CN107480617A - A kind of adaptive unit analysis method and system of Face Detection - Google Patents

A kind of adaptive unit analysis method and system of Face Detection Download PDF

Info

Publication number
CN107480617A
CN107480617A CN201710650617.1A CN201710650617A CN107480617A CN 107480617 A CN107480617 A CN 107480617A CN 201710650617 A CN201710650617 A CN 201710650617A CN 107480617 A CN107480617 A CN 107480617A
Authority
CN
China
Prior art keywords
gmb
present image
block
size
colourity
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
Application number
CN201710650617.1A
Other languages
Chinese (zh)
Other versions
CN107480617B (en
Inventor
舒倩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen mengwang video Co., Ltd
Original Assignee
Shenzhen Monternet Encyclopedia Information Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Monternet Encyclopedia Information Technology Co Ltd filed Critical Shenzhen Monternet Encyclopedia Information Technology Co Ltd
Priority to CN201710650617.1A priority Critical patent/CN107480617B/en
Publication of CN107480617A publication Critical patent/CN107480617A/en
Application granted granted Critical
Publication of CN107480617B publication Critical patent/CN107480617B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06V40/161—Detection; Localisation; Normalisation
    • G06V40/162—Detection; Localisation; Normalisation using pixel segmentation or colour matching
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • 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
    • G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Image Analysis (AREA)

Abstract

The present invention proposes a kind of adaptive unit analysis method and system of Face Detection.The inventive method devises a kind of adaptive Face Detection that should determine that method, complete image of the Face Detection unit based on colorimetric analysis according to the characteristics of Face Detection, then by the temporal correlation between image, completes the Face Detection of associated subsequent picture.So as to set suitable piecemeal size for the video without compression information, while boosting algorithm performs speed, ensure higher judgment accuracy.

