CN103996018A - Human-face identification method based on 4DLBP - Google Patents
Human-face identification method based on 4DLBP Download PDFInfo
- Publication number
- CN103996018A CN103996018A CN201410072699.2A CN201410072699A CN103996018A CN 103996018 A CN103996018 A CN 103996018A CN 201410072699 A CN201410072699 A CN 201410072699A CN 103996018 A CN103996018 A CN 103996018A
- Authority
- CN
- China
- Prior art keywords
- 4dlbp
- image
- feature
- face
- pixel
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a human-face identification method based on 4DLBP (4-Dimnesional Local Binary Patterns). The human-face identification method based on 4DLBP comprises the following steps including image preprocessing, 4DLBP characteristic extraction and ELM (Extreme Learning Machine) training classification, wherein the ELM training classification comprises two stages: a training image set stage and a test image set stage, human-face images are respectively subjected to image preprocessing in each stage, subblock sub images of the images are subjected to 4DLBP characteristic extraction, characteristics obtained from each block of sub images are connected in series for obtaining the 4DLBP characteristics of the whole human face, then, the extracted human-face 4DLBP characteristics are utilized as the input of an ELM classifier for carrying on training and testing, finally, classes of the images are obtained, and the identifying is completed. The human-face identification method has the advantages that center point pixel values are added on the basis of 3DLBP (3-Dimensional Local Binary Patterns) for forming the 4DLBP characteristics, the human-face image local characteristics can be effectively extracted for improving the human-face identification performance.
Description
Technical field
The invention belongs to living things feature recognition field, particularly a kind of face identification method based on 4DLBP.
Background technology
Face recognition technology is an important topic in biological identification technology, is current very active research direction.Carry out compared with identification with utilizing other biological feature, recognition of face has directly, facilitates, friendly, non-offensive advantage, thereby has application prospect extremely widely.But recognition of face also exists some difficult problems in the process of development.First face is the irregular surface of a three-dimensional non-rigid body; Secondly, face can change along with the variation of age, health and expression; Again, in the time gathering facial image, different illumination, angle all can affect recognition of face ground accuracy.Because human brain is still unknowable to the mechanism of recognition of face, machine recognition of face is also in groping and the stage of innovating, and relates to many-sided many knowledge such as computer vision, pattern-recognition, physiology and psychology.The recognition of face that all of these factors taken together is all becomes and has challenge, but the valuable problem of tool very.
Local binary (Local Binary Patterns, LBP), fixes and extracts the defect that range scale is little and propose in order to solve yardstick.But LBP operator, using feature extraction as a single process, has not made full use of the relation between sample.LBP operator has been described the relation between center pixel value and field pixel number thereof, and 3DLBP(3Dimensions Local Binary Patterns) operator has been specialized this relation more, but the common ground of the two is all to have ignored center pixel value.
Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art part, a kind of stronger face identification method based on 4DLBP of distinguishing ability that utilizes image processing techniques and intellectual technology to solve recognition of face problem is provided.
The technical solution adopted for the present invention to solve the technical problems is:
Advantage of the present invention and good effect are:
The present invention has added that on the basis of 3DLBP central point pixel value has formed 4DLBP feature, can effectively extract facial image local feature to improve the performance of recognition of face.
The present invention adopts mean variance normalization method to carry out pre-service to image to image respectively in training plan image set stage and test pattern image set stage, effectively eliminates the impact of illumination variation on facial image.
It is that the little image of each piecemeal is carried out to 4DLBP feature extraction that the present invention extracts 4DLBP feature, obtains 16 4LBP proper vectors, then these 4DLBP proper vectors are together in series, and obtains the 4DLBP feature of view picture facial image, makes eigenwert more accurate.
Brief description of the drawings
Fig. 1 is the FB(flow block) of carrying out recognition of face by the present invention;
Fig. 2 is the forward and backward image comparison figure of image pre-service in the present invention;
Fig. 3 is the schematic diagram of the 4DLBP feature extraction in the present invention;
Fig. 4 be in the 4DLBP characteristic extraction procedure in the present invention each 4DLBP histogram extract schematic diagram.
