CN109558813A - A kind of AI depth based on pulse signal is changed face video evidence collecting method - Google Patents
A kind of AI depth based on pulse signal is changed face video evidence collecting method Download PDFInfo
- Publication number
- CN109558813A CN109558813A CN201811352507.8A CN201811352507A CN109558813A CN 109558813 A CN109558813 A CN 109558813A CN 201811352507 A CN201811352507 A CN 201811352507A CN 109558813 A CN109558813 A CN 109558813A
- Authority
- CN
- China
- Prior art keywords
- video
- face
- signal
- pulse signal
- power spectrum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V40/161—Detection; Localisation; Normalisation
-
- 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
- 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
- G06V40/15—Biometric patterns based on physiological signals, e.g. heartbeat, blood flow
Abstract
The AI depth based on pulse signal that the invention discloses a kind of is changed face video evidence collecting method, since the cardiovascular pulse wave propagated in human body periodically can cause vascular wall to stretch, it is fluctuated so that the tissue containing a large amount of blood vessels is also synchronous to the absorbability of light, to reflect regular pulse signal.In face video shooting process, these minor changes being invisible to the naked eye but can destroy these regular governed minor changes by the forgery face that common camera is recorded, and generated using AI method.According to this characteristic, the classifier that present invention combination machine learning algorithm SVM training obtains, effectively identification forge the improper due pulse signal of human body in video by the face true to nature of representative of deep learning, to be achieved the purpose of video evidence obtaining.The present invention does not need that video to be detected is required to contain digital watermark information, does not need to install other additional detection devices, therefore with good application prospect yet.
Description
Technical field
The invention belongs to field of information security technology, are related to a kind of evidence collecting method based on pulse signal, more particularly to
It is a kind of based on pulse signal, for depth AI forge video of changing face evidence collecting method.
Background technique
With the continuous development of computer vision and image processing techniques, we can accomplish to pass through deep learning now
Method generate face extremely true to nature in video.The behind of this technology is unlimited application space, but therein certain
Abuse has but sounded safety alarm for us, and the video after distorting misleads to the public by Internet communication, not only can
Daily life is upset, and the harmony that can seriously threaten society is stablized, therefore designs a reliable video evidence obtaining side
Method is extremely urgent.
In order to guarantee multimedia content safety, academia has carried out a large amount of research to video evidence obtaining work, at this stage
Technology be broadly divided into following a few classes:
1, video active forensic technologies, such technical requirements content providers must pre-process video, such as extract
Abstract or insertion verification information etc., representative method is digital signature and video watermark technology.Such methods are not applicable
The evidence obtaining work of multitude of video from a wealth of sources in internet.
2, the passive forensic technologies of video, such technology mainly pass through compression domain, time domain or the sky for differentiating video to be detected
Characteristic of field to video carries out evidence obtaining work.Such method multitude of video data from a wealth of sources suitable for internet, and have
Biggish exploration and perfect space.
Summary of the invention
The characteristics of forging evidence obtaining research for existing video and deficiency, the present invention provides a kind of taking based on pulse signal
Card method is collected evidence mainly for the forgery face video generated using deep learning, and this method does not need to require to be detected
Video contains digital watermark information, does not also need to install other additional detection devices, can the multimedia content peace of effective guarantee
Entirely.
The technical scheme adopted by the invention is that: a kind of AI depth based on pulse signal is changed face video evidence collecting method,
It is characterized in that, comprising the following steps:
Step 1: collecting the video for containing true and false face, video data is divided into training set and test set two parts, and pre-
Region of interest ROI is selected in processing；
Step 2: extracting pretreated data-signal, obtain corresponding power spectrum；
Step 3: according to power spectrum, selecting feature vector, input the feature into and carry out model training in SVM；
Step 4: the model obtained using step 3 predicts video to be detected, determines whether forgery face video.
