CN107844785B - A kind of method for detecting human face based on size estimation - Google Patents
A kind of method for detecting human face based on size estimation Download PDFInfo
- Publication number
- CN107844785B CN107844785B CN201711294249.8A CN201711294249A CN107844785B CN 107844785 B CN107844785 B CN 107844785B CN 201711294249 A CN201711294249 A CN 201711294249A CN 107844785 B CN107844785 B CN 107844785B
- Authority
- CN
- China
- Prior art keywords
- model
- face
- training
- result
- size estimation
- 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.)
- Active
Links
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
- G06V40/168—Feature extraction; Face representation
-
- 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
-
- 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)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of method for detecting human face based on size estimation.The present invention first uses estimation of the face size estimation to face size on image, further according to face scaling image, quickly does Suggestion box extraction using full convolutional network, finally does cascade classification and recurrence twice and obtains Face datection result.The present invention does Face datection using the method that face size estimation and concatenated convolutional neural network combine, and reduces the calculation amount of Face datection entirety, reduces the time-consuming of Face datection entirety, and guarantees the effect of Face datection.
Description
Technical field
The invention belongs to technical field of video monitoring, are related to a kind of method for detecting human face based on size estimation.
Background technique
The function of method for detecting human face is to judge whether there is face in picture or video, if there is face, predicts face
Position and size.Face datection is the basis that various analyses are carried out to face.The time-consuming of Face datection is its crucial problem
One of.In most cases, in order to reduce time-consuming, it has to sacrifice the detection effect of a part.It can to the size estimation of face
To greatly reduce the time-consuming of Face datection.
In current existing technology, " a kind of method for detecting human face and device -201510639824.8 " uses AdaBoost
Method, quick multi-Scale Pyramid characteristic extract on the basis of classify, ensure accuracy in detection premise
Under can be effectively reduced calculation amount, but the method for AdaBoost effectively cannot promote detection using current a large amount of training data
Secondary performance." method and device -201610206093.2 of Face datection " extracts Suggestion box using AdaBoost and uses convolution refreshing again
Face datection is done through network, the threshold value selection that the Suggestion box of the method first step is extracted needs to adjust under different scenes, than
It is cumbersome." method and device -201710618497.7 of Face datection " does Face datection using cascade convolutional neural networks,
Because overall calculation amount is big, time-consuming problem is than more serious.
Summary of the invention
In view of the deficiencies of the prior art, the present invention provides a kind of method for detecting human face based on size estimation.
The present invention first uses estimation of the face size estimation to face size on image, further according to face scaling figure
Picture quickly does Suggestion box extraction using full convolutional network, finally does cascade classification and recurrence twice and obtains Face datection result.
The method of the present invention mainly comprises the steps that
Step 1: off-line training
1.1, training face size estimation
Scale space is divided into multiple sections by face size estimation.Each section is made a decision, on input picture whether
In the presence of the face for belonging to this section.In simple terms, face size estimation be exactly do it is multiple two classification, obtain be to multiple sections
It is no that there are the scores vectors of this scale face.
1.1.1, original image is scaled；
1.1.2, according to the wide high mean value computation face scale scores vector of the face on zoomed image.For an area
Between, if there is the face for belonging to this interval scale, corresponding score is set to 1 on scores vector, is positive sample；If do not deposited
Corresponding score is set to 0 on the face for belonging to this interval scale, scores vector, is negative sample.
1.1.3, the loss function of training uses the cross entropy loss function weighted:
Loss indicates loss.N indicates Scaling interval quantity.N indicates Scaling interval serial number.wnIndicate n-th of Scaling interval
Weight.pnIndicate the score of n-th of Scaling interval.pnIndicate the score of n-th of Scaling interval.Indicate n-th of Scaling interval
Estimated result.