Embodiment
Below by specific embodiment, the invention will be further described, and following examples are descriptive, is not determinate, can not limit protection scope of the present invention with this.
A kind of face identification method based on 4DLBP, see Fig. 1, comprise the steps: image pre-service, 4DLBP feature extraction, ELM trains classification, wherein ELM(Extrem Learning Machine, extreme learning machine device) train classification to comprise two stages: training plan image set stage and test pattern image set stage, each stage is carried out pre-service to facial image respectively, the piecemeal subimage of image is extracted to its 4DLBP feature, the feature that every subimage is obtained is together in series and obtains the 4DLBP feature of view picture face, then the face 4DLBP feature that utilization is extracted is as the input of ELM sorter, carry out training and testing, finally obtain the classification of image, complete thus mark.
Its concrete steps are:
(1) pre-service
In order to eliminate the impact of illumination variation on facial image, adopt mean variance normalization method to carry out pre-service to image herein.This process is divided two steps: greyscale transformation and gray scale stretch.If the pixel value of certain pixel is f (x, y) in image, in a so whole image, the mean value of all pixel gray-scale values is
Standard deviation is:
Greyscale transformation can be expressed as:
f(x,y)=(f(x,y)-aver)/σ
Recalculate the gray-scale value of each pixel according to formula.Pixels all in image is carried out after greyscale transformation, just there is variation in the scope of pixel grayscale, in 0~255 grey level range of not expecting at us, so also need to carry out gray scale stretching.
An image is carried out after greyscale transformation, and the maximal value in all grey scale pixel values of our mark is max, and minimum value is min,, gray scale stretching formula can be expressed as:
f(x,y)=(f(x,y)-min)*255/(max-min)
Image is carried out, after pre-service, can finding out compared with original image, and it is very even that the illumination of image becomes, and sees Fig. 2, and shade has obtained good removal.
(2) 4DLBP feature extraction
Calculate the LBP value of each pixel in facial image with LBP operator, obtain the response image of a LBP, the then histogram of calculated response image, the histogram of this response image is exactly the feature of the facial image of LBP extraction.The histogram of this response image is exactly the feature of the facial image of LBP extraction.As the first-order statistical properties of image, histogram has been described the various small pattern in entire image, such as the frequency of occurrences at bright spot, dim spot, smooth region, edge etc., but structural information that cannot Description Image.And the local feature of image regional differs greatly, if one whole image only generates a LBP histogram, can cause local different information to lose.In order to address this problem, first, original facial image to be carried out to piecemeal, and then calculate the LBP operator histogram of every.Method (seeing Fig. 3~4) is herein that the little image of each piecemeal is carried out to 4DLBP feature extraction, obtains 16 4DLBP proper vectors, then these 4DLBP proper vectors are together in series, and obtains the 4DLBP feature of view picture facial image.The concrete steps of calculating 4DLBP feature are as follows:
Step1: calculate LBP feature.
LBP feature is for the relation of the each pixel value of Description Image and its neighborhood territory pixel value.In gray level image, the gray scale that pixel value is pixel.LBP operator by neighborhood territory pixel value (f=l, 2 ..., 8) compare with center pixel value, carry out thresholding processing.By b
i(i=l, 2 ..., 8) arrange and obtain 8 bits in the direction of the clock, then be translated into decimal number, the mark using this value as center pixel, this mark is exactly the result of LBP operator after to center pixel effect.LBP can represent with the neighborhood of different size, adopts circular neighborhood in conjunction with bilinear interpolation, and the radius of neighborhood and number of pixels can be chosen arbitrarily.
Step2: calculate 3DLBP feature.
