CN106250836B - Two benches facial image sorting technique and system under a kind of condition of small sample - Google Patents
Two benches facial image sorting technique and system under a kind of condition of small sample Download PDFInfo
- Publication number
- CN106250836B CN106250836B CN201610598570.4A CN201610598570A CN106250836B CN 106250836 B CN106250836 B CN 106250836B CN 201610598570 A CN201610598570 A CN 201610598570A CN 106250836 B CN106250836 B CN 106250836B
- Authority
- CN
- China
- Prior art keywords
- sample
- subset
- unmarked
- unlabeled
- new
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000001815 facial effect Effects 0.000 title abstract 4
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000012360 testing method Methods 0.000 claims description 46
- 230000004186 co-expression Effects 0.000 claims description 5
- 230000014509 gene expression Effects 0.000 abstract description 3
- 239000003550 marker Substances 0.000 abstract 2
- 238000003909 pattern recognition Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002195 synergetic effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses the two benches facial image sorting technique and system under a kind of condition of small sample, wherein method includes：Sample expands the stage, so that the unmarked sample of facial image is carried out collaboration expression to marker samples by semi-supervised mode, obtains collaboration and indicates coefficient；Obtain the unmarked sample corresponding to maximum collaboration expression coefficient；Maximum cooperate with is indicated that the unmarked sample corresponding to coefficient is added in label subset, marker samples are expanded, using the label subset after expansion as training sample；Simultaneously using remaining unmarked sample as new unmarked sample；Facial image sorting phase expands the new unmarked sample that the stage obtains to sample using the label subset after expansion based on collaboration presentation class device and classifies, and obtains final classification results.The present invention improves the accuracy of supervised classification method, while making full use of the judgement information of unmarked sample, and semi-supervised learning problem is converted to supervised learning problem using sample extended mode.
Description
Technical Field
The invention relates to the field of pattern recognition, in particular to a two-stage face image classification method and a two-stage face image classification system under a small sample condition.
Background
With the rapid development of pattern recognition and computer vision technology, human face recognition has gained much attention from researchers in various fields due to its wide application, and becomes an important aspect in the research of modern pattern recognition technology.
However, the face recognition is a small sample problem in practical application, and many conventional face recognition methods are based on the premise of a large number of training samples, so that in the case of extreme lack of labeled samples, a large number of supervised recognition methods are limited, and the recognition capability is weakened. Currently, face recognition methods mainly used include KNN (k nearest neighbor), LDA (linear discriminant analysis), SRC (sparse representation classifier), CRC (collaborative representation classifier), and the like. In euclidean space, KNN is a strategy for classification based on distance similarity. It can effectively maintain the structural relationship between local neighbors of the sample, but it is sensitive to noise and depends on the euclidean distance. LDA is used as a supervised classification method, marking information is utilized, a projection vector is selected to enable the same type of points to be as close as possible after projection, and different types of points are dispersed as far as possible after projection, so that the classification function is realized. However, LDA and KNN have the same disadvantage that they do not achieve robustness to noise.
In view of robustness to noise, many research efforts have emerged. The idea of SRC is to use training data to linearly and sparsely reconstruct a test sample, and obtain a classification result by comparing reconstructed residuals. Sparse methods are widely used in various fields due to their success in pattern recognition. But solve for l1However, it takes a lot of time, and thus the classification result cannot be obtained quickly. Recently, researchers propose a CRC method for classification, and the final classification result can be rapidly obtained. However, the classification effect of SRC and CRC depends on the number of training samples, and thus the condition of small samples cannot be satisfied in practical use.
Compared with labeled samples which are difficult to obtain, a large number of unlabeled samples are simple and easy to obtain, so that the semi-supervised method can be directly used for classifying in consideration of judgment information in the large number of unlabeled samples, but if a semi-supervised classification strategy is directly used, classification errors of the semi-supervised classification strategy are accumulated along with the progress of classification, and thus, the final classification result is greatly influenced.
Although there are many face classification methods, it is difficult to fully use the decision information of a large number of unlabeled samples and the semantic information of a small number of labeled samples for the problem of small sample classification when face data is actually classified. Therefore, classifying a large number of unlabeled samples under small sample conditions is a major and difficult point of current research.
Disclosure of Invention
The invention aims to solve the problem of classifying a large number of unlabelled samples under the condition of small samples, and provides a two-stage face image classification method and a two-stage face image classification system under the condition of small samples.