Further, it when needing to do fine accurate size estimation, needs closeer to divide Scaling interval, i.e. N meeting
Become very big.But positive sample quantity is basically unchanged, this will cause, and positive negative sample is uneven, and training is difficult to restrain.To positive sample
This plus a weight bigger than negative sample can be conducive to train convergence.The face scale scores vector side that labeled data generates
Difference is very big, and in order to alleviate this problem, the negative sample weight near positive sample is set to 0.
1.1.4, training face scale classifier does flip horizontal disturbance to image when training.
1.2, training Stage1 model.Stage1 model is a multi task model, and task is classification and returns.Classification is appointed
Classification chart tile of being engaged in is face, returns the position that task returns two points in face bounding box upper left and bottom right.Use face
Labeled data generates the training sample of Stage1 model.
1.3, training Stage2 model.Stage2 model is a multi task model, and task is classification and returns.Relative to
Stage1 model, the classification of Stage2 model and regression capability are stronger, and model size is also bigger.Use the result of Stage1 model
Multiple dimensioned scanning is carried out on original image, classification score is greater than training sample of the scan box result as Stage2 model of threshold value
This.
1.4, training Stage3 model.Stage2 model is a multi task model, and task is classification and returns.Relative to
Stage1 model and Stage2 model, its classification and regression capability is stronger, and model size is also bigger.Using Stage1 model and
The result of Stage2 model obtains the training sample of Stage3 model, and the training sample input of Stage3 model is two images
Block, one be Stage2 model regression result, one is that the regression result of Stage2 model diffuses into its twice of size outward
Image block.After the completion of Stage3 training result, then difficult example is done to Stage3 and is excavated, finely tunes model.
Step 2: on-line checking
2.1, input picture.
2.2, the size estimation of face:
2.2.1, input picture is scaled, inputs trained face size estimation model, obtains face size estimation
Scores vector.
2.2.2, the scores vector of face size estimation is done smoothly.
2.2.3, non-maxima suppression is done to the scores vector of face size estimation, obtains face scale.
2.3, Stage1 model is a full convolutional network.Using face size estimation as a result, being scaled to input picture,
Input Stage1 model.
2.4, the result of Stage1 model is inputted into Stage2 model, does and classifies and return.
2.5, the result of Stage2 model is inputted into Stage3 model, inputted as two image blocks, one is Stage2 mould
The regression result of type, one is that the regression result of Stage2 model diffuses into the frame of its twice of size outward, does and classifies and return.
2.6, the detection block for obtaining Stage3 model merges, and consolidation strategy uses non-maxima suppression.
2.7, output test result.
Beneficial effects of the present invention:
1, using face size estimation, the image pyramid number of plies of subsequent Stage1 model is reduced, accelerates average detected speed
Degree.
2, face size estimation makes training process be easy convergence using the cross entropy loss function of weighting, solves sample mark
Infuse variance problems of too, available more accurate size estimation result.
3, Stage1 model uses full convolutional network, reduces convolutional calculation amount, promotes detection speed.
4, using cascade detection method, the model structure of three Stage is from gently to again, available entirety is preferably examined
Degree of testing the speed and performance.
5, the input of Stage3 model also inputs the image block of its twice of size in addition to the regression result of Stage2,
The information for making Stage3 mode input face near zone, promotes the accuracy rate of detection.
Detailed description of the invention
Fig. 1 is the method for the present invention flow chart.
Fig. 2 is face scale classifier network structure.
Fig. 3 is the network structure of Stage1 model.
Fig. 4 is the network structure of Stage2 model.
Fig. 5 is the network structure of Stage3 model.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, the technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
It is only a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiment of the present invention, ordinary skill people
Member's every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Specific implementation step of the invention is:
Step 1: off-line training
1.1, training face size estimation.
1.1.1, original image is zoomed into 224X224,224 scaling figure is zoomed to by long side, short side is mended 0 and filled out
It fills.