Description

A kind of adaptive unit analysis method and system of Face Detection
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of adaptive unit analysis method of Face Detection and it is System.
Background technology
With developing rapidly for multimedia technology and computer networking technology, the main flow that video is increasingly becoming information propagation carries One of body.Either face video retrieval or Online Video U.S. face, accurate quickly Face Detection technology can all strengthen its thing The effect of half work(times.
If, can be accurate although judging using the unified determining method based on pixel, due to judging language in algorithm performs The speed of service of sentence is far longer than the speed of conventional addition subtraction multiplication and division, large-scale using sentence is judged, will greatly reduce calculation The execution speed of method, it is this in high definition, the video image application of the big resolution ratio of ultra high-definition so as to influence the ageing of algorithm Negative effect is especially prominent.
If, can be with the speed of service of boosting algorithm using unified block-based determining method.Notice and actually should In, often scene is more complicated, exist more people, one, situations such as different resolution.The block division of solidification, can not meet reality The complex situations of border application.
The content of the invention
The purpose of the embodiment of the present invention is to propose a kind of Face Detection unit analysis method based on graphical analysis, it is intended to Solve the problems, such as that either efficiency is low or accuracy is low for prior art Face Detection technology.
The embodiment of the present invention is achieved in that a kind of adaptive unit analysis method of Face Detection, and methods described includes Following steps：
Step0：Make t=1；T represents video frame number, also into current frame number；
Step1：The size of Generalized Block is determined according to present image resolution ratio；
Step2:Calculate the first colourity intensive variable Inuk, InukRepresent the u colourity intensive variables of present image；
Step3：If Inuk>=Thres, then sizef (gmb are set firstk)=1, then sets k=k+1, subsequently enters Step6；Otherwise, then into Step4；
Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；gmbkRepresent present image k-th just Square block, Thres represent threshold value, Thres≤size (gmbk), size (gmbk) represent Generalized Block one-dimensional size；
Step4：Calculate the second colourity intensive variable Invk, InvkRepresent the v colourity intensive variables of present image；
Step5：If Invk<Thres, then sizef (gmb are set firstk)=size (gmbk), then k=k+1 is set； Otherwise, then sizef (gmb are set firstk)=1, then sets k=k+1；
Step6：If k≤num, reenter Step2；Otherwise, into Step7；
Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block；
Step7：If next two field picture of present image is not present, terminate；Otherwise, then t=t+1 is made, and will be current The next frame of image is arranged to present image, subsequently into Step8；
Step8：Size (gmb are pressed firstk) size the former frame of present image and present image are divided it is blocking, so The block statistical variable of each block is calculated afterwards；
Step9：Give present image pictAll pieces make time similarity mark；pictVideo t frames are represented, also referred to as Present image；
Step10：If the time similarity identifier of all pieces of present image is 0, Step2 is reentered；It is no Then, then into Step11；
Step11：First according to the time similarity identifier of block, the Face Detection unit of all pieces of present image of calculating One-dimensional size, then, reenter Step7.
Another object of the present invention is to propose a kind of adaptive unit analysis system of Face Detection, the system includes：
Frame number setup module, for Step0：Make t=1, pictRepresent video t frames, also referred to as present image；
Broad sense block size confirms module, for determining the size of Generalized Block according to present image resolution ratio；
Wherein, QVGA, VGA, 720P are disclosed graphics standard size in the industry；gmbkRepresent k-th of pros of present image Shape block, referred to as Generalized Block, k initial value is 1；size(gmbk) represent Generalized Block one-dimensional size；
First colourity intensive variable computing module, for calculating the first colourity intensive variable, Inuk=std (u (i, j) | u (i,j)∈gmbk)；
Wherein, u (i, j) represents the u chromatic values that present image arranges in the i-th row jth；Std represents to seek mean square deviation；InukRepresent The u colourity intensive variables of present image；