In order to achieve the purpose, the invention adopts the following technical scheme:
a two-stage face image classification method under a small sample condition comprises the following steps:
in the sample expansion stage, a face image unmarked sample carries out collaborative representation on a marked sample in a semi-supervised mode to obtain a collaborative representation coefficient; obtaining an unmarked sample corresponding to the maximum collaborative representation coefficient; adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, expanding the marked sample, and taking the expanded marked subset as a training sample; meanwhile, taking the residual unlabeled samples as new unlabeled samples and storing the new unlabeled samples in a new unlabeled subset;
and a face image classification stage, namely classifying the new unlabeled samples in the new unlabeled subsets obtained in the sample expansion stage by using the expanded labeled subsets based on the collaborative representation classifier, and obtaining a final classification result.
And in the sample expansion stage, adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, acquiring marks corresponding to the unmarked samples added into the marked subset, and if the unmarked samples correspond to a plurality of marks, storing the unmarked samples marked as a plurality of marks into a new unmarked subset.
The sample expansion phase comprises the following steps:
step (1): acquiring a face image data set, wherein the mark subset is ItrainUnmarked subset is UtestC is the number of sample categories;
step (2): respectively expanding the mark subsets of the ith class, wherein the value range of i is from 1 to C, and respectively expanding the jth mark sample of the ith classUsing unlabeled subset UtestAcquiring corresponding collaborative representation coefficients, and acquiring unmarked samples corresponding to the maximum collaborative representation coefficientsAnd the maximum collaborative representation coefficient is corresponding to the unmarked sampleAdding to the mark subset ItrainPerforming the following steps; at the same timeUnmarked subset UtestUnmarked sample corresponding to medium maximum collaborative representation coefficientSet to all 0 s;
and (3): performing iterative loop on the step (2) until the mark subset ItrainIs terminated when the number of marked samples reaches a set threshold.
In the step (3):
obtaining the classification label of the unmarked sample corresponding to the maximum collaborative representation coefficient in the iterative loop process, and judging the number of the labels, thereby determining whether the unmarked sample corresponding to the maximum collaborative representation coefficient is put into the mark subset ItrainIn (1).
The method for judging the number of the labels comprises the following steps:
if the unlabeled sample has multiple labels or is unlabeled, the unlabeled sample is labeled Inf, and the sample labeled Inf is placed in the unlabeled subset to obtain a new unlabeled subset Unew_test；
If the unlabeled sample has a tag, the unlabeled sample is added directly to the labeled subset ItrainIn (2), finally obtaining an expanded mark subset Inew_train。
The set threshold value of the step (3) is an unmarked subset UtestHalf the number of unlabeled samples.
The face image classification stage comprises the following steps:
extended tag subset Inew_trainAs a training sample, a new unlabeled subset Unew_testFor testing the samples, use Inew_trainAs training samples and using co-representation based classifiers on new unlabeled subsets Unew_testAnd (5) classifying to obtain a final classification result.
The collaborative representation-based classifier is as follows:
wherein, αiA presentation coefficient representing the i-th class,a co-expression coefficient representing the ith class,
next, identity (y) ═ argmin | e is usediCalculating a reconstructed residual error, and classifying according to the size of the reconstructed residual error;
wherein,
the calculation step of the collaborative representation coefficient in the step (2) is as follows:
wherein, UtestIn order to be an unlabeled subset of the,for the jth labeled sample of class i, αijAnd the co-expression coefficients corresponding to the jth mark sample of the ith class are represented, lambda is a regularization parameter used for controlling the calculation of the co-expression coefficients and preventing overfitting, and α is a representation coefficient.
A two-stage face image classification system under a small sample condition comprises:
the sample expansion module is configured to enable the unlabeled samples of the face image to cooperatively represent the labeled samples in a semi-supervised mode, and obtain a cooperative representation coefficient; obtaining an unmarked sample corresponding to the maximum collaborative representation coefficient; adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, expanding the marked sample, and taking the expanded marked subset as a training sample; meanwhile, taking the residual unlabeled samples as new unlabeled samples and storing the new unlabeled samples in a new unlabeled subset;
and the face image classification module is used for classifying the new unlabeled samples in the new unlabeled subsets obtained in the sample expansion stage by using the expanded labeled subsets based on the collaborative representation classifier, and obtaining a final classification result.