1.1.2, according to the wide high mean value computation face scaled target scores vector of the face on zoomed image.Face scale
Target zone is [22.6,28], space ratio 20.1, it is 55 two classification in total.Than if any 24One or more this scale
Face, then for this picture 24This classification is positive sample.In order to solve the problems, such as positive and negative imbalanced training sets, positive sample
Classified weight is 16, and the classified weight of negative sample is 1.It, will be in 1 field of positive sample in order to solve the problems, such as that labeled data variance is big
The classified weight of negative sample be set to 0
1.1.3, using caffe training face scale classifier, flip horizontal disturbance, network structure are done to image when training
See Fig. 2.Conv_blokA contains one convolutional layer, active coating a ReLU, a normalization layer BN.Conv_blokA's
The convolution kernel size of convolutional layer is 3x3, and step-length 1 is filled with 1.Conv_blokB contains one convolutional layer, an activation
Layer ReLU, a normalization layer BN.The convolution kernel size of the convolutional layer of conv_blokB is 5x5, and step-length 1 is filled with 2.
Conv_blokC contains one convolutional layer, active coating a ReLU, a normalization layer BN.The convolutional layer of conv_blokC
Convolution kernel size be 1x1, step-length 1 is filled with 0.Inception is the complicated convolution knot of a multiple convolutional coding structure compositions
Structure, structure are shown in figure.Concat is an articulamentum in Inception, by conv_blokC1, conv_blokB2, conv_
BlokC3_2, conv_blokC4, totally 4 convolution connect together.It is 224x224x8 that conv_blokA1, which exports dimension, other
Convolution block output size is shown in network structure.Conv_cls is that a convolution kernel size is 3x3, and step-length 1 is filled with 1 volume
Lamination.Global_max_pool is global maximum pond layer.Prob is softMax layers.
1.2, training Stage1 model.Stage1 model is a multi task model, and task is classification and returns.Classification is appointed
Classification chart tile of being engaged in is face, returns the position that task returns two points in face bounding box upper left and bottom right.Use face
Labeled data, generates the training sample of Stage1 model, and tile size zooms to 12X12, positive and negative sample proportion 1:3.
The network structure of Stage1 model is shown in Fig. 3.
Conv_blok contains one convolutional layer, active coating a ReLU, a normalization layer BN.Conv_blok's
The convolution kernel size of convolutional layer is 3x3, and step-length 1 is filled with 1.It is 12x12x8, other convolution that conv_blok1, which exports dimension,
Block output size is shown in network structure.Conv_cls is that a convolution kernel size is 3x3, and step-length 1 is filled with 0 convolutional layer.
Fc_cls is that a convolution kernel size is 1x1, and step-length 1 is filled with 0 convolutional layer.Prob is softMax layers.conv_reg
It is a convolution kernel size for 3x3, step-length 1 is filled with 0 convolutional layer.Fc_reg is that a convolution kernel size is 1x1, step
A length of 1, it is filled with 0 convolutional layer.
1.3, training Stage2 model.Stage2 model is a multi task model, and task is classification and returns.Relative to
Stage1 model, the classification of Stage2 model and regression capability are stronger, and model size is also bigger.Use the result of Stage1 model
Multiple dimensioned scanning, training sample of scan box result of the classification score greater than 0.5 as Stage2 model are carried out on original image
This, training sample image block size zooms to 24X24, positive and negative sample proportion 1:1.The network structure of Stage2 is shown in Fig. 4.Mark with
Fig. 3 is similar.
1.4, training Stage3 model.Stage2 model is a multi task model, and task is classification and returns.Relative to
Stage1 model and Stage2 model, its classification and regression capability is stronger, and model size is also bigger.Using Stage1 model and
The result of Stage2 model obtains the training sample of Stage3 model, and the training sample input of Stage3 model is two images
Block, one be Stage2 model regression result, one is that the regression result of Stage2 model diffuses into its twice of size outward
Image block, two image blocks all zoom to 32X32.After the completion of Stage3 training result, then difficult example is done to Stage3 and is excavated, fine tuning
Model.The network structure of Stage3 is shown in Fig. 5.There are two data input layer data1 and data2.Concat is an articulamentum, will
Conv_block1_1 and conv_block1_2 are connected in series together.Other marks are similar to Fig. 3.