First colourity intensive variable threshold decision processing module, if for judging Inuk>=Thres, then set first sizef(gmbk)=1, then sets k=k+1, subsequently enters judging treatmenting module；Otherwise, then become into the second colourity intensity Measure computing module；
Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；Thres represents threshold value, Thres ≤size(gmbk)；
Second colourity intensive variable computing module, for the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈gmbk)；
Wherein, v (i, j) represents the v chromatic values that present image arranges in the i-th row jth respectively；InvkRepresent the v of present image Colourity intensive variable；
Second colourity intensive variable threshold decision processing module, if for judging Invk<Thres, then set first sizef(gmbk)=size (gmbk), then k=k+1 is set；Otherwise, then sizef (gmb are set firstk)=1, then sets k =k+1；
Judging treatmenting module, if for judging k≤num, into the first colourity intensive variable computing module；Otherwise, Into the next two field picture judging treatmenting module of present image；
Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block；
The next two field picture judging treatmenting module of present image, if for judging that next two field picture of present image is not deposited Then terminating；Otherwise, then t=t+1 is made, and the next frame of present image is arranged to present image, counts and becomes subsequently into block Measure computing module；
Block statistical variable computing module, for pressing size (gmb firstk) size by the former frame of present image and current Image divides blocking, then calculates the block statistical variable of each block
tik=std (y (i, j)-py (i, j) | y (i, j) ∈ gmbkAnd py (i, j) ∈ pgmbk)；
Wherein, y (i, j) is pictThe brightness value of i-th row jth row；Py (i, j) is pictThe i-th row of former frame jth row Brightness value；pgmbkRepresent pictK-th of square block of former frame；tikRepresent gmbkBlock statistical variable；
Time similarity mark module, for making time similarity mark to all pieces of present image；
Specific method is as follows：
Wherein, Thres1Represent the first decision threshold, Thres1=8* (1+24/fps), fps represent video acquisition frame rate； notekRepresent gmbkTime similarity identifier；
Time similarity identifier judge module, if the time similarity identifier for judging all pieces of present image It is 0, then reenters the first colourity intensive variable computing module；Otherwise, then into the one-dimensional size meter of block Face Detection unit Calculate module；
The one-dimensional Size calculation module of block Face Detection unit, for according to the time similarity identifier of block, calculating first The one-dimensional size of the Face Detection unit of all pieces of present image, then, reenter at the next two field picture judgement of present image Manage module.
Beneficial effects of the present invention
The present invention proposes a kind of adaptive unit analysis method of Face Detection.The inventive method is according to the spy of Face Detection Point, a kind of adaptive Face Detection that should determine that method, complete image of the Face Detection unit based on colorimetric analysis is devised, then By the temporal correlation between image, the Face Detection of associated subsequent picture is completed.So as to for without the video for compressing information Suitable piecemeal size is set, while boosting algorithm performs speed, ensures higher judgment accuracy.
Brief description of the drawings
Fig. 1 is a kind of adaptive unit analysis method flow diagram of Face Detection of the preferred embodiment of the present invention；
Fig. 2 is a kind of adaptive unit analysis system construction drawing of Face Detection of the preferred embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and examples The present invention is further elaborated, and for convenience of description, illustrate only the part related to the embodiment of the present invention.It should manage Solution, the specific embodiment that this place is described, it is used only for explaining the present invention, is not intended to limit the invention.
The present invention proposes a kind of adaptive unit analysis method and system of Face Detection.The inventive method is according to Face Detection The characteristics of, a kind of adaptive Face Detection that should determine that method, complete image of the Face Detection unit based on colorimetric analysis is devised, Then by the temporal correlation between image, the Face Detection of associated subsequent picture is completed.So as to for without compression information Video sets suitable piecemeal size, while boosting algorithm performs speed, ensures higher judgment accuracy.
