weight calculation unit connected to the image feature input unit and configured to calculate the classification weight of the few-sample commodity as the corresponding commodity class according to the input image feature;
the inter-class relevance analysis unit is connected with the image feature input unit and used for analyzing and obtaining the inter-class relevance of the few-sample commodity and the basic commodity according to the input image features;
a classification weight updating unit, respectively connected to the classification weight calculating unit and the inter-class correlation analyzing unit, for updating the classification weight of the few-sample commodity as the corresponding commodity class according to the inter-class correlation;
and the classification unit is connected with the image characteristic input unit and the classification weight updating unit and is used for classifying the commodities with few samples according to the input image characteristics and the classification weight after final updating and obtaining the classification result.
As a preferable aspect of the present invention, the few-sample commodity classification system obtains the inter-class association between the few-sample commodity and the base commodity by analyzing the image features associated with the few-sample commodity and an attention mechanism.
The invention also provides a few-sample commodity classification method based on the cascade classifier, which is realized by applying the few-sample commodity classification system and comprises the following steps:
step S1, the few-sample commodity classification system acquires the commodity image of the few-sample commodity;
step S2, the few-sample commodity classification system extracts the image features of the commodity images based on the pre-trained feature extractor, and extracts the multilayer image features related to the few-sample commodities;
step S3, the few-sample commodity classification system correspondingly inputs the image features of each layer into the classifier of each level of the cascade classifier;
step S4, the cascade classifier calculates the classification weight of the low-sample commodity in each classification layer as the corresponding commodity class according to the input image features, and updates the classification weight in a cascade manner by analyzing the inter-class correlation between the low-sample commodity and the basic commodity;
step S5, the cascade classifier performs commodity classification on the few-sample commodity based on the image features extracted at the last layer and the classification weights finally obtained through cascade update, and finally outputs the classification result of the few-sample commodity.
In a preferred embodiment of the present invention, in step S2, the image feature includes a hidden feature extracted by the feature extractor and associated with the few-sample commodity.
As a preferable aspect of the present invention, in step S4, the method for calculating the classification weight of the few-sample commodity as the corresponding commodity class by the cascade classifier specifically includes the following steps:
step L1, calculating a feature mean value F of the few-sample commodity according to the input image features;
step L2, putting the characteristic mean value F into a full connection layer to obtain a corresponding first characteristic W1;
step L3, calculating the feature mean value F and the corresponding correlation coefficient of the basic commodity by using an attention mechanism, and then multiplying the correlation coefficient by the preset class weight corresponding to the basic commodity to obtain a second feature W2 corresponding to the few-sample commodity;
and L4, adding the first characteristic W1 and the second characteristic W2 to obtain the classification weight W of the few-sample commodity as the commodity class.
As a preferable aspect of the present invention, in step S4, the method for analyzing the inter-class association between the small sample commodity and the base commodity by the cascade classifier specifically includes the following steps:
step M1, putting the characteristic mean value F of the few-sample commodity into a full-connection layer to obtain the first characteristic W1;
step M2, multiplying the first characteristic W1 of the few-sample commodity with the commodity characteristic of the basic commodity to obtain a first product, and then calculating a first logistic regression (Softmax) value of the first product;
step M3, multiplying the first logistic regression value calculated in step M2 by the class weight corresponding to the base commodity itself to obtain a second product, then calculating a second logistic regression value of the second product, and using the second logistic regression value as an index for evaluating the strength of the correlation between the classes of the few-sample commodity and the base commodity.
As a preferable aspect of the present invention, in step S4, the cascade classifier updates the classification weight corresponding to the low-sample commodity by the following formula:
Wn＝W1+0.5Wn-1；
in the above formula, WnThe classification weight used for representing the sample-less commodity after the nth grade classifier weight in the cascade classifier is updated is corresponding to the commodity class;
W1the classification weight used for representing the low-sample commodity calculated by the first-stage classifier in the cascade classifier is corresponding to the commodity class.
The invention has the beneficial effects that:
1. the method and the device not only classify the image features finally output by the feature extractor, but also classify the image features extracted by each hidden layer of the feature extractor through the cascade classifier, thereby enhancing the inter-class distinction degree of the hidden features of the feature extractor, being beneficial to further improving the inter-class distinction degree of the image features finally extracted by the feature extractor, and improving the accuracy and the classification efficiency of commodity classification on few-sample commodities.