And the sample expansion module is used for adding the unmarked samples corresponding to the maximum collaborative representation coefficient into the marked subset, acquiring marks corresponding to the unmarked samples added into the marked subset, and if the unmarked samples correspond to a plurality of marks, storing the unmarked samples marked as a plurality of marks into a new unmarked subset.
A sample expansion module comprising:
a data acquisition unit: configured to acquire a face image dataset, wherein the subset of labels is ItrainUnmarked subset is UtestC is the number of sample categories;
tag subset expansion unit: configured to respectively expand the mark subsets of the ith class, wherein the value range of i is 1 to C, and the jth mark sample of the ith class is subjected toUsing unlabeled subset UtestAcquiring corresponding collaborative representation coefficients, and acquiring unmarked samples corresponding to the maximum collaborative representation coefficientsAnd the maximum collaborative representation coefficient is corresponding to the unmarked sampleAdding to the mark subset ItrainPerforming the following steps; while unmarked subset UtestMiddle maximumCo-representing unlabeled samples corresponding to coefficientsSet to all 0 s;
an iteration unit: configured to iteratively loop the work of the tag subset expansion unit until the tag subset ItrainIs terminated when the number of marked samples reaches a set threshold.
The iteration unit: configured to obtain the classification label of the unmarked sample corresponding to the maximum collaborative representation coefficient in the iterative loop process, and judge the number of the labels, thereby determining whether the unmarked sample corresponding to the maximum collaborative representation coefficient is put into the mark subset ItrainIn (1).
The judging of the number of the tags is as follows:
if the unlabeled sample has multiple labels or is unlabeled, the unlabeled sample is labeled Inf, and the sample labeled Inf is placed in the unlabeled subset to obtain a new unlabeled subset Unew_test；
If the unlabeled sample has a tag, the unlabeled sample is added directly to the labeled subset ItrainIn (2), finally obtaining an expanded mark subset Inew_train。
The set threshold of the iteration unit is an unmarked subset UtestHalf the number of unlabeled samples.
A face image classification module configured to:
extended tag subset Inew_trainAs a training sample, a new unlabeled subset Unew_testFor testing the samples, use Inew_trainAs training samples and using co-representation based classifiers on new unlabeled subsets Unew_testAnd (5) classifying to obtain a final classification result.
The invention has the beneficial effects that:
1. the method considers the relationship between the cooperative expression neighbors and fully utilizes the subspace structure of the data. And the cooperative sparse information and the subspace keeping information are fully utilized for pattern recognition.
2. The sample expansion is performed in a semi-supervised manner, taking into account the error accumulation of the pure semi-supervised method.
3. And the classification is carried out by using a supervision method, so that the error accumulation can be effectively slowed down, and the recognition effect of the supervision algorithm is far superior to that of an unsupervised algorithm and a semi-supervised algorithm, so that the recognition precision can be fully enhanced by using the supervision method finally.
Drawings
FIG. 1 is an overall flow chart of the present invention;
FIG. 2 is a system framework of the present invention.
Detailed Description
The invention is further described with reference to the following figures and examples.
As shown in fig. 1, includes:
obtaining an image data set I comprising a training image data subset ItrainAnd a test image data subset Utest；
Assume that the face image dataset I consists of two parts, I ═ X, label]Wherein X ═ X1，x2，…，xn]Each sample xi{ i ═ 1 … n } is a vector of dimensions p × 1, p being the dimension of the samples, and C being the number of samples.
The sample data set I can be divided into I according to requirementstrainAnd Utest. The invention uses 10 times cross validation, and the sample data set is randomly and uniformlyDividing into 10 parts, and taking one part at a time as a test sample set ItrainThe other nine parts are used as UtestThe experiment can be repeated 10 times.
Obtaining a sample set I according to the step (1)trainFirstly to ItrainOf a certain sample point xi{ i ═ 1L n }, all samples of a certain class of tags are taken.