Embodiment one
Fig. 1 is a kind of adaptive unit analysis method flow diagram of Face Detection of the preferred embodiment of the present invention；Methods described bag Include following steps：
Step0：Make t=1, pictRepresent video t frames, also referred to as present image.
Step1：The size of Generalized Block is determined according to present image resolution ratio.
Wherein, QVGA, VGA, 720P are disclosed graphics standard size in the industry；gmbkRepresent k-th of pros of present image Shape block, referred to as Generalized Block, k initial value is 1；size(gmbk) represent Generalized Block one-dimensional size.
Step2:Calculate the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈ gmbk)。
Wherein, u (i, j) represents the u chromatic values that present image arranges in the i-th row jth；Std represents to seek mean square deviation；InukRepresent The u colourity intensive variables of present image.
Step3：If Inuk>=Thres, then sizef (gmb are set firstk)=1, then sets k=k+1, subsequently enters Step6；Otherwise, then into Step4.
Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；Thres represents threshold value, typically Take Thres≤size (gmbk)。
Step4：Calculate the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)。
Wherein, v (i, j) represents the v chromatic values that present image arranges in the i-th row jth respectively；InvkRepresent the v of present image Colourity intensive variable.
Step5：If Invk<Thres, then sizef (gmb are set firstk)=size (gmbk), then k=k+1 is set； Otherwise, then sizef (gmb are set firstk)=1, then sets k=k+1.
Wherein, the selection of first the second colourity of colourity is random, i.e. the first colourity can be u colourities, and the second colourity is v colourities, It can be v colourities that the first colourity, which can also be set, and the second colourity is u colourities, if second of method to set up of selection, as long as will be above-mentioned Step u, v is exchanged.
Step6：If k≤num, reenter Step2；Otherwise, into Step7.
Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block.
Step7：If next two field picture of present image is not present, terminate；Otherwise, then t=t+1 is made, and will be current The next frame of image is arranged to present image, subsequently into Step8.
Step8：Size (gmb are pressed firstk) size the former frame of present image and present image are divided it is blocking, so The block statistical variable of each block is calculated afterwards
tik=std (y (i, j)-py (i, j) | y (i, j) ∈ gmbkAnd py (i, j) ∈ pgmbk)。
Wherein, y (i, j) is pictThe brightness value of i-th row jth row；Py (i, j) is pictThe i-th row of former frame jth row Brightness value；pgmbkRepresent pictK-th of square block of former frame；tikRepresent gmbkBlock statistical variable.
Step9：All pieces to present image are made time similarity mark, and specific method is as follows：
Wherein, Thres1The first decision threshold is represented, can use Thres1=8* (1+24/fps), fps represent video acquisition frame Rate；notekRepresent gmbkTime similarity identifier.
Step10：If the time similarity identifier of all pieces of present image is 0, Step2 is reentered；It is no Then, then into Step11.
Step11：First according to the time similarity identifier of block, the Face Detection unit of all pieces of present image of calculating One-dimensional size, then, reenter Step7.
Pattern one：Suitable for notek=1 block
If notek=1, then sizef (gmb are setk)=sizef (pgmbk)
Wherein, sizef (pgmbk) represent pgmbkThe one-dimensional size of Face Detection unit.
Step A：Calculate the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈ gmbk)。
Step B：If Inuk>=Thres, then sizef (gmb are setk)=1；Otherwise, then into step C.
Step C：Calculate the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)。
Step D：If Invk<Thres, then sizef (gmb are setk)=size (gmbk)；Otherwise, then sizef is set (gmbk)=1.
Embodiment two
Fig. 2 is a kind of adaptive unit analysis system construction drawing of Face Detection of the preferred embodiment of the present invention, the system bag Include：
Frame number setup module, for Step0：Make t=1, pictRepresent video t frames, also referred to as present image；
Broad sense block size confirms module, for determining the size of Generalized Block according to present image resolution ratio；
Wherein, QVGA, VGA, 720P are disclosed graphics standard size in the industry；gmbkRepresent k-th of pros of present image Shape block, referred to as Generalized Block, k initial value is 1；size(gmbk) represent Generalized Block one-dimensional size；