2. According to the invention, the inter-class relevance between the few-sample commodities and the basic commodities is obtained through an attention mechanism, and then the classification weight of the few-sample commodities as the corresponding commodity class is updated in a cascading manner based on the inter-class relevance, so that the accuracy of classifying the few-sample commodities is effectively improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments of the present invention will be briefly described below. It is obvious that the drawings described below are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
Fig. 1 is a schematic structural diagram of a cascade classifier based low-sample commodity classification system according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the internal structure of each classifier in the low-sample product classification system according to an embodiment of the invention;
fig. 3 is a diagram of the steps of a method for classifying commodities of a few sample commodities by using the system for classifying commodities of a few sample based on a cascade classifier according to an embodiment of the present invention:
FIG. 4 is a diagram of the method steps for calculating classification weights of low-sample commodities to corresponding commodity classes by the low-sample commodity classification system based on the cascade classifiers according to an embodiment of the present invention;
FIG. 5 is a flowchart illustrating the method steps for analyzing the inter-class association between the low-sample commodities and the basic commodities by the low-sample commodity classification system based on the cascade classifier according to an embodiment of the present invention;
fig. 6 is a block diagram of a flow chart of a less-sample commodity classification system based on a cascade classifier according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is further explained by the specific implementation mode in combination with the attached drawings.
Wherein the showings are for the purpose of illustration only and are shown by way of illustration only and not in actual form, and are not to be construed as limiting the present patent; to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "inner", "outer", etc. are used for indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, it is only for convenience of description and simplification of description, but it is not indicated or implied that the referred device or element must have a specific orientation, be constructed in a specific orientation and be operated, and therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and are not to be construed as limitations of the present patent, and the specific meanings of the terms may be understood by those skilled in the art according to specific situations.
In the description of the present invention, unless otherwise explicitly specified or limited, the term "connected" or the like, if appearing to indicate a connection relationship between the components, is to be understood broadly, for example, as being fixed or detachable or integral; can be mechanically or electrically connected; they may be directly connected or indirectly connected through intervening media, or may be connected through one or more other components or may be in an interactive relationship with one another. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Referring to fig. 1, the cascade classifier based few-sample commodity classification system according to the embodiment of the present invention is used for classifying commodities lacking sample features, and the few-sample commodity classification system specifically includes:
the characteristic extractor training module 1 is used for training and forming a characteristic extractor according to a plurality of input basic commodity characteristic samples and storing the characteristic extractor;
the classifier training module 2 is used for training and forming a cascade classifier according to a plurality of input basic commodity class samples and storing the cascade classifier;
the few-sample commodity image acquisition module 3 is used for acquiring commodity images of few-sample commodities;
the characteristic extraction module 4 is respectively connected with the few-sample commodity image acquisition module 3 and the characteristic extractor training module 1 and is used for inputting the commodity images of the few-sample commodities into the characteristic extractor, and then the characteristic extractor extracts multilayer image characteristics related to the few-sample commodities in a multilayer image characteristic extraction mode;
a classification module 5 respectively connected with the feature extraction module 4 and the classifier training module 2 and used for correspondingly inputting the extracted image features of each layer into each level of classifier of the cascade classifier,
the cascade classifier calculates classification weights of the few-sample commodities as corresponding commodities according to the input image features, updates the classification weights in a cascade mode by analyzing the relevance between the few-sample commodities and the basic commodities, and finally outputs classification results of the few-sample commodities based on the image features input by each layer and the classification weights after the cascade updating.
In a preferred embodiment of this embodiment, the image features extracted by the feature extractor include hidden features of the commodity image of the few-sample commodity after image feature extraction.
Referring to fig. 2 and 6, each classifier 100 in the cascade of classifiers specifically includes:
an image feature input unit 101 for inputting the image features extracted by the feature extractor;
a classification weight calculation unit 102, connected to the image feature input unit 101, for calculating a classification weight of the few-sample commodity as a corresponding commodity class according to the input image feature;
the inter-class relevance analysis unit 103 is connected with the image feature input unit 101 and is used for analyzing the inter-class relevance between the few-sample commodities and the basic commodities according to the input image features;
a classification weight updating unit 104, respectively connected to the classification weight calculating unit 102 and the inter-class correlation analyzing unit 103, for updating and storing the classification weight of the few-sample commodity as the classification weight of the corresponding commodity class according to the inter-class correlation;
and the classification unit 105 is connected with the image feature input unit 101 and the classification weight updating unit 102, and is used for classifying the commodities with few samples according to the input image features and the classification weights after final updating to obtain a classification result.