Since the cardiovascular pulse wave propagated in human body periodically can cause vascular wall to stretch, so that containing a large amount of blood vessels
Tissue it is also synchronous to the absorbability of light fluctuate, to reflect regular pulse signal.In face video shooting process
In, these minor changes being invisible to the naked eye, the adulterator that but can be recorded by common camera, and be generated using AI method
Face can destroy these regular governed minor changes.According to this characteristic, present invention combination machine learning algorithm SVM is trained
The classifier arrived, effectively identification forge the improper due pulse of human body in video by the face true to nature of representative of deep learning
Signal, to be achieved the purpose of video evidence obtaining.
The present invention has the advantage that compared with existing evidence collecting method
1, a kind of evidence collecting method based on pulse signal is proposed, although the method that can use deep learning at this stage is pseudo-
The face video for being difficult to distinguish is produced, but the method for this deep learning can destroy the pulse signal of normal human, so as to
As a kind of evidence collecting method；
2, the method proposed does not need to carry out complicated human-computer interaction, better user experience；
3, the method proposed does not need that video to be detected is required to contain digital watermark information；
4, the method proposed does not need to install other additional detection devices, easy to operate, at low cost；
5, propose method combine the SVM classifier in machine learning, using sample training model, reduce manually at
This, while detection efficiency and accuracy rate can be made to be promoted.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention；
Fig. 2 is the procedure chart to power spectrum of signal processing simultaneously in the embodiment of the present invention.
Fig. 3 is the training flow chart of SVM of the embodiment of the present invention；
Fig. 4 is the differentiation procedure chart of video of embodiment of the present invention evidence obtaining.
Specific embodiment
Understand for the ease of those of ordinary skill in the art and implement the present invention, with reference to the accompanying drawings and embodiments to this hair
It is bright to be described in further detail, it should be understood that implementation example described herein is merely to illustrate and explain the present invention, not
For limiting the present invention.
The video evidence collecting method referring to Fig.1, a kind of AI depth based on pulse signal provided by the invention is changed face, including it is following
Step:
Step 1: collecting the video ([document 1]) for containing true and false face, video data is divided into training set and test set two
Part, and pre-process and select region of interest ROI；
Training set and test set are divided into the true of acquisition and forgery video data, and pre-processed to obtain region of interest
Domain ROI, specific implementation include following sub-step:
Step 1.1: according to machine learning to the conventional treatment method of data set, 2/3~4/5 sample of data set being regarded
Training set, remaining sample regard test set；
Step 1.2: ([document 2]) is pre-processed to video using Haar feature, detects the face in video, due to
The volumetric blood pulse BVP signal of estimation pulse is distributed uneven suction according to face skin to illumination on face skin
The face detected is further processed using the detectMultiScale function in OpenCV, determines volume for receipts ability
Head and cheek are required region of interest ROI.
Step 2: extracting pretreated data-signal, obtain corresponding power spectrum；
What Fig. 2 was indicated is to extract pretreated data-signal, and obtain the process of corresponding power spectrum, and specific steps are such as
Under:
Step 2.1: every frame region of interest ROI image that step 1 is chosen carries out 3 primary colours separation, generates R, G, B tri-
Then channel image takes signal value of the average gray of each channel all pixels as the frame image, form 3 original letters
Number, as original pulse signal；
Step 2.2: the original pulse signal extracted in step 2.1 being subjected to trending and is handled, in order to observe arteries and veins
It fights the inherent characteristic of signal；
Step 2.3: trending will be removed in step 2.2 treated signal, use independent component analysis ICA method ([document
3]) isolated pure pulse original signal；
Step 2.4: the signal that step 2.3 is obtained carries out signal screening using correlation analysis；
Step 2.5: by the signal after step 2.4 screening, changing FFT using fast Fourier, obtain corresponding power spectrum.