Then according to step (2) for each marked sample of each classUsing UtestObtaining corresponding co-expression coefficient and adding the unmarked sample corresponding to the maximum coefficient valueThe specific method comprises the following steps:
wherein U istestIn order to be an unlabeled subset of the,for the jth labeled sample of class i, αijThe corresponding synergistic coefficient of the jth marked sample of the ith class; obtaining the unmarked sample corresponding to the maximum value of the synergistic coefficientAdding the sample to ItrainAnd is paired with UtestIn (1)Setting to all 0 s is performed;
according to steps (3) to ItrainUp to the original UtestEnding to obtain an extended mark sample set when the number of the extended mark samples is half of the number of the extended mark samples;
according toAnd (3) obtaining the classification label of the extended sample, wherein if a certain sample has a plurality of label marks Inf, otherwise, directly adding Itrain；
The image data in the face data set (YaleB) was verified, and the YaleB face database contained 38 persons, and there were 2414 samples of known data types, with 32x32 pixels. The experiment adopts 10 times of cross validation, all data are randomly and uniformly divided into 10 parts, one group is selected as test data each time, the rest data are selected as training data, the experiment is repeated for 10 times, the average value of 10 times is taken as the final identification accuracy, and the accuracy is shown in table 1.
The face image classification stage: according to the expansion mark subset I obtained in the step (3)new_trainAs a training sample, a new unlabeled subset Unew_testFor testing the samples, use Inew_trainUsing collaborative representation-based classifier pairs Unew_testAnd (5) classifying to obtain a final classification result. And acquiring the identification precision (%) under different samples.
As shown in fig. 2, a two-stage face image classification system under a small sample condition includes:
the sample expansion module is configured to enable the unlabeled samples of the face image to cooperatively represent the labeled samples in a semi-supervised mode, and obtain a cooperative representation coefficient; obtaining an unmarked sample corresponding to the maximum collaborative representation coefficient; adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, expanding the marked sample, and taking the expanded marked subset as a training sample; meanwhile, taking the residual unlabeled samples as new unlabeled samples and storing the new unlabeled samples in a new unlabeled subset;
and the face image classification module is used for classifying the new unlabeled samples in the new unlabeled subsets obtained in the sample expansion stage by using the expanded labeled subsets based on the collaborative representation classifier, and obtaining a final classification result.
And the sample expansion module is used for adding the unmarked samples corresponding to the maximum collaborative representation coefficient into the marked subset, acquiring marks corresponding to the unmarked samples added into the marked subset, and if the unmarked samples correspond to a plurality of marks, storing the unmarked samples marked as a plurality of marks into a new unmarked subset.
A sample expansion module comprising:
a data acquisition unit: configured to acquire a face image dataset, wherein the subset of labels is ItrainUnmarked subset is UtestC is the number of sample categories;
tag subset expansion unit: configured to respectively expand the mark subsets of the ith class, wherein the value range of i is 1 to C, and the jth mark sample of the ith class is subjected toUsing unlabeled subset UtestAcquiring corresponding collaborative representation coefficients, and acquiring unmarked samples corresponding to the maximum collaborative representation coefficientsAnd the maximum collaborative representation coefficient is corresponding to the unmarked sampleAdding to the mark subset ItrainPerforming the following steps; while unmarked subset UtestUnmarked sample corresponding to medium maximum collaborative representation coefficientSet to all 0 s;
an iteration unit: configured to iteratively loop the work of the tag subset expansion unit until the tag subset ItrainIs terminated when the number of marked samples reaches a set threshold.
Table 1 comparison of recognition accuracy of six classification methods on face data set YaleB
In order to test the performance of the classification method, the classifier of the invention is used on a plurality of data sets to classify the low-dimensional sample data after the dimensionality reduction by PCA, and the identification precision is obtained by comparing the low-dimensional sample data with the original class mark. Because a semi-supervised sample expansion mode is used, a large number of marked samples can be obtained, and in order to inhibit accumulated errors generated by the semi-supervised classification mode, the method is converted into a supervised mode for classification in the last step, so that a better classification result can be obtained.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (10)
1. A two-stage face image classification method under a small sample condition is characterized by comprising the following steps:
in the sample expansion stage, a face image unmarked sample carries out collaborative representation on a marked sample in a semi-supervised mode to obtain a collaborative representation coefficient; obtaining an unmarked sample corresponding to the maximum collaborative representation coefficient; adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, expanding the marked sample, and taking the expanded marked subset as a training sample; meanwhile, taking the residual unlabeled samples as new unlabeled samples and storing the new unlabeled samples in a new unlabeled subset;
and a face image classification stage, namely classifying the new unlabeled samples in the new unlabeled subsets obtained in the sample expansion stage by using the expanded labeled subsets based on the collaborative representation classifier, and obtaining a final classification result.
2. The method according to claim 1, wherein in the sample expansion stage, the unmarked sample corresponding to the maximum collaborative representation coefficient is added into the marked subset, the mark corresponding to the unmarked sample added into the marked subset is obtained, and if the unmarked sample corresponds to a plurality of marks, the unmarked sample marked as a plurality of marks is saved into a new unmarked subset.