In the above technical solution, the low-sample commodity classification system preferably obtains the inter-class relevance between the low-sample commodity and the base commodity through attention mechanism analysis according to the image features related to the low-sample commodity.
Referring to fig. 3 and fig. 6, the invention further provides a few-sample commodity classification method based on a cascade classifier, which is implemented by applying the few-sample commodity classification system, and specifically includes the following steps:
step S1, the low-sample commodity classification system obtains commodity images of low-sample commodities;
step S2, the few-sample commodity classification system extracts the image features of the commodity images based on the pre-trained feature extractor, and extracts the multilayer image features related to the few-sample commodities;
step S3, the few-sample commodity classification system correspondingly inputs the image characteristics of each layer into the classifier of each level of the cascade classifier;
step S4, calculating the classification weight of the few-sample commodities as the corresponding commodity class in each classification layer according to the input image characteristics by the cascade classifier, and updating the classification weight in a cascade mode by analyzing the relevance between the few-sample commodities and the basic commodities;
and step S5, the cascade classifier classifies the commodities of the few-sample commodities based on the image features extracted from the last layer and the classification weights finally obtained through cascade updating, and finally outputs the classification results of the few-sample commodities.
In the above technical solution, the training process of the feature extractor and the cascade classifier is the existing method, for example, the feature extractor and the cascade classifier can be obtained by using deep learning convolutional neural network training, and the specific training process is not described here.
In step S2, the extracted image features include hidden features associated with the sample-less commodity. In general, the image features are extracted through a convolutional neural network, the finally output image features do not include image hidden features, the image hidden features are generally discarded by default, and in the embodiment, the low-sample commodity classification system takes the hidden image features as classification samples, so that the accuracy of commodity classification of low-sample commodities is improved.
Referring to fig. 4, in step S4, the method for calculating the classification weight of the few-sample commodity as the corresponding commodity class by the cascade classifier specifically includes the following steps:
step L1, calculating a characteristic mean value F of the few-sample commodity according to the input image characteristics;
step L2, putting the characteristic mean value F into the full connection layer to obtain a corresponding first characteristic W1;
l3, calculating a correlation coefficient between the feature mean value F and the corresponding basic commodity by using an attention mechanism, and multiplying the correlation coefficient by a preset class weight corresponding to the basic commodity to obtain a second feature W2 corresponding to the few-sample commodity;
and L4, adding the first characteristic W1 and the second characteristic W2 to obtain the classification weight W with the commodity with less samples as the corresponding commodity class.
W＝W1+W2。
It should be noted that, the method for calculating the feature mean of the few-sample commodity by the few-sample commodity classification system through each stage of classifier is an existing feature mean calculation method, and since the feature mean calculation method is not within the scope of the claimed invention, the process for calculating the feature mean F corresponding to the few-sample commodity by the few-sample commodity classification system according to the image features related to the few-sample commodity is not described here.
Referring to fig. 5, in step S4, the method for analyzing the association between the classes of the few-sample commodities and the basic commodities by the cascade classifier specifically includes the following steps:
step M1, putting the characteristic mean value F of the few-sample commodity into the full-connection layer to obtain a corresponding first characteristic W1;
step M2, multiplying the first characteristic W1 corresponding to the few-sample commodity with the commodity characteristic of the basic commodity to obtain a first product, and then calculating a first logistic regression (Softmax) value of the first product;
and step M3, multiplying the first logistic regression value obtained by calculation in the step M2 by the corresponding category weight of the basic commodity to obtain a second product, then calculating a second logistic regression value of the second product, and taking the second logistic regression value as an index for evaluating the relevance between the categories of the few-sample commodity and the basic commodity.
It should be noted that the first logistic regression value calculated in step M2 is the correlation coefficient between the feature mean value F calculated in step L3 and the base product (i.e., the correlation coefficient between the few-sample product and the base product).
The second logistic regression value calculated in step M3 is the second characteristic W2 corresponding to the few-sample commodity calculated in step L3.
In the above technical solution, the methods for calculating the first feature W1, the first logistic regression value, and the second logistic regression value are all existing calculation methods, and the calculation processes thereof are not described herein.