Step 2: on-line checking
2.1, input picture.
2.2, the size estimation of face:
2.2.1, the longest edge of input picture is zoomed to 224, inputs trained face size estimation model, obtains 55
The score of a face size estimation.
2.2.2, to the score of 55 face size estimations do window be 3 it is smooth.
2.2.3, the one-dimensional non-maxima suppression that window is 5 is done to the score of 55 face size estimations, obtains face ruler
Degree.
2.3, using face size estimation as a result, scaling to input picture, input Stage1 model, Stage1 model is
One full convolutional network, obtains first step testing result.
2.4, the result of Stage1 model is inputted into Stage2 model, input picture block zooms to 24X24, does and classify and return
Return, obtains second step testing result.
2.5, the result of Stage2 model is inputted into Stage3 model, inputted as two image blocks, one is Stage2 mould
The regression result of type, one is that the regression result of Stage2 model diffuses into the frame of its twice of size outward, and two image blocks all contract
It is put into 32X32, is done and is classified and return, third portion testing result is obtained.
2.6, the detection block for obtaining Stage3 model merges, and consolidation strategy uses non-maxima suppression.
2.7, output test result.
To sum up, the present invention does Face datection using the method that face size estimation and concatenated convolutional neural network combine, and subtracts
The calculation amount of few Face datection entirety, reduces the time-consuming of Face datection entirety, and guarantees the effect of Face datection.
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the scope of the present invention, should
Understand, the present invention is not limited to implementation as described herein, the purpose of these implementations description is to help this field
In technical staff practice the present invention.
Claims (4)
1. a kind of method for detecting human face based on size estimation, it is characterised in that method includes the following steps:
Off-line training model and on-line checking face two parts；Wherein
Off-line training model is specifically:
Step 1.1, training face size estimation:
Step 1.1.1, original image is scaled；
Step 1.1.2, according to the wide high mean value computation face scale scores vector of the face on zoomed image；For an area
Between, if there is the face for belonging to this interval scale, corresponding score is set to 1 on scores vector, is positive sample；If do not deposited
Corresponding score is set to 0 on the face for belonging to this interval scale, scores vector, is negative sample；
Step 1.1.3, use the cross entropy loss function of weighting as the loss function of training:
Wherein Loss indicates loss, and N indicates Scaling interval quantity, and n indicates Scaling interval serial number, wnIndicate n-th of Scaling interval power
Weight, pnIndicate the score of n-th of Scaling interval,Indicate the estimated result of n-th of Scaling interval；
Step 1.1.4, training face scale classifier does flip horizontal disturbance to image when training；
Step 1.2, training Stage1 model；Stage1 model is a full convolutional network；Using face labeled data, generate
The training sample of Stage1 model；
Step 1.3, training Stage2 model；Multiple dimensioned scanning is carried out on original image using the result of Stage1 model, point
Class score is greater than training sample of the scan box result as Stage2 model of threshold value；
Step 1.4, training Stage3 model；The instruction of Stage3 model is obtained using the result of Stage1 model and Stage2 model
Practice sample, Stage3 model training sample input be two image blocks, one be Stage2 model regression result, one is
The regression result of Stage2 model diffuses into the image block of its twice of size outward；After the completion of Stage3 training result, then it is right
Stage3 does difficult example and excavates, and finely tunes model；
On-line checking face is specifically:
Step 2.1, input picture；
The size estimation of step 2.2, face:
Step 2.2.1, input picture is scaled, inputs trained face size estimation model, obtains face size estimation
Scores vector；
Step 2.2.2, the scores vector of face size estimation is done smoothly；
Step 2.2.3, non-maxima suppression is done to the scores vector of face size estimation, obtains face scale；
Step 2.3, using face size estimation as a result, scaled to input picture, input Stage1 model, the Stage1
Model is a full convolutional network；
The result of Stage1 model is inputted Stage2 model by step 2.4, is done and is classified and return；
The result of Stage2 model is inputted Stage3 model by step 2.5, is inputted as two image blocks, one is Stage2 mould
The regression result of type, one is that the regression result of Stage2 model diffuses into the frame of its twice of size outward, does and classifies and return；
Step 2.6, the detection block for obtaining Stage3 model merge, and consolidation strategy uses non-maxima suppression；
Step 2.7, output test result.