3. The method of claim 1, wherein the sample expansion phase comprises the steps of:
step (1): acquiring a face image data set, wherein the mark subset is ItrainUnmarked subset is UtestC is the number of sample categories;
step (2): respectively expanding the mark subsets of the ith class, wherein the value range of i is from 1 to C, and respectively expanding the jth mark sample of the ith classUsing unlabeled subset UtestAcquiring corresponding collaborative representation coefficients, and acquiring unmarked samples corresponding to the maximum collaborative representation coefficientsAnd the maximum collaborative representation coefficient is corresponding to the unmarked sampleAdding to the mark subset ItrainPerforming the following steps; while unmarked subset UtestUnmarked sample corresponding to medium maximum collaborative representation coefficientSet to all 0 s;
and (3): performing iterative loop on the step (2) until the mark subset ItrainIs terminated when the number of marked samples reaches a set threshold.
4. The method as claimed in claim 3, wherein in the step (3):
obtaining the classification label of the unmarked sample corresponding to the maximum collaborative representation coefficient in the iterative loop process, and judging the number of the labels, thereby determining whether the unmarked sample corresponding to the maximum collaborative representation coefficient is put into the mark subset ItrainIn (1).
5. The method of claim 4, wherein said determining the number of tags comprises:
if the unlabeled sample has multiple labels or is unlabeled, the unlabeled sample is labeled Inf, and the sample labeled Inf is placed in the unlabeled subset to obtain a new unlabeled subset Unew_test；
If the unlabeled sample has a tag, the unlabeled sample is added directly to the labeled subset ItrainIn (2), finally obtaining an expanded mark subset Inew_train。
6. The method as claimed in claim 1, wherein the face image classification stage comprises the steps of:
extended tag subset Inew_trainAs a training sample, a new unlabeled subset Unew_testFor testing the samples, use Inew_trainAs training samples and using co-representation based classifiers on new unlabeled subsets Unew_testAnd (5) classifying to obtain a final classification result.
7. A two-stage face image classification system under the condition of small samples is characterized by comprising the following steps:
the sample expansion module is configured to enable the unlabeled samples of the face image to cooperatively represent the labeled samples in a semi-supervised mode, and obtain a cooperative representation coefficient; obtaining an unmarked sample corresponding to the maximum collaborative representation coefficient; adding the unmarked sample corresponding to the maximum collaborative representation coefficient into the marked subset, expanding the marked sample, and taking the expanded marked subset as a training sample; meanwhile, taking the residual unlabeled samples as new unlabeled samples and storing the new unlabeled samples in a new unlabeled subset;
and the face image classification module is used for classifying the new unlabeled samples in the new unlabeled subsets obtained in the sample expansion stage by using the expanded labeled subsets based on the collaborative representation classifier, and obtaining a final classification result.
8. The system of claim 7, wherein,
and the sample expansion module is used for adding the unmarked samples corresponding to the maximum collaborative representation coefficient into the marked subset, acquiring marks corresponding to the unmarked samples added into the marked subset, and if the unmarked samples correspond to a plurality of marks, storing the unmarked samples marked as a plurality of marks into a new unmarked subset.
9. The system of claim 7, wherein the sample expansion module comprises:
a data acquisition unit: configured to acquire a face image dataset, wherein the subset of labels is ItrainUnmarked subset is UtestC is the number of sample categories;
tag subset expansion unit: configured to respectively expand the mark subsets of the ith class, wherein the value range of i is 1 to C, and the jth mark sample of the ith class is subjected toUsing unlabeled subset UtestObtaining corresponding co-expression coefficients, obtainingTaking the unmarked sample corresponding to the maximum collaborative representation coefficientAnd the maximum collaborative representation coefficient is corresponding to the unmarked sampleAdding to the mark subset ItrainPerforming the following steps; while unmarked subset UtestUnmarked sample corresponding to medium maximum collaborative representation coefficientSet to all 0 s;
an iteration unit: configured to iteratively loop the work of the tag subset expansion unit until the tag subset ItrainIs terminated when the number of marked samples reaches a set threshold.