Step 3: according to power spectrum, selecting feature vector, input the feature into progress model training ([document in SVM
4])；
Step 3.1: according to power spectrum, selecting feature vector；
In terms of feature vector, the feature of selection has the corresponding pulse signal peak value P (r) of RGB triple channel, P (g), P (b),
And RGB triple channel is in summation S (r), the S of the corresponding power spectrum of 0.7HZ to 4HZ (i.e. 42bpm to 240bpm Pulse Rate)
(g),S(b)；
Step 3.2: inputting the feature into and carry out model training in SVM；
The feature that step 3.1 is proposed, which is input in SVM, carries out model training, and training process is as shown in figure 3, training goal
It is to obtain one video sample can be divided into the hyperplane of true, pseudo- two class, and be spaced maximum, formula is expressed as follows:
Wherein ω is normal vector, and b is displacement item, determines the distance between hyperplane and origin, (χi,γi) it is corresponding
Data set, n are video sample quantity.
Step 4: the model obtained using step 3 predicts video to be detected, determines whether forgery face video；
As shown in Figure 4, the specific steps are as follows:
Step 4.1: video to be detected being pre-processed first, selects area-of-interest；
Step 4.2: extracting the signal of pretreatment rear video, mainly have primary colours separation, remove trending, Blind Signal Separation, letter
It number screens and extracts power spectrum；
Step 4.3: according to power spectrum, the corresponding pulse signal peak value P (r) of RGB triple channel, P (g), P (b) are chosen, with
And RGB triple channel is in summation S (r), the S of the corresponding power spectrum of 0.7HZ to 4HZ (i.e. 42bpm to 240bpm Pulse Rate)
(g),S(b)；
Step 4.4, it inputs the feature into the trained classifier of step 3, calculates confidence level and be compared with threshold value, from
And it is able to determine whether video to be detected is to forge face video；
If confidence level is greater than threshold value, for real human face video；It otherwise, is forgery face video.
It should be understood that the part that this specification does not elaborate belongs to the prior art.
The characteristics of forging evidence collecting method for the above video and deficiency, the present invention propose a kind of AI based on pulse signal
Depth is changed face video evidence collecting method.Proposed method is primarily directed to be taken using the forgery face that deep learning generates
Card, since the cardiovascular pulse wave propagated in human body periodically can cause vascular wall to stretch, internal volumetric blood and blood oxygen contain
Amount changes therewith, fluctuates so that the tissue containing a large amount of blood vessels is also synchronous to the absorbability of light, to realize optical body
The purpose of product cardiotokography (PPG) non-contact detection video human face heart rate ([document 5]), and network (GAN) is fought with production
Do not have regular governed arteries and veins possessed by real human face for the depth AI of the representative forgery face video ([document 6]) synthesized
It fights information, so method can be used as a kind of new method that face forges evidence obtaining.Meanwhile this method does not need to require view to be detected
Frequency contains digital watermark information, does not also need to install other additional detection devices, therefore with good application prospect.
[1]Thies J,Zollhofer M,Stamminger M,et al.Face2Face:Real-Time Face
Capture and Reenactment of RGB Videos[C]//Computer Vision and Pattern
Recognition.
IEEE,2016:1-2.
[2]Mita T,Kaneko T,Hori O.Joint Haar-like features for face detection
[C]//Tenth IEEE International Conference on Computer Vision.2005:1619-1626.
[3]Alghoul K,Alharthi S,Osman H A,et al.Heart Rate Variability
Extraction From Videos Signals:ICA vs.EVM Comparison[J].IEEE Access,2017,5
(99):4711-4719.
[4]Bennett K P,Campbell C.Support vector machines:hype or hallelujah
[J].Acm Sigkdd Explorations Newsletter,2000,2(2):1-13.
[5]Verkruysse W,Svaasand LO,Nelson JS.Remote plethysmographic imaging
using ambient light[J].Optics Express,2008,16(26):21434-45.