2. a kind of method for detecting human face based on size estimation according to claim 1, it is characterised in that: when needing to do essence
When thin accurate size estimation, the weight bigger than negative sample to positive sample plus one.
3. a kind of method for detecting human face based on size estimation according to claim 2, it is characterised in that: positive sample is attached
Close negative sample weight is set to 0.
4. a kind of method for detecting human face based on size estimation according to any one of claim 1 to 3, feature exist
In: the task of the Stage1 model is classification and returns；Classification task: classification chart tile is face, returns task:
Return the position of two points in face bounding box upper left and bottom right.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711294249.8A CN107844785B (en) | 2017-12-08 | 2017-12-08 | A kind of method for detecting human face based on size estimation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711294249.8A CN107844785B (en) | 2017-12-08 | 2017-12-08 | A kind of method for detecting human face based on size estimation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107844785A CN107844785A (en) | 2018-03-27 |
| CN107844785B true CN107844785B (en) | 2019-09-24 |
Family
ID=61663261
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711294249.8A Active CN107844785B (en) | 2017-12-08 | 2017-12-08 | A kind of method for detecting human face based on size estimation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107844785B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110555334B (en) * | 2018-05-30 | 2022-06-07 | 东华软件股份公司 | Face feature determination method and device, storage medium and electronic equipment |
| CN109117886B (en) * | 2018-08-17 | 2022-02-18 | 浙江捷尚视觉科技股份有限公司 | Method for estimating target dimension and region in image frame |
| CN109829371B (en) * | 2018-12-26 | 2022-04-26 | 深圳云天励飞技术有限公司 | Face detection method and device |
| CN109815332B (en) * | 2019-01-07 | 2023-06-20 | 平安科技（深圳）有限公司 | Loss function optimization method, loss function optimization device, computer equipment and storage medium |
| CN111488517A (en) * | 2019-01-29 | 2020-08-04 | 北京沃东天骏信息技术有限公司 | Method and device for training click rate estimation model |
| CN109961006A (en) * | 2019-01-30 | 2019-07-02 | 东华大学 | A kind of low pixel multiple target Face datection and crucial independent positioning method and alignment schemes |
| CN110188720A (en) * | 2019-06-05 | 2019-08-30 | 上海云绅智能科技有限公司 | A kind of object detection method and system based on convolutional neural networks |
| CN110580445B (en) * | 2019-07-12 | 2023-02-07 | 西北工业大学 | Face key point detection method based on GIoU and weighted NMS improvement |
| CN110619350B (en) * | 2019-08-12 | 2021-06-18 | 北京达佳互联信息技术有限公司 | Image detection method, device and storage medium |
| CN112434178A (en) * | 2020-11-23 | 2021-03-02 | 北京达佳互联信息技术有限公司 | Image classification method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105912990A (en) * | 2016-04-05 | 2016-08-31 | 深圳先进技术研究院 | Face detection method and face detection device |
| CN105975961A (en) * | 2016-06-28 | 2016-09-28 | 北京小米移动软件有限公司 | Human face recognition method, device and terminal |
| CN106056101A (en) * | 2016-06-29 | 2016-10-26 | 哈尔滨理工大学 | Non-maximum suppression method for face detection |
| CN106897732A (en) * | 2017-01-06 | 2017-06-27 | 华中科技大学 | Multi-direction Method for text detection in a kind of natural picture based on connection word section |
| CN107103281A (en) * | 2017-03-10 | 2017-08-29 | 中山大学 | Face identification method based on aggregation Damage degree metric learning |
-
2017
- 2017-12-08 CN CN201711294249.8A patent/CN107844785B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105912990A (en) * | 2016-04-05 | 2016-08-31 | 深圳先进技术研究院 | Face detection method and face detection device |