The method for updating the classification weight corresponding to the few-sample commodity through cascade connection of the cascade classifier is realized through the following formula calculation:
Wn＝W1+0.5Wn-1；
in the above formula, WnThe classification weight is used for representing the few-sample commodities after the weight of the nth-level classifier in the cascade classifier is updated as the classification weight of the corresponding commodity class;
W1and the classification weight is used for indicating the low-sample commodities calculated by the first-stage classifier in the cascade classifier as the corresponding commodity class.
It should be understood that the above-described embodiments are merely preferred embodiments of the invention and the technical principles applied thereto. It will be understood by those skilled in the art that various modifications, equivalents, changes, and the like can be made to the present invention. However, such variations are within the scope of the invention as long as they do not depart from the spirit of the invention. In addition, certain terms used in the specification and claims of the present application are not limiting, but are used merely for convenience of description.
Claims (9)
1. A few-sample commodity classification system based on a cascade classifier, which is used for commodity classification of commodities lacking sample features, and is characterized by comprising:
the characteristic extractor training module is used for training and forming a characteristic extractor according to a plurality of input basic commodity characteristic samples and storing the characteristic extractor;
the classifier training module is used for training a cascade classifier according to a plurality of input basic commodity class samples and storing the cascade classifier;
the few-sample commodity image acquisition module is used for acquiring commodity images of few-sample commodities;
the characteristic extraction module is respectively connected with the few-sample commodity image acquisition module and the characteristic extractor training module and is used for inputting the commodity images of the few-sample commodities into the characteristic extractor, and then the characteristic extractor extracts multilayer image characteristics related to the few-sample commodities in a multilayer image characteristic extraction mode;
a classification module, respectively connected to the feature extraction module and the classifier training module, for correspondingly inputting the extracted image features of each layer into the classifier of each stage of the cascade classifier,
the cascade classifier calculates and obtains classification weights of the few-sample commodities as corresponding commodity classes in each classification layer according to input image features, updates the classification weights in a cascade mode by analyzing the relevance between the few-sample commodities and basic commodities, classifies the few-sample commodities based on the image features input at the last layer and the classification weights finally obtained through the cascade updating, and finally outputs classification results of the few-sample commodities.
2. The few-sample commodity classification system according to claim 1, wherein the image features extracted by the feature extractor include hidden features of the commodity image of the few-sample commodity after image feature extraction.
3. The few-sample merchandise classification system of claim 1, wherein each stage of said cascade of classifiers comprises in particular:
an image feature input unit for inputting the image features extracted by the feature extractor correspondingly;
a classification weight calculation unit connected to the image feature input unit and configured to calculate the classification weight of the few-sample commodity as the corresponding commodity class according to the input image feature;
the inter-class relevance analysis unit is connected with the image feature input unit and used for analyzing and obtaining the inter-class relevance of the few-sample commodity and the basic commodity according to the input image features;
a classification weight updating unit, respectively connected to the classification weight calculating unit and the inter-class correlation analyzing unit, for updating the classification weight of the few-sample commodity as the corresponding commodity class according to the inter-class correlation;
and the classification unit is connected with the image characteristic input unit and the classification weight updating unit and is used for classifying the commodities with few samples according to the input image characteristics and the classification weight after final updating and obtaining the classification result.
4. The system for few-sample merchandise classification of claim 1, wherein the system for few-sample merchandise classification obtains the inter-class association of the few-sample merchandise with the base merchandise based on the image features associated with the few-sample merchandise and through attention mechanism analysis.
5. A few-sample commodity classification method based on a cascade classifier is realized by applying the few-sample commodity classification system as any one of the right 1 to the right 4, and is characterized by comprising the following steps of:
step S1, the few-sample commodity classification system acquires the commodity image of the few-sample commodity;
step S2, the few-sample commodity classification system extracts the image features of the commodity images based on the pre-trained feature extractor, and extracts the multilayer image features related to the few-sample commodities;
step S3, the few-sample commodity classification system correspondingly inputs the image features of each layer into the classifier of each level of the cascade classifier;
step S4, the cascade classifier calculates the classification weight of the low-sample commodity in each classification layer as the corresponding commodity class according to the input image features, and updates the classification weight in a cascade manner by analyzing the inter-class correlation between the low-sample commodity and the basic commodity;
step S5, the cascade classifier performs commodity classification on the few-sample commodity based on the image features extracted at the last layer and the classification weights finally obtained through cascade update, and finally outputs the classification result of the few-sample commodity.