First colourity intensive variable computing module, for calculating the first colourity intensive variable, Inuk=std (u (i, j) | u (i,j)∈gmbk)；
Wherein, u (i, j) represents the u chromatic values that present image arranges in the i-th row jth；Std represents to seek mean square deviation；InukRepresent The u colourity intensive variables of present image；
First colourity intensive variable threshold decision processing module, if for judging Inuk>=Thres, then set first sizef(gmbk)=1, then sets k=k+1, subsequently enters judging treatmenting module；Otherwise, then become into the second colourity intensity Measure computing module；
Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；Thres represents threshold value, typically Take Thres≤size (gmbk)；
Second colourity intensive variable computing module, for the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈gmbk)；
Wherein, v (i, j) represents the v chromatic values that present image arranges in the i-th row jth respectively；InvkRepresent the v of present image Colourity intensive variable；
Second colourity intensive variable threshold decision processing module, if for judging Invk<Thres, then set first sizef(gmbk)=size (gmbk), then k=k+1 is set；Otherwise, then sizef (gmb are set firstk)=1, then sets k =k+1；
Wherein, the selection of first the second colourity of colourity is random, i.e. the first colourity can be u colourities, and the second colourity is v colourities, It can be v colourities that the first colourity, which can also be set, and the second colourity is u colourities, if second of method to set up of selection, as long as will be above-mentioned Step u, v is exchanged；
Judging treatmenting module, if for judging k≤num, into the first colourity intensive variable computing module；Otherwise, Into the next two field picture judging treatmenting module of present image；
Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block；
The next two field picture judging treatmenting module of present image, if for judging that next two field picture of present image is not deposited Then terminating；Otherwise, then t=t+1 is made, and the next frame of present image is arranged to present image, counts and becomes subsequently into block Measure computing module；
Block statistical variable computing module, for pressing size (gmb firstk) size by the former frame of present image and current Image divides blocking, then calculates the block statistical variable of each block
tik=std (y (i, j)-py (i, j) | y (i, j) ∈ gmbkAnd py (i, j) ∈ pgmbk)；
Wherein, y (i, j) is pictThe brightness value of i-th row jth row；Py (i, j) is pictThe i-th row of former frame jth row Brightness value；pgmbkRepresent pictK-th of square block of former frame；tikRepresent gmbkBlock statistical variable；
Time similarity mark module, for making time similarity mark to all pieces of present image, specific method is such as Under：
Wherein, Thres1The first decision threshold is represented, can use Thres1=8* (1+24/fps), fps represent video acquisition frame Rate；notekRepresent gmbkTime similarity identifier；
Time similarity identifier judge module, if the time similarity identifier for judging all pieces of present image It is 0, then reenters the first colourity intensive variable computing module；Otherwise, then into the one-dimensional size meter of block Face Detection unit Calculate module.
The one-dimensional Size calculation module of block Face Detection unit, for according to the time similarity identifier of block, calculating first The one-dimensional size of the Face Detection unit of all pieces of present image, then, reenter at the next two field picture judgement of present image Manage module.
Pattern one：Suitable for notek=1 block
If notek=1, then sizef (gmb are setk)=sizef (pgmbk)
Wherein, sizef (pgmbk) represent pgmbkThe one-dimensional size of Face Detection unit.
Pattern two：Suitable for notek=0 block
Step A：Calculate the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈ gmbk)。
Step B：If Inuk>=Thres, then sizef (gmb are setk)=1；Otherwise, then into step C.
Step C：Calculate the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)。
Step D：If Invk<Thres, then sizef (gmb are setk)=size (gmbk)；Otherwise, then sizef is set (gmbk)=1.
Can it will be understood by those skilled in the art that realizing that all or part of step in above-described embodiment method is So that by programmed instruction related hardware, come what is completed, described program can be stored in a computer read/write memory medium, Described storage medium can be ROM, RAM, disk, CD etc..
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.

Claims (8)

  1. A kind of 1. adaptive unit analysis method of Face Detection, it is characterised in that the described method comprises the following steps：
    Step0：Make t=1；T represents video frame number, also into current frame number；
    Step1：The size of Generalized Block is determined according to present image resolution ratio；
    Step2:Calculate the first colourity intensive variable Inuk, InukRepresent the u colourity intensive variables of present image；
    Step3：If Inuk>=Thres, then sizef (gmb are set firstk)=1, then sets k=k+1, subsequently enters Step6；Otherwise, then into Step4；
    Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；gmbkRepresent k-th of square of present image Block, Thres represent threshold value, Thres≤size (gmbk), size (gmbk) represent Generalized Block one-dimensional size；
    Step4：Calculate the second colourity intensive variable Invk, InvkRepresent the v colourity intensive variables of present image；
    Step5：If Invk<Thres, then sizef (gmb are set firstk)=size (gmbk), then k=k+1 is set；It is no Then, then sizef (gmb are set firstk)=1, then sets k=k+1；
    Step6：If k≤num, reenter Step2；Otherwise, into Step7；
    Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block；
    Step7：If next two field picture of present image is not present, terminate；Otherwise, then t=t+1 is made, and by present image Next frame be arranged to present image, subsequently into Step8；
    Step8：Size (gmb are pressed firstk) size the former frame of present image and present image are divided into blocking, Ran Houji Calculate the block statistical variable of each block；
    Step9：Give present image pictAll pieces make time similarity mark；pictVideo t frames are represented, it is also referred to as current Image；
    Step10：If the time similarity identifier of all pieces of present image is 0, Step2 is reentered；Otherwise, then Into Step11；
    Step11：First according to the time similarity identifier of block, the one of the Face Detection unit of all pieces of present image is calculated Size is tieed up, then, reenters Step7.
  2. 2. the adaptive unit analysis method of Face Detection as claimed in claim 1, it is characterised in that
    It is described Generalized Block is determined according to present image resolution ratio size be specially：
    Wherein, QVGA, VGA, 720P are disclosed graphics standard size in the industry；gmbkK-th of square block of present image is represented, Referred to as Generalized Block, k initial value is 1；size(gmbk) represent Generalized Block one-dimensional size；
    The first colourity intensive variable Inu of the calculatingkSpecially：
    Inuk=std (u (i, j) | u (i, j) ∈ gmbk)；
    Wherein, u (i, j) represents the u chromatic values that present image arranges in the i-th row jth；Std represents to seek mean square deviation；InukRepresent current The u colourity intensive variables of image；
    The second colourity intensive variable Inv of the calculatingkSpecially：
    Invk=std (v (i, j) | v (i, j) ∈ gmbk)；
    Wherein, v (i, j) represents the v chromatic values that present image arranges in the i-th row jth respectively；InvkRepresent that the v colourities of present image are strong Spend variable；
    It is described to press size (gmb firstk) size the former frame of present image and present image are divided blocking, then calculate The block statistical variable of each block is specially：
    tik=std (y (i, j)-py (i, j) | y (i, j) ∈ gmbkAnd py (i, j) ∈ pgmbk)；
    Wherein, y (i, j) is pictThe brightness value of i-th row jth row；Py (i, j) is pictThe i-th row of former frame jth row brightness Value；pgmbkRepresent pictK-th of square block of former frame；tikRepresent gmbkBlock statistical variable.
  3. 3. the adaptive unit analysis method of Face Detection as claimed in claim 2, it is characterised in that
    It is as follows that all pieces to present image make time similarity mark specific method：
    <mrow> <msub> <mi>note</mi> <mi>k</mi> </msub> <mo>=</mo> <mi>s</mi> <mi>i</mi> <mi>g</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> <mo>,</mo> <mi>s</mi> <mi>i</mi> <mi>g</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mn>1</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>&lt;</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>0</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>e</mi> <mi>l</mi> <mi>s</mi> <mi>e</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow>
    Wherein, Thres1Represent the first decision threshold, Thres1=8* (1+24/fps), fps represent video acquisition frame rate；notek Represent gmbkTime similarity identifier.
  4. 4. the adaptive unit analysis method of Face Detection as claimed in claim 3, it is characterised in that
    It is described first according to the time similarity identifier of block, calculate the one-dimensional chi of the Face Detection unit of all pieces of present image It is very little to be specially：
    Pattern one：Suitable for notek=1 block
    If notek=1, then sizef (gmb are setk)=sizef (pgmbk)
    Wherein, sizef (pgmbk) represent pgmbkThe one-dimensional size of Face Detection unit；
    Pattern two：Suitable for notek=0 block
    Step A：Calculate the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈ gmbk)。
    Step B：If Inuk>=Thres, then sizef (gmb are setk)=1；Otherwise, then into step C.
    Step C：Calculate the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)。
    Step D：If Invk<Thres, then sizef (gmb are setk)=size (gmbk)；Otherwise, then sizef (gmb are setk)= 1。
  5. 5. the adaptive unit analysis method of Face Detection as claimed in claim 4, it is characterised in that first colourity second The selection of colourity is random；
    It is v colourities to set the first colourity, and the second colourity is u colourities, and u, v in Step1-Step11 are exchanged.
  6. 6. a kind of adaptive unit analysis system of Face Detection, it is characterised in that the system includes：
    Frame number setup module, for Step0：Make t=1, pictRepresent video t frames, also referred to as present image；
    Broad sense block size confirms module, for determining the size of Generalized Block according to present image resolution ratio；
    Wherein, QVGA, VGA, 720P are disclosed graphics standard size in the industry；gmbkK-th of square block of present image is represented, Referred to as Generalized Block, k initial value is 1；size(gmbk) represent Generalized Block one-dimensional size；
    First colourity intensive variable computing module, for calculating the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈gmbk)；
    Wherein, u (i, j) represents the u chromatic values that present image arranges in the i-th row jth；Std represents to seek mean square deviation；InukRepresent current The u colourity intensive variables of image；
    First colourity intensive variable threshold decision processing module, if for judging Inuk>=Thres, then set sizef first (gmbk)=1, then sets k=k+1, subsequently enters judging treatmenting module；Otherwise, then calculated into the second colourity intensive variable Module；
    Wherein, sizef (gmbk) represent gmbkThe one-dimensional size of Face Detection unit；Thres expression threshold values, Thres≤ size(gmbk)；
    Second colourity intensive variable computing module, for the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)；
    Wherein, v (i, j) represents the v chromatic values that present image arranges in the i-th row jth respectively；InvkRepresent that the v colourities of present image are strong Spend variable；
    Second colourity intensive variable threshold decision processing module, if for judging Invk<Thres, then sizef is set first (gmbk)=size (gmbk), then k=k+1 is set；Otherwise, then sizef (gmb are set firstk)=1, then sets k=k+ 1；
    Judging treatmenting module, if for judging k≤num, into the first colourity intensive variable computing module；Otherwise, enter The next two field picture judging treatmenting module of present image；
    Wherein, num represents the broad sense number of blocks that image is divided by the size of Generalized Block；
    The next two field picture judging treatmenting module of present image, if for judging that next two field picture of present image is not present, Terminate；Otherwise, then t=t+1 is made, and the next frame of present image is arranged to present image, subsequently into block statistical variable meter Calculate module；
    Block statistical variable computing module, for pressing size (gmb firstk) size by the former frame and present image of present image Divide blocking, then calculate the block statistical variable ti of each blockk=std (y (i, j)-py (i, j) | y (i, j) ∈ gmbkAnd py(i,j)∈pgmbk)；
    Wherein, y (i, j) is pictThe brightness value of i-th row jth row；Py (i, j) is pictThe i-th row of former frame jth row brightness Value；pgmbkRepresent pictK-th of square block of former frame；tikRepresent gmbkBlock statistical variable；
    Time similarity mark module, for making time similarity mark to all pieces of present image；
    Specific method is as follows：
    <mrow> <msub> <mi>note</mi> <mi>k</mi> </msub> <mo>=</mo> <mi>s</mi> <mi>i</mi> <mi>g</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> <mo>,</mo> <mi>s</mi> <mi>i</mi> <mi>g</mi> <mi>n</mi> <mrow> <mo>(</mo> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mn>1</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <msub> <mi>ti</mi> <mi>k</mi> </msub> <mo>&lt;</mo> <msub> <mi>Thres</mi> <mn>1</mn> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>0</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>e</mi> <mi>l</mi> <mi>s</mi> <mi>e</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> </mrow>
    Wherein, Thres1Represent the first decision threshold, Thres1=8* (1+24/fps), fps represent video acquisition frame rate；notek Represent gmbkTime similarity identifier；
    Time similarity identifier judge module, if the time similarity identifier for all pieces of present image of judgement is 0, then reenter the first colourity intensive variable computing module；Otherwise, then into the one-dimensional Size calculation mould of block Face Detection unit Block；
    The one-dimensional Size calculation module of block Face Detection unit, for according to the time similarity identifier of block, calculating current first The one-dimensional size of the Face Detection unit of all pieces of image, then, reenter the next two field picture of present image and judge processing mould Block.
  7. 7. the adaptive unit analysis system of Face Detection as claimed in claim 6, it is characterised in that
    Among the one-dimensional Size calculation module of described piece of Face Detection unit, for first according to the time similarity identifier of block, The one-dimensional size of Face Detection unit for calculating all pieces of present image is specially：
    Pattern one：Suitable for notek=1 block
    If notek=1, then sizef (gmb are setk)=sizef (pgmbk)
    Wherein, sizef (pgmbk) represent pgmbkThe one-dimensional size of Face Detection unit；
    Pattern two：Suitable for notek=0 block
    Step A：Calculate the first colourity intensive variable, Inuk=std (u (i, j) | u (i, j) ∈ gmbk)；
    Step B：If Inuk>=Thres, then sizef (gmb are setk)=1；Otherwise, then into step C；
    Step C：Calculate the second colourity intensive variable Invk=std (v (i, j) | v (i, j) ∈ gmbk)；
    Step D：If Invk<Thres, then sizef (gmb are setk)=size (gmbk)；Otherwise, then sizef (gmb are setk)= 1。
  8. 8. the adaptive unit analysis system of Face Detection as claimed in claim 7, it is characterised in that
    It is v colourities to set the first colourity, and the second colourity is u colourities, and u, v in the claims 7 are exchanged.
CN201710650617.1A 2017-08-02 2017-08-02 Skin color detection self-adaptive unit analysis method and system Active CN107480617B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710650617.1A CN107480617B (en) 2017-08-02 2017-08-02 Skin color detection self-adaptive unit analysis method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710650617.1A CN107480617B (en) 2017-08-02 2017-08-02 Skin color detection self-adaptive unit analysis method and system

Publications (2)

Publication Number Publication Date
CN107480617A true CN107480617A (en) 2017-12-15
CN107480617B CN107480617B (en) 2020-03-17

Family

ID=60597502

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710650617.1A Active CN107480617B (en) 2017-08-02 2017-08-02 Skin color detection self-adaptive unit analysis method and system

Country Status (1)

Country Link
CN (1) CN107480617B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1751522A (en) * 2002-12-18 2006-03-22 罗伯特·戈登大学 Video encoding with skipping motion estimation for selected macroblocks
CN101650830A (en) * 2009-08-06 2010-02-17 中国科学院声学研究所 Compressed domain video lens mutation and gradient union automatic segmentation method and system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1751522A (en) * 2002-12-18 2006-03-22 罗伯特·戈登大学 Video encoding with skipping motion estimation for selected macroblocks
CN101650830A (en) * 2009-08-06 2010-02-17 中国科学院声学研究所 Compressed domain video lens mutation and gradient union automatic segmentation method and system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张小红 等: "自适应方向插值及相似性约束人脸区域的H.264视频流错误隐藏研究", 《计算机应用研究》 *
纪晓宁: "基于肤色的人脸检测及其在红眼消除中的应用研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Also Published As

Publication number Publication date
CN107480617B (en) 2020-03-17

Similar Documents

Publication Publication Date Title
US11468585B2 (en) Pseudo RGB-D for self-improving monocular slam and depth prediction
CN110717527B (en) Method for determining target detection model by combining cavity space pyramid structure
CN110909594A (en) Video significance detection method based on depth fusion
CN108012202B (en) Video concentration method, device, computer readable storage medium and computer device
EP2706507A1 (en) Method and apparatus for generating morphing animation
CN110189294B (en) RGB-D image significance detection method based on depth reliability analysis
CN102968814B (en) A kind of method and apparatus of image rendering
KR102669454B1 (en) Activity recognition technique in video image sequences using depth information
CN101236657A (en) Single movement target track tracking and recording method
CN106997478B (en) RGB-D image salient target detection method based on salient center prior
CN110827312B (en) Learning method based on cooperative visual attention neural network
CN102231829B (en) Display format identification method and device of video file as well as video player
WO2014036813A1 (en) Method and device for extracting image features
CN110610486A (en) Monocular image depth estimation method and device
Zha et al. A real-time global stereo-matching on FPGA
CN101582171A (en) Method and device for creating depth maps
CN104185012B (en) 3 D video form automatic testing method and device
CN104065954A (en) Method for quickly detecting parallax scope of high-definition stereoscopic video
Martin et al. Optimal choice of motion estimation methods for fine-grained action classification with 3d convolutional networks
Xu et al. Robust motion compensation for event cameras with smooth constraint
CN108447060A (en) Front and back scape separation method based on RGB-D images and its front and back scene separation device
CN109871790A (en) A kind of video decolorizing method based on hybrid production style
CN101833760A (en) Background modeling method and device based on image blocks
CN114419102B (en) Multi-target tracking detection method based on frame difference time sequence motion information
Hu et al. A low illumination video enhancement algorithm based on the atmospheric physical model

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
CP01 Change in the name or title of a patent holder

Address after: 518057 Guangdong city of Shenzhen province Nanshan District Guangdong streets high in the four Longtaili Technology Building Room 325 No. 30

Patentee after: Shenzhen mengwang video Co., Ltd

Address before: 518057 Guangdong city of Shenzhen province Nanshan District Guangdong streets high in the four Longtaili Technology Building Room 325 No. 30

Patentee before: SHENZHEN MONTNETS ENCYCLOPEDIA INFORMATION TECHNOLOGY Co.,Ltd.

CP01 Change in the name or title of a patent holder