10. The system of claim 9, wherein,
the iteration unit: configured to obtain the classification label of the unmarked sample corresponding to the maximum collaborative representation coefficient in the iterative loop process, and judge the number of the labels, thereby determining whether the unmarked sample corresponding to the maximum collaborative representation coefficient is put into the mark subset ItrainPerforming the following steps;
the judging of the number of the tags is as follows:
if the unlabeled sample has multiple labels or is unlabeled, the unlabeled sample is labeled Inf, and the sample labeled Inf is placed in the unlabeled subset to obtain a new unlabeled subset Unew_test；
If the unlabeled sample has a tag, the unlabeled sample is added directly to the labeled subset ItrainIn (2), finally obtaining an expanded mark subset Inew_train。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610598570.4A CN106250836B (en) | 2016-07-26 | 2016-07-26 | Two benches facial image sorting technique and system under a kind of condition of small sample |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610598570.4A CN106250836B (en) | 2016-07-26 | 2016-07-26 | Two benches facial image sorting technique and system under a kind of condition of small sample |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106250836A CN106250836A (en) | 2016-12-21 |
| CN106250836B true CN106250836B (en) | 2018-09-14 |
Family
ID=57604775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610598570.4A Expired - Fee Related CN106250836B (en) | 2016-07-26 | 2016-07-26 | Two benches facial image sorting technique and system under a kind of condition of small sample |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106250836B (en) |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10339464B2 (en) * | 2012-06-21 | 2019-07-02 | Philip Morris Products S.A. | Systems and methods for generating biomarker signatures with integrated bias correction and class prediction |
| CN104598917B (en) * | 2014-12-08 | 2018-04-27 | 上海大学 | A kind of support vector machine classifier IP kernel |
| CN105046673B (en) * | 2015-07-13 | 2017-11-03 | 哈尔滨工业大学 | High spectrum image and visual image fusion sorting technique based on self study |
| CN105303198B (en) * | 2015-11-17 | 2018-08-17 | 福州大学 | A kind of remote sensing image semisupervised classification method learnt from fixed step size |
| CN105740908B (en) * | 2016-01-31 | 2017-05-24 | 中国石油大学(华东) | Classifier design method based on kernel space self-explanatory sparse representation |
-
2016
- 2016-07-26 CN CN201610598570.4A patent/CN106250836B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN106250836A (en) | 2016-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110414462B (en) | Unsupervised cross-domain pedestrian re-identification method and system | |
| Yang et al. | Retracted: Face recognition attendance system based on real-time video processing | |
| CN103605972B (en) | Non-restricted environment face verification method based on block depth neural network | |
| Almazan et al. | Handwritten word spotting with corrected attributes | |
| CN109376796A (en) | Image classification method based on active semi-supervised learning | |
| Alnajar et al. | Learning-based encoding with soft assignment for age estimation under unconstrained imaging conditions | |
| CN103400144B (en) | Active learning method based on K-neighbor for support vector machine (SVM) | |
| CN107943856A (en) | A kind of file classification method and system based on expansion marker samples | |
| Escalera et al. | Boosted Landmarks of Contextual Descriptors and Forest-ECOC: A novel framework to detect and classify objects in cluttered scenes | |
| CN113887661B (en) | Image set classification method and system based on representation learning reconstruction residual analysis | |
| Zweig et al. | Hierarchical regularization cascade for joint learning | |
| US20150131899A1 (en) | Devices, systems, and methods for learning a discriminant image representation | |
| Zhang et al. | Automatic discrimination of text and non-text natural images | |
| CN105930792A (en) | Human action classification method based on video local feature dictionary | |
| CN110414587A (en) | Depth convolutional neural networks training method and system based on progressive learning | |
| CN103310208A (en) | Identifiability face pose recognition method based on local geometrical visual phrase description | |
| Duin et al. | Non-euclidean dissimilarities: causes, embedding and informativeness | |
| CN112257600B (en) | Face recognition method and system | |
| Singhal et al. | Image classification using bag of visual words model with FAST and FREAK | |
| Ma et al. | Shiss: Supervised hashing with informative set selection | |
| CN106250836B (en) | Two benches facial image sorting technique and system under a kind of condition of small sample | |
| CN107085700A (en) | A kind of face identification method being combined based on rarefaction representation with neural networks with single hidden layer technology | |
| Tian et al. | Self-supervised learning based on discriminative nonlinear features for image classification | |
| Beltrán et al. | Semi-supervised dimensionality reduction via multimodal matrix factorization | |
| Wu et al. | Towards self-exploring discriminating features for visual learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180914 |