6. The few-sample commodity classification method according to claim 5, wherein in step S2, the image features include hidden features extracted by the feature extractor and associated with the few-sample commodity.
7. The method for classifying a few-sample commodity according to claim 5, wherein in said step S4, said step of calculating said classification weight of said few-sample commodity as said commodity class by said cascade classifier specifically comprises the steps of:
step L1, calculating a feature mean value F of the few-sample commodity according to the input image features;
step L2, putting the characteristic mean value F into a full connection layer to obtain a corresponding first characteristic W1;
step L3, calculating the feature mean value F and the corresponding correlation coefficient of the basic commodity by using an attention mechanism, and then multiplying the correlation coefficient by the preset class weight corresponding to the basic commodity to obtain a second feature W2 corresponding to the few-sample commodity;
and L4, adding the first characteristic W1 and the second characteristic W2 to obtain the classification weight W of the few-sample commodity as the commodity class.
8. The method for classifying a few-sample commodity according to claim 7, wherein in said step S4, the method for analyzing said inter-class association between said few-sample commodity and said base commodity by said cascade classifier specifically comprises the steps of:
step M1, putting the characteristic mean value F of the few-sample commodity into a full-connection layer to obtain the first characteristic W1;
step M2, multiplying the first characteristic W1 of the few-sample commodity with the commodity characteristic of the basic commodity to obtain a first product, and then calculating a first logistic regression (Softmax) value of the first product;
step M3, multiplying the first logistic regression value calculated in step M2 by the class weight corresponding to the base commodity itself to obtain a second product, then calculating a second logistic regression value of the second product, and using the second logistic regression value as an index for evaluating the strength of the correlation between the classes of the few-sample commodity and the base commodity.
9. The method for classifying a few-sample commodity according to claim 8, wherein in said step S4, said cascade classifier updating said classification weight corresponding to said few-sample commodity is implemented by calculating according to the following formula:
Wn＝W1+0.5Wn-1；
in the above formula, WnThe classification weight used for representing the sample-less commodity after the nth grade classifier weight in the cascade classifier is updated is corresponding to the commodity class;
W1the classification weight used for representing the low-sample commodity calculated by the first-stage classifier in the cascade classifier is corresponding to the commodity class.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911398741.9A CN111191723B (en) | 2019-12-30 | 2019-12-30 | Cascade classifier-based few-sample commodity classification system and classification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911398741.9A CN111191723B (en) | 2019-12-30 | 2019-12-30 | Cascade classifier-based few-sample commodity classification system and classification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111191723A true CN111191723A (en) | 2020-05-22 |
| CN111191723B CN111191723B (en) | 2023-06-20 |
Family
ID=70707901
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911398741.9A Active CN111191723B (en) | 2019-12-30 | 2019-12-30 | Cascade classifier-based few-sample commodity classification system and classification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111191723B (en) |
Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5793891A (en) * | 1994-07-07 | 1998-08-11 | Nippon Telegraph And Telephone Corporation | Adaptive training method for pattern recognition |
| US20080114564A1 (en) * | 2004-11-25 | 2008-05-15 | Masayoshi Ihara | Information Classifying Device, Information Classifying Method, Information Classifying Program, Information Classifying System |
| US20110250626A1 (en) * | 2002-09-09 | 2011-10-13 | Reactive Surfaces, Ltd. | Visual Assays for Coatings Incorporating Bioactive Enzymes for Catalytic Functions |
| CN103077399A (en) * | 2012-11-29 | 2013-05-01 | 西交利物浦大学 | Biological microscopic image classification method based on integrated cascade structure |