| CN105975961A (en) * | 2016-06-28 | 2016-09-28 | 北京小米移动软件有限公司 | Human face recognition method, device and terminal |
| CN106056101A (en) * | 2016-06-29 | 2016-10-26 | 哈尔滨理工大学 | Non-maximum suppression method for face detection |
| CN106897732A (en) * | 2017-01-06 | 2017-06-27 | 华中科技大学 | Multi-direction Method for text detection in a kind of natural picture based on connection word section |
| CN107103281A (en) * | 2017-03-10 | 2017-08-29 | 中山大学 | Face identification method based on aggregation Damage degree metric learning |
Non-Patent Citations (1)
| Title |
|---|
| "A Multi-Scale Fusion Convolutional Neural Network for Face Detection";Qiaosong Chen etal.;《IEEE》;20171008;第1013-1018页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107844785A (en) | 2018-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107844785B (en) | A kind of method for detecting human face based on size estimation | |
| CN104463117B (en) | A kind of recognition of face sample collection method and system based on video mode | |
| Zhang et al. | Pedestrian detection method based on Faster R-CNN | |
| CN103839065B (en) | Extraction method for dynamic crowd gathering characteristics | |
| US20200005022A1 (en) | Method, terminal, and storage medium for tracking facial critical area | |
| CN106446930B (en) | Robot operative scenario recognition methods based on deep layer convolutional neural networks | |
| CN109635694B (en) | Pedestrian detection method, device and equipment and computer readable storage medium | |
| CN102542289B (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
| CN105160317B (en) | One kind being based on area dividing pedestrian gender identification method | |
| CN100361138C (en) | Method and system of real time detecting and continuous tracing human face in video frequency sequence | |
| CN105913003B (en) | A kind of pedestrian detection method of multiple features multi-model | |
| CN102682287B (en) | Pedestrian detection method based on saliency information | |
| WO2019114036A1 (en) | Face detection method and device, computer device, and computer readable storage medium | |
| CN103093212B (en) | The method and apparatus of facial image is intercepted based on Face detection and tracking | |
| CN103440645A (en) | Target tracking algorithm based on self-adaptive particle filter and sparse representation | |
| CN109446922B (en) | Real-time robust face detection method | |
| WO2019080203A1 (en) | Gesture recognition method and system for robot, and robot | |
| CN105260749B (en) | Real-time target detection method based on direction gradient binary pattern and soft cascade SVM | |
| CN107909027A (en) | It is a kind of that there is the quick human body target detection method for blocking processing | |
| Ren et al. | A novel squeeze YOLO-based real-time people counting approach | |
| CN109670405A (en) | A kind of complex background pedestrian detection method based on deep learning | |
| CN101196991A (en) | Close passenger traffic counting and passenger walking velocity automatic detection method and system thereof | |
| CN112528913A (en) | Grit particulate matter particle size detection analytic system based on image | |
| CN101364263A (en) | Method and system for detecting skin texture to image | |
| CN102693427A (en) | Method and device for forming detector for detecting images |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230428 Address after: Room 319-2, 3rd Floor, Building 2, No. 262 Wantang Road, Xihu District, Hangzhou City, Zhejiang Province, 310012 Patentee after: Hangzhou Jieshang Safety Equipment Co.,Ltd. Address before: 311121 East Building, building 7, No. 998, Wenyi West Road, Wuchang Street, Yuhang District, Hangzhou City, Zhejiang Province Patentee before: ZHEJIANG ICARE VISION TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right |