[6]Korshunova I,Shi W,Dambre J,et al.Fast Face-Swap Using
Convolutional Neural Networks[J].2016:3697-3705.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, those skilled in the art under the inspiration of the present invention, are not departing from power of the present invention
Benefit requires to make replacement or deformation under protected ambit, fall within the scope of protection of the present invention, this hair
It is bright range is claimed to be determined by the appended claims.
Claims (5)
- The video evidence collecting method 1. a kind of AI depth based on pulse signal is changed face, which comprises the following steps:Step 1: collecting the video for containing true and false face, video data is divided into training set and test set two parts, and pre-process Obtain region of interest ROI；Step 2: extracting pretreated data-signal, obtain corresponding power spectrum；Step 3: according to power spectrum, selecting feature vector, input the feature into and carry out model training in SVM；Step 4: the model obtained using step 3 predicts video to be detected, determines whether forgery face video.
- The video evidence collecting method 2. the AI depth according to claim 1 based on pulse signal is changed face, it is characterised in that: step In 1, training set and test set are divided into the true of acquisition and forgery video data, and pre-processed to obtain area-of-interest ROI, specific implementation include following sub-step:Step 1.1: 2/3~4/5 sample of data set being regarded into training set, remaining sample regards test set；Step 1.2: video being pre-processed using Haar feature, detects the face in video, due to estimating the blood of pulse Liquid volume pulse BVP signal be distributed on face skin it is uneven, according to face skin to the absorbability of illumination, to detection To face be further processed, to obtain region of interest ROI.
- The video evidence collecting method 3. the AI depth according to claim 1 based on pulse signal is changed face, which is characterized in that step 2 specific implementation includes following sub-step:Step 2.1: every frame region of interest ROI image that step 1 is chosen carries out 3 primary colours separation, generates R, G, B triple channel figure Then picture takes signal value of the average gray of each channel all pixels as the frame image, forms 3 original signals, as Original pulse signal；Step 2.2: the original pulse signal extracted in step 2.1 being subjected to trending and is handled, in order to observe that pulse is believed Number inherent characteristic；Step 2.3: trending will be removed in step 2.2 treated signal, isolated pure pulse original signal；Step 2.4: the signal that step 2.3 is obtained, the signal screening more refined；Step 2.5: using the signal after step 2.4 screening, obtaining corresponding power spectrum.
- The video evidence collecting method 4. the AI depth according to claim 1 based on pulse signal is changed face, which is characterized in that step 3 specific implementation includes following sub-step:Step 3.1: according to power spectrum, selecting feature vector；In terms of feature vector, the feature of selection has the corresponding pulse signal peak value P (r) of RGB triple channel, P (g), P (b), and Summation S (r), S (g), S (b) of the RGB triple channel in the corresponding power spectrum of 0.7HZ to 4HZ；Step 3.2: inputting the feature into and carry out model training in SVM；The feature that step 3.1 is proposed, which is input in SVM, carries out model training, so that video sample can be divided by obtaining one Very, the hyperplane of pseudo- two classes, and it is spaced maximum；Formula is expressed as follows:Wherein ω is normal vector, and b is displacement item, determines the distance between hyperplane and origin, (χi,γi) it is corresponding data Collection, n are video sample quantity.