| CN104182621A (en) * | 2014-08-08 | 2014-12-03 | 同济大学 | DBN based ADHD discriminatory analysis method |
| CN104751163A (en) * | 2013-12-27 | 2015-07-01 | 同方威视技术股份有限公司 | Fluoroscopy examination system and method for carrying out automatic classification recognition on goods |
| CN106415594A (en) * | 2014-06-16 | 2017-02-15 | 北京市商汤科技开发有限公司 | A method and a system for face verification |
| CN106446895A (en) * | 2016-10-28 | 2017-02-22 | 安徽四创电子股份有限公司 | License plate recognition method based on deep convolutional neural network |
| CN106874934A (en) * | 2017-01-12 | 2017-06-20 | 华南理工大学 | Sewage disposal method for diagnosing faults based on weighting extreme learning machine Integrated Algorithm |
| CN107155360A (en) * | 2014-08-12 | 2017-09-12 | 西门子保健有限责任公司 | Multilayer polymeric for object detection |
| CN107798283A (en) * | 2016-08-31 | 2018-03-13 | 西安英诺视通信息技术有限公司 | A kind of neural network failure multi classifier based on the acyclic figure of decision-directed |
| US20180157938A1 (en) * | 2016-12-07 | 2018-06-07 | Samsung Electronics Co., Ltd. | Target detection method and apparatus |
| CN108647706A (en) * | 2018-04-24 | 2018-10-12 | 广州大学 | Article identification classification based on machine vision and flaw detection method |
| CN108921198A (en) * | 2018-06-08 | 2018-11-30 | 山东师范大学 | commodity image classification method, server and system based on deep learning |
| US20190122111A1 (en) * | 2017-10-24 | 2019-04-25 | Nec Laboratories America, Inc. | Adaptive Convolutional Neural Knowledge Graph Learning System Leveraging Entity Descriptions |
| CN109784421A (en) * | 2019-01-30 | 2019-05-21 | 北京朗镜科技有限责任公司 | A kind of construction method and device of identification model |
| CN109816028A (en) * | 2019-01-18 | 2019-05-28 | 创新奇智(南京)科技有限公司 | A kind of unbalanced dataset classification Model Fusion method of Partial Feature migration |
| CN109934261A (en) * | 2019-01-31 | 2019-06-25 | 中山大学 | A kind of Knowledge driving parameter transformation model and its few sample learning method |
| CN109993235A (en) * | 2019-04-10 | 2019-07-09 | 苏州浪潮智能科技有限公司 | A kind of multivariate data classification method and device |
| CN110084318A (en) * | 2019-05-07 | 2019-08-02 | 哈尔滨理工大学 | A kind of image-recognizing method of combination convolutional neural networks and gradient boosted tree |
| CN110209811A (en) * | 2019-04-30 | 2019-09-06 | 上海艾瑞市场咨询股份有限公司 | A kind of commodity automatic classification method and system |
| CN110210522A (en) * | 2019-05-10 | 2019-09-06 | 无线生活(北京)信息技术有限公司 | The training method and device of picture quality Fraction Model |
| CN110414414A (en) * | 2019-07-25 | 2019-11-05 | 合肥工业大学 | SAR image Ship Target discrimination method based on the fusion of multi-layer depths of features |
-
2019
- 2019-12-30 CN CN201911398741.9A patent/CN111191723B/en active Active
Patent Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5793891A (en) * | 1994-07-07 | 1998-08-11 | Nippon Telegraph And Telephone Corporation | Adaptive training method for pattern recognition |
| US20110250626A1 (en) * | 2002-09-09 | 2011-10-13 | Reactive Surfaces, Ltd. | Visual Assays for Coatings Incorporating Bioactive Enzymes for Catalytic Functions |
| US20080114564A1 (en) * | 2004-11-25 | 2008-05-15 | Masayoshi Ihara | Information Classifying Device, Information Classifying Method, Information Classifying Program, Information Classifying System |
| CN103077399A (en) * | 2012-11-29 | 2013-05-01 | 西交利物浦大学 | Biological microscopic image classification method based on integrated cascade structure |
| CN104751163A (en) * | 2013-12-27 | 2015-07-01 | 同方威视技术股份有限公司 | Fluoroscopy examination system and method for carrying out automatic classification recognition on goods |
| CN106415594A (en) * | 2014-06-16 | 2017-02-15 | 北京市商汤科技开发有限公司 | A method and a system for face verification |
| CN104182621A (en) * | 2014-08-08 | 2014-12-03 | 同济大学 | DBN based ADHD discriminatory analysis method |
| CN107155360A (en) * | 2014-08-12 | 2017-09-12 | 西门子保健有限责任公司 | Multilayer polymeric for object detection |
| CN107798283A (en) * | 2016-08-31 | 2018-03-13 | 西安英诺视通信息技术有限公司 | A kind of neural network failure multi classifier based on the acyclic figure of decision-directed |
| CN106446895A (en) * | 2016-10-28 | 2017-02-22 | 安徽四创电子股份有限公司 | License plate recognition method based on deep convolutional neural network |
| US20180157938A1 (en) * | 2016-12-07 | 2018-06-07 | Samsung Electronics Co., Ltd. | Target detection method and apparatus |
| CN106874934A (en) * | 2017-01-12 | 2017-06-20 | 华南理工大学 | Sewage disposal method for diagnosing faults based on weighting extreme learning machine Integrated Algorithm |