According to statistics, for the face gray level image (normalizing to 0-255) after normalization, gray scale difference value between 2 within pixel has 93% to be less than 7, thereby available triad number is encoded to the gray scale difference value of gray level image each point and neighborhood thereof.The absolute value of triad number ({ i2i3i4}) corresponding grey scale difference (depth difference, DD) is 0~7. all | DD| >=7 distribute digital 7.Represent the symbol (corresponding with LBP) of gray scale difference value with binary number i1, so just with tetrad { i
1i
2i
3i
4dD can be expressed as:
|DD|=2
2i
2+2i
3+i
4
4 binary numbers are divided into 4 layers, and the binary cell of every layer is by arranged clockwise.Finally, can obtain 48 bits, 4 decimal number IV 1, IV 2, IV 3 and the IV 4 of its correspondence, as the expression of pixel, are referred to as 3DLBP.Structure 4 map images corresponding with IV 1, IV 2, IV 3, IV 4 represent original gray-scale map, are called 3DLBPMapl (corresponding LBP), 3DLBPMap2,3DLBPMap3 and 3DLBPMap4.Similar with LBP, for retaining the spatial information of face, face gray level image is carried out to piecemeal, as shown in Figure 3.To the every local 3DLBP of structure, the histogram of more local 3DLBP is connected into a proper vector.
Step3: calculate 4DLBP feature.
3DLBP has only considered the big or small partial structurtes relation of central pixel point and field pixel, has well described the Local textural feature of image.But 3DLBP has but ignored the pixel value of the central point of image own for the impact of local detail feature.4DLBP feature is exactly the pixel value that adds central point on the basis of 3DLBP feature, forms new 4DLBP feature.As shown in Figure 3.
(3) training of the ELM based on variation inference pattern classification
The ELM sorter that the feature that facial image extracts by 4DLBP is input to based on variation inference pattern is classified, and identifies every face and belongs to which classification.Wherein, the hidden layer node number of ELM sorter is determined by variation inference pattern.Whole process comprises two stages: training stage and test phase.ELM training plan image set Stage Summary is as follows:
Step 1: training plan image set: Data_Train=[M
face1, M
face2..., M
faceN], N is all training sample numbers.
Step 2: stray parameter is set, input weight vectors A={a
1, a
2..., a
n, bias vector B={b
1, b
2..., b
n.
Step 3:
Step 4:H β=T, β={ β
1, β
2..., β
l.
Step 5: calculate β=H
+y, Y={y
1, y
2..., y
n, Y is the mark of facial image.
Step 6: calculate T, obtain the classification of training objective.
Three steps in test phase are general and training process is similar, and we obtain parameter by test model from training pattern, and then by test pattern, we just can obtain actual output.
ELM test pattern image set Stage Summary is as follows:
Step 1: obtain optimized parameter A' and B', β ' (result of training stage) and count N based on the definite hidden layer node of variation inference pattern.
Step 2: test set: Data_Test=[M
face1, M
face2..., M
facem], m is all test sample book numbers.
Step 3:
Step 4: calculate T', obtain the classification of test target.
Claims (7)
1. the face identification method based on 4DLBP, it is characterized in that: comprise the steps: image pre-service, 4DLBP feature extraction, ELM trains classification, wherein ELM training classification comprises two stages: training plan image set stage and test pattern image set stage, each stage is carried out image pre-service to facial image respectively, the piecemeal subimage of image is extracted to its 4DLBP feature, the feature that every subimage is obtained is together in series and obtains the 4DLBP feature of view picture face, then the face 4DLBP feature that utilization is extracted is as the input of ELM sorter, carry out training and testing, finally obtain the classification of image, complete mark.
2. the face identification method based on 4DLBP according to claim 1, it is characterized in that: described image pre-service adopts mean variance normalization method to carry out pre-service to image, this process is divided two steps: greyscale transformation and gray scale stretch, if the pixel value of certain pixel is f (x in image, y), in a so whole image, the mean value of all pixel gray-scale values is:
n is the total quantity of pixel
Standard deviation is:
Greyscale transformation is expressed as:
f(x,y)=(f(x,y)-aver)/σ
An image is carried out after greyscale transformation, and the maximal value in all grey scale pixel values of mark is max, and minimum value is min,, gray scale stretching formulae express is:
f(x,y)=(f(x,y)-min)*255/(max-min)。
3. the face identification method based on 4DLBP according to claim 1, it is characterized in that: described 4DLBP feature extraction is that the little image of each piecemeal is carried out to 4DLBP feature extraction, obtain 16 4LBP proper vectors, again these 4DLBP proper vectors are together in series, obtain the 4DLBP feature of view picture facial image.
4. the face identification method based on 4DLBP according to claim 1, it is characterized in that: facial image extracts by 4DLBP the ELM sorter that feature is input to based on variation inference pattern and classifies, identify the affiliated corresponding classification of every face, the hidden layer node number of described ELM sorter is determined by variation inference pattern.
5. the face identification method based on 4DLBP according to claim 1, is characterized in that: the ELM training stage step in described training plan image set stage is:
(1) training plan image set: Data_Train=[M
face1, M
face2..., M
faceN], N is all training sample numbers;
(2) stray parameter is set: input weight vectors A={a
1, a
2..., a
n, bias vector B={b
1, b
2..., b
n;
⑶
⑷Hβ=T,β={β
1,β
2,...,β
L};
(5) calculate β=H
+y, Y={y
1, y
2..., y
n, Y is the mark of facial image;
(6) calculate T, obtain the classification of training objective.
6. the face identification method based on 4DLBP according to claim 5, is characterized in that: the ELM test phase step in described training plan image set stage is:
(1) obtain optimized parameter A' and B', β ' (result of training stage) and count N based on the definite hidden layer node of variation inference pattern;
(2) test pattern image set: Data_Test=[M
face1, M
face2..., M
facem], m is all test sample book numbers;
⑶
(4) calculate T', obtain the classification of test target.
7. the face identification method based on 4DLBP according to claim 3, is characterized in that: described 4DLBP characteristic extraction step is:
(1) calculate LBP feature: LBP feature is used for the relation of the each pixel value of Description Image and its neighborhood territory pixel value, in gray level image, the gray scale that pixel value is pixel, LBP operator by neighborhood territory pixel value (f=l, 2 ... 8) compare with center pixel value, carry out thresholding processing.By b
i(i=l, 2,8) arrange and obtain 8 bits in the direction of the clock, then be translated into decimal number, the mark using this value as center pixel, this mark is exactly the result of LBP operator after to center pixel effect, LBP can represent with the neighborhood of different size, adopts circular neighborhood in conjunction with bilinear interpolation, and the radius of neighborhood and number of pixels can be chosen arbitrarily;
(2) calculate 3DLBP feature:
According to statistics, for the face gray level image after normalization, normalize to 0-255, gray scale difference value between 2 within pixel has 93% to be less than 7, thereby available triad number is encoded to the gray scale difference value of gray level image each point and neighborhood thereof, the absolute value of triad number ({ i2i3i4}) corresponding grey scale difference (depth difference, DD) is 0~7. all | DD|>=7 distribute digital 7, use binary number i
1the symbol that represents gray scale difference value, corresponding with LBP, so just with tetrad { i
1i
2i
3i
4dD can be expressed as:
|DD|=2
2i
2+2i
3+i
4
4 binary numbers are divided into 4 layers, the binary cell of every layer is by arranged clockwise, finally, can obtain 48 bits, 4 decimal number IV 1, IV 2, IV 3 and the IV 4 of its correspondence is as the expression of pixel, be referred to as 3DLBP, structure 4 map images corresponding with IV 1, IV 2, IV 3, IV 4 represent original gray-scale map, be called 3DLBPMapl, 3DLBPMap2,3DLBPMap3 and 3DLBPMap4, face gray level image is carried out to piecemeal, to the every local 3DLBP of structure, the histogram of more local 3DLBP is connected into a proper vector;
(3) calculate 4DLBP feature:
On the basis of 3DLBP feature, add the pixel value of central point, form new 4DLBP feature.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410072699.2A CN103996018B (en) | 2014-03-03 | 2014-03-03 | Face identification method based on 4DLBP |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410072699.2A CN103996018B (en) | 2014-03-03 | 2014-03-03 | Face identification method based on 4DLBP |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103996018A true CN103996018A (en) | 2014-08-20 |
CN103996018B CN103996018B (en) | 2018-03-02 |
Family
ID=51310179
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410072699.2A Expired - Fee Related CN103996018B (en) | 2014-03-03 | 2014-03-03 | Face identification method based on 4DLBP |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103996018B (en) |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104657718A (en) * | 2015-02-13 | 2015-05-27 | 武汉工程大学 | Face recognition method based on face image feature extreme learning machine |
CN104751141A (en) * | 2015-03-30 | 2015-07-01 | 东南大学 | ELM gesture recognition algorithm based on feature image full pixel gray values |
CN104778705A (en) * | 2015-04-17 | 2015-07-15 | 许书华 | Bilinear interpolation-based color image edge detection method |
CN106127674A (en) * | 2016-06-15 | 2016-11-16 | 公安部第三研究所 | Image white balance method based on LBP |
CN106156857A (en) * | 2015-03-31 | 2016-11-23 | 日本电气株式会社 | The method and apparatus selected for mixed model |
CN106326843A (en) * | 2016-08-15 | 2017-01-11 | 武汉工程大学 | Face recognition method |
CN107067029A (en) * | 2017-03-20 | 2017-08-18 | 新智认知数据服务有限公司 | The image classification method that a kind of ELM and DE based on multi-channel feature are combined |
CN108062747A (en) * | 2017-12-28 | 2018-05-22 | 江西师范大学 | The processing method and system of a kind of remote sensing image |
CN108460325A (en) * | 2018-01-10 | 2018-08-28 | 西安理工大学 | A kind of crowd's demographic method of the two-way fusion based on ELM |
CN108986124A (en) * | 2018-06-20 | 2018-12-11 | 天津大学 | In conjunction with Analysis On Multi-scale Features convolutional neural networks retinal vascular images dividing method |
CN109344758A (en) * | 2018-09-25 | 2019-02-15 | 厦门大学 | Based on the face identification method for improving local binary patterns |
CN109447978A (en) * | 2018-11-09 | 2019-03-08 | 河北工业大学 | A kind of photovoltaic solar cell piece electroluminescent image defect classification method |
CN109670412A (en) * | 2018-11-30 | 2019-04-23 | 天津大学 | Improve the 3D face identification method of LBP |
CN110363083A (en) * | 2019-06-06 | 2019-10-22 | 广智微芯(扬州)有限公司 | A kind of facial image gender identification method and device |
CN110781800A (en) * | 2019-10-23 | 2020-02-11 | 北京远舢智能科技有限公司 | Image recognition system |
CN113642503A (en) * | 2021-08-23 | 2021-11-12 | 国网山东省电力公司金乡县供电公司 | Window service scoring method and system based on image and voice recognition |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105913024B (en) * | 2016-04-12 | 2019-05-24 | 上海交通大学 | The Android mobile phone endpoint detection methods of resistance Replay Attack based on LAP operator |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080166026A1 (en) * | 2007-01-10 | 2008-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for generating face descriptor using extended local binary patterns, and method and apparatus for face recognition using extended local binary patterns |
CN101706964A (en) * | 2009-08-27 | 2010-05-12 | 北京交通大学 | Color constancy calculating method and system based on derivative structure of image |
CN102024141A (en) * | 2010-06-29 | 2011-04-20 | 上海大学 | Face recognition method based on Gabor wavelet transform and local binary pattern (LBP) optimization |
CN102156887A (en) * | 2011-03-28 | 2011-08-17 | 湖南创合制造有限公司 | Human face recognition method based on local feature learning |
WO2011119117A1 (en) * | 2010-03-26 | 2011-09-29 | Agency For Science, Technology And Research | Facial gender recognition |
US8594432B2 (en) * | 2009-12-18 | 2013-11-26 | Canon Kabushiki Kaisha | Image processing apparatus and image processing method |
CN103455805A (en) * | 2013-09-27 | 2013-12-18 | 厦门大学 | Novel method for describing facial features |
-
2014
- 2014-03-03 CN CN201410072699.2A patent/CN103996018B/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080166026A1 (en) * | 2007-01-10 | 2008-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for generating face descriptor using extended local binary patterns, and method and apparatus for face recognition using extended local binary patterns |
CN101706964A (en) * | 2009-08-27 | 2010-05-12 | 北京交通大学 | Color constancy calculating method and system based on derivative structure of image |
US8594432B2 (en) * | 2009-12-18 | 2013-11-26 | Canon Kabushiki Kaisha | Image processing apparatus and image processing method |
WO2011119117A1 (en) * | 2010-03-26 | 2011-09-29 | Agency For Science, Technology And Research | Facial gender recognition |
CN102024141A (en) * | 2010-06-29 | 2011-04-20 | 上海大学 | Face recognition method based on Gabor wavelet transform and local binary pattern (LBP) optimization |
CN102156887A (en) * | 2011-03-28 | 2011-08-17 | 湖南创合制造有限公司 | Human face recognition method based on local feature learning |
CN103455805A (en) * | 2013-09-27 | 2013-12-18 | 厦门大学 | Novel method for describing facial features |
Non-Patent Citations (2)
Title |
---|
付晓峰: "基于二元模式的人脸识别与表情识别研究", 《中国博士学位论文全文数据库》 * |
袁理等: "基于自适应3DLBP特征的人脸深度图像识别", 《电视技术》 * |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104657718A (en) * | 2015-02-13 | 2015-05-27 | 武汉工程大学 | Face recognition method based on face image feature extreme learning machine |
CN104657718B (en) * | 2015-02-13 | 2018-12-14 | 武汉工程大学 | A kind of face identification method based on facial image feature extreme learning machine |
CN104751141A (en) * | 2015-03-30 | 2015-07-01 | 东南大学 | ELM gesture recognition algorithm based on feature image full pixel gray values |
CN106156857B (en) * | 2015-03-31 | 2019-06-28 | 日本电气株式会社 | The method and apparatus of the data initialization of variation reasoning |
CN106156857A (en) * | 2015-03-31 | 2016-11-23 | 日本电气株式会社 | The method and apparatus selected for mixed model |
CN104778705A (en) * | 2015-04-17 | 2015-07-15 | 许书华 | Bilinear interpolation-based color image edge detection method |
CN106127674A (en) * | 2016-06-15 | 2016-11-16 | 公安部第三研究所 | Image white balance method based on LBP |
CN106326843A (en) * | 2016-08-15 | 2017-01-11 | 武汉工程大学 | Face recognition method |
CN106326843B (en) * | 2016-08-15 | 2019-08-16 | 武汉工程大学 | A kind of face identification method |
CN107067029A (en) * | 2017-03-20 | 2017-08-18 | 新智认知数据服务有限公司 | The image classification method that a kind of ELM and DE based on multi-channel feature are combined |
CN108062747A (en) * | 2017-12-28 | 2018-05-22 | 江西师范大学 | The processing method and system of a kind of remote sensing image |
CN108460325B (en) * | 2018-01-10 | 2021-07-20 | 西安理工大学 | Double-path fusion crowd counting method based on ELM |
CN108460325A (en) * | 2018-01-10 | 2018-08-28 | 西安理工大学 | A kind of crowd's demographic method of the two-way fusion based on ELM |
CN108986124A (en) * | 2018-06-20 | 2018-12-11 | 天津大学 | In conjunction with Analysis On Multi-scale Features convolutional neural networks retinal vascular images dividing method |
CN109344758A (en) * | 2018-09-25 | 2019-02-15 | 厦门大学 | Based on the face identification method for improving local binary patterns |
CN109447978A (en) * | 2018-11-09 | 2019-03-08 | 河北工业大学 | A kind of photovoltaic solar cell piece electroluminescent image defect classification method |
CN109447978B (en) * | 2018-11-09 | 2021-11-02 | 河北工业大学 | Method for classifying defects of electroluminescence image of photovoltaic solar cell |
CN109670412A (en) * | 2018-11-30 | 2019-04-23 | 天津大学 | Improve the 3D face identification method of LBP |
CN109670412B (en) * | 2018-11-30 | 2023-04-28 | 天津大学 | 3D face recognition method for improving LBP |
CN110363083A (en) * | 2019-06-06 | 2019-10-22 | 广智微芯(扬州)有限公司 | A kind of facial image gender identification method and device |
CN110781800A (en) * | 2019-10-23 | 2020-02-11 | 北京远舢智能科技有限公司 | Image recognition system |
CN110781800B (en) * | 2019-10-23 | 2022-04-12 | 北京远舢智能科技有限公司 | Image recognition system |
CN113642503A (en) * | 2021-08-23 | 2021-11-12 | 国网山东省电力公司金乡县供电公司 | Window service scoring method and system based on image and voice recognition |
CN113642503B (en) * | 2021-08-23 | 2024-03-15 | 国网山东省电力公司金乡县供电公司 | Window service scoring method and system based on image and voice recognition |
Also Published As
Publication number | Publication date |
---|---|
CN103996018B (en) | 2018-03-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103996018A (en) | Human-face identification method based on 4DLBP | |
CN107506761B (en) | Brain image segmentation method and system based on significance learning convolutional neural network | |
CN112529042B (en) | Medical image classification method based on dual-attention multi-example deep learning | |
CN107610087B (en) | Tongue coating automatic segmentation method based on deep learning | |
CN103886330B (en) | Sorting technique based on semi-supervised SVM integrated study | |
CN109101938B (en) | Multi-label age estimation method based on convolutional neural network | |
CN104598885B (en) | The detection of word label and localization method in street view image | |
CN105956560A (en) | Vehicle model identification method based on pooling multi-scale depth convolution characteristics | |
CN105160310A (en) | 3D (three-dimensional) convolutional neural network based human body behavior recognition method | |
CN103136504A (en) | Face recognition method and device | |
CN103984959A (en) | Data-driven and task-driven image classification method | |
CN105320945A (en) | Image classification method and apparatus | |
CN100414562C (en) | Method for positioning feature points of human face in human face recognition system | |
CN105825233B (en) | A kind of pedestrian detection method based on on-line study random fern classifier | |
CN109583375B (en) | Multi-feature fusion face image illumination identification method and system | |
CN103106265A (en) | Method and system of classifying similar images | |
CN108734138A (en) | A kind of melanoma skin disease image classification method based on integrated study | |
CN114998220B (en) | Tongue image detection and positioning method based on improved Tiny-YOLO v4 natural environment | |
CN103440510A (en) | Method for positioning characteristic points in facial image | |
CN102156871A (en) | Image classification method based on category correlated codebook and classifier voting strategy | |
CN103745233B (en) | The hyperspectral image classification method migrated based on spatial information | |
CN106204651A (en) | A kind of method for tracking target based on the judgement improved with generation conjunctive model | |
CN107590821A (en) | A kind of method for tracking target and system based on track optimizing | |
CN105976397A (en) | Target tracking method based on half nonnegative optimization integration learning | |
CN104200233A (en) | Clothes classification and identification method based on Weber local descriptor |
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 |
Granted publication date: 20180302 Termination date: 20190303 |
|
CF01 | Termination of patent right due to non-payment of annual fee |