- The video evidence collecting method 5. the AI depth according to claim 1 based on pulse signal is changed face, which is characterized in that step 4 specific implementation includes following sub-step:Step 4.1: video to be detected being pre-processed first, selects area-of-interest；Step 4.2: extracting the signal of pretreatment rear video, mainly have primary colours separation, remove trending, Blind Signal Separation, signal sieve It selects and extracts power spectrum；Step 4.3: according to power spectrum, choosing the corresponding pulse signal peak value P (r) of RGB triple channel, P (g), P (b) and RGB Summation S (r), S (g), S (b) of the triple channel in the corresponding power spectrum of 0.7HZ to 4HZ；Step 4.4, it inputs the feature into the trained classifier of step 3, calculates confidence level and be compared with threshold value, thus To determine whether video to be detected is to forge face video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352507.8A CN109558813A (en) | 2018-11-14 | 2018-11-14 | A kind of AI depth based on pulse signal is changed face video evidence collecting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352507.8A CN109558813A (en) | 2018-11-14 | 2018-11-14 | A kind of AI depth based on pulse signal is changed face video evidence collecting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109558813A true CN109558813A (en) | 2019-04-02 |
Family
ID=65866229
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811352507.8A Pending CN109558813A (en) | 2018-11-14 | 2018-11-14 | A kind of AI depth based on pulse signal is changed face video evidence collecting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109558813A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110279406A (en) * | 2019-05-06 | 2019-09-27 | 苏宁金融服务(上海)有限公司 | A kind of touchless pulse frequency measurement method and device based on camera |
| CN111311549A (en) * | 2020-01-20 | 2020-06-19 | 中国人民解放军国防科技大学 | Image authentication method, system and storage medium |
| CN111783644A (en) * | 2020-06-30 | 2020-10-16 | 百度在线网络技术（北京）有限公司 | Detection method, device, equipment and computer storage medium |
| CN111797735A (en) * | 2020-06-22 | 2020-10-20 | 深圳壹账通智能科技有限公司 | Face video recognition method, device, equipment and storage medium |
| CN113963427A (en) * | 2021-12-22 | 2022-01-21 | 浙江工商大学 | Method and system for rapid in vivo detection |
| CN114402359A (en) * | 2019-07-18 | 2022-04-26 | 纽洛斯公司 | System and method for detecting composite video of person |
| CN115410260A (en) * | 2022-09-05 | 2022-11-29 | 中国科学院自动化研究所 | Face counterfeit discrimination and evidence obtaining method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150193669A1 (en) * | 2011-11-21 | 2015-07-09 | Pixart Imaging Inc. | System and method based on hybrid biometric detection |
| CN105989357A (en) * | 2016-01-18 | 2016-10-05 | 合肥工业大学 | Human face video processing-based heart rate detection method |
| CN106473750A (en) * | 2016-10-08 | 2017-03-08 | 西安电子科技大学 | Personal identification method based on photoplethysmographic optimal period waveform |
| CN107506713A (en) * | 2017-08-15 | 2017-12-22 | 哈尔滨工业大学深圳研究生院 | Living body faces detection method and storage device |
| CN108038456A (en) * | 2017-12-19 | 2018-05-15 | 中科视拓（北京）科技有限公司 | A kind of anti-fraud method in face identification system |
-
2018
- 2018-11-14 CN CN201811352507.8A patent/CN109558813A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150193669A1 (en) * | 2011-11-21 | 2015-07-09 | Pixart Imaging Inc. | System and method based on hybrid biometric detection |
| CN105989357A (en) * | 2016-01-18 | 2016-10-05 | 合肥工业大学 | Human face video processing-based heart rate detection method |
| CN106473750A (en) * | 2016-10-08 | 2017-03-08 | 西安电子科技大学 | Personal identification method based on photoplethysmographic optimal period waveform |
| CN107506713A (en) * | 2017-08-15 | 2017-12-22 | 哈尔滨工业大学深圳研究生院 | Living body faces detection method and storage device |
| CN108038456A (en) * | 2017-12-19 | 2018-05-15 | 中科视拓（北京）科技有限公司 | A kind of anti-fraud method in face identification system |
Non-Patent Citations (6)
| Title |
|---|
| GUILLAUME GIBERT等: "Face detection method based on photoplethysmography", 《2013 10TH IEEE INTERNATIONAL CONFERENCE ON ADVANCED VIDEO AND SIGNAL BASED SURVEILLANCE》 * |
| KARIM ALGHOUL等: "Heart Rate Variability Extraction From Videos Signals: ICA vs. EVM Comparison", 《IEEE ACCESS》 * |
| MING-ZHER POH等: "Advancements in Noncontact, Multiparameter Physiological Measurements Using a Webcam", 《IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING》 * |
| MING-ZHER POH等: "Non-contact, automated cardiac pulse measurements using video imaging and blind source separation", 《HTTPS://DOI.ORG/10.1364/OE.18.010762》 * |
| TAKESHI MITA等: "Joint Haar-like Features for Face Detection", 《TENTH IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION》 * |
| 王利文等: "基于支持向量机的云图自动识别和提取方法", 《天文学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110279406A (en) * | 2019-05-06 | 2019-09-27 | 苏宁金融服务(上海)有限公司 | A kind of touchless pulse frequency measurement method and device based on camera |
| CN114402359A (en) * | 2019-07-18 | 2022-04-26 | 纽洛斯公司 | System and method for detecting composite video of person |
| US11676690B2 (en) | 2019-07-18 | 2023-06-13 | Nuralogix Corporation | System and method for detection of synthesized videos of humans |
| CN114402359B (en) * | 2019-07-18 | 2023-11-17 | 纽洛斯公司 | System and method for detecting a composite video of a person |
| CN111311549A (en) * | 2020-01-20 | 2020-06-19 | 中国人民解放军国防科技大学 | Image authentication method, system and storage medium |
| CN111797735A (en) * | 2020-06-22 | 2020-10-20 | 深圳壹账通智能科技有限公司 | Face video recognition method, device, equipment and storage medium |
| CN111783644A (en) * | 2020-06-30 | 2020-10-16 | 百度在线网络技术（北京）有限公司 | Detection method, device, equipment and computer storage medium |
| CN113963427A (en) * | 2021-12-22 | 2022-01-21 | 浙江工商大学 | Method and system for rapid in vivo detection |
| CN115410260A (en) * | 2022-09-05 | 2022-11-29 | 中国科学院自动化研究所 | Face counterfeit discrimination and evidence obtaining method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109558813A (en) | A kind of AI depth based on pulse signal is changed face video evidence collecting method | |
| Jee et al. | Liveness detection for embedded face recognition system | |
| CN108921041A (en) | A kind of biopsy method and device based on RGB and IR binocular camera | |
| CN111967427A (en) | Fake face video identification method, system and readable storage medium | |
| CN105518710B (en) | Video detecting method, video detection system and computer program product | |
| CN108446690B (en) | Human face in-vivo detection method based on multi-view dynamic features | |
| CN105975938A (en) | Smart community manager service system with dynamic face identification function | |
| Monwar et al. | Pain recognition using artificial neural network | |
| CN110287918A (en) | Vivo identification method and Related product | |
| CN107862298B (en) | Winking living body detection method based on infrared camera device | |
| CN109409343A (en) | A kind of face identification method based on In vivo detection | |
| Das et al. | A framework for liveness detection for direct attacks in the visible spectrum for multimodal ocular biometrics | |
| Shivakumara et al. | A new RGB based fusion for forged IMEI number detection in mobile images | |
| CN117095471B (en) | Face counterfeiting tracing method based on multi-scale characteristics | |
| CN106709480B (en) | Intersected human face recognition methods based on weighed intensities PCNN models | |
| CN109522865A (en) | A kind of characteristic weighing fusion face identification method based on deep neural network | |
| Bhaskar et al. | Advanced algorithm for gender prediction with image quality assessment | |
| HR et al. | A novel hybrid biometric software application for facial recognition considering uncontrollable environmental conditions | |
| Sikander et al. | Facial feature detection: A facial symmetry approach | |
| CN110135362A (en) | A kind of fast face recognition method based under infrared camera | |
| CN112861588A (en) | Living body detection method and device | |
| Das et al. | Face liveness detection based on frequency and micro-texture analysis | |
| El-Sayed et al. | An identification system using eye detection based on wavelets and neural networks | |
| Salih et al. | Color model based convolutional neural network for image spam classification | |
| Yu et al. | Research on face anti-spoofing algorithm based on image fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190402 |
|
| RJ01 | Rejection of invention patent application after publication |