| US20190122111A1 (en) * | 2017-10-24 | 2019-04-25 | Nec Laboratories America, Inc. | Adaptive Convolutional Neural Knowledge Graph Learning System Leveraging Entity Descriptions |
| CN108647706A (en) * | 2018-04-24 | 2018-10-12 | 广州大学 | Article identification classification based on machine vision and flaw detection method |
| CN108921198A (en) * | 2018-06-08 | 2018-11-30 | 山东师范大学 | commodity image classification method, server and system based on deep learning |
| CN109816028A (en) * | 2019-01-18 | 2019-05-28 | 创新奇智(南京)科技有限公司 | A kind of unbalanced dataset classification Model Fusion method of Partial Feature migration |
| CN109784421A (en) * | 2019-01-30 | 2019-05-21 | 北京朗镜科技有限责任公司 | A kind of construction method and device of identification model |
| CN109934261A (en) * | 2019-01-31 | 2019-06-25 | 中山大学 | A kind of Knowledge driving parameter transformation model and its few sample learning method |
| CN109993235A (en) * | 2019-04-10 | 2019-07-09 | 苏州浪潮智能科技有限公司 | A kind of multivariate data classification method and device |
| CN110209811A (en) * | 2019-04-30 | 2019-09-06 | 上海艾瑞市场咨询股份有限公司 | A kind of commodity automatic classification method and system |
| CN110084318A (en) * | 2019-05-07 | 2019-08-02 | 哈尔滨理工大学 | A kind of image-recognizing method of combination convolutional neural networks and gradient boosted tree |
| CN110210522A (en) * | 2019-05-10 | 2019-09-06 | 无线生活(北京)信息技术有限公司 | The training method and device of picture quality Fraction Model |
| CN110414414A (en) * | 2019-07-25 | 2019-11-05 | 合肥工业大学 | SAR image Ship Target discrimination method based on the fusion of multi-layer depths of features |
Non-Patent Citations (1)
| Title |
|---|
| 王忠民;王科;贺炎;: "高可信度加权的多分类器融合行为识别模型" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111191723B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086756B (en) | Text detection analysis method, device and equipment based on deep neural network | |
| CN106951825B (en) | Face image quality evaluation system and implementation method | |
| CN109492750B (en) | Zero sample image classification method based on convolutional neural network and factor space | |
| CN104484461B (en) | A kind of method and system classified based on encyclopaedia data to entity | |
| CN110717554B (en) | Image recognition method, electronic device, and storage medium | |
| CN103632168A (en) | Classifier integration method for machine learning | |
| CN110377727B (en) | Multi-label text classification method and device based on multi-task learning | |
| CN109492105B (en) | Text emotion classification method based on multi-feature ensemble learning | |
| US11335086B2 (en) | Methods and electronic devices for automated waste management | |
| CN106250925B (en) | A kind of zero Sample video classification method based on improved canonical correlation analysis | |
| CN115205521B (en) | Kitchen waste detection method based on neural network | |
| WO2022062419A1 (en) | Target re-identification method and system based on non-supervised pyramid similarity learning | |
| CN110532421A (en) | A kind of gender based on people, the music recommended method and system of mood and age | |
| CN115131580B (en) | Space target small sample identification method based on attention mechanism | |
| CN111967433A (en) | Action identification method based on self-supervision learning network | |
| CN110516622A (en) | A kind of gender of occupant, age and emotional intelligence recognition methods and system | |
| CN111914902A (en) | Traditional Chinese medicine identification and surface defect detection method based on deep neural network | |
| CN113033321A (en) | Training method of target pedestrian attribute identification model and pedestrian attribute identification method | |
| CN110413825B (en) | Street-clapping recommendation system oriented to fashion electronic commerce | |
| CN111401454A (en) | Few-sample target identification method based on transfer learning | |
| CN111723302A (en) | Recommendation method based on collaborative dual-model deep representation learning | |
| CN111191723A (en) | Few-sample commodity classification system and method based on cascade classifier | |
| CN116680578A (en) | Cross-modal model-based deep semantic understanding method | |
| CN115797637A (en) | Semi-supervised segmentation model based on uncertainty between models and in models | |
| CN111681748B (en) | Medical behavior action normalization evaluation method based on intelligent visual perception |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |