CN105426816A - Method and device of processing face images - Google Patents
Method and device of processing face images Download PDFInfo
- Publication number
- CN105426816A CN105426816A CN201510724126.8A CN201510724126A CN105426816A CN 105426816 A CN105426816 A CN 105426816A CN 201510724126 A CN201510724126 A CN 201510724126A CN 105426816 A CN105426816 A CN 105426816A
- Authority
- CN
- China
- Prior art keywords
- image
- area
- skin
- skin color
- colour
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention discloses a method and device of processing face images, comprising: converting a pending image into an HSV color space from an RGB color space; obtaining the skin color area of the pending image according to the H component threshold in the HSV color space; converting the HSV color space into a skin color likelihood image, and performing binary segmentation on the skin color likelihood image; obtaining a face candidate image having a skin color area and a non skin color area separated according to the skin color likelihood image after binary segmentation and the skin color area of the pending image; and marking the skin color area in the face candidate image. The method and device utilize the chroma, brightness and saturation in the HSV color space to perform sound separation, realize less overlapping between skin color points and non skin color points, and improve detection precision; in addition, the method and device utilize an H-SGM skin color model to eliminate the interference of the non skin color area and most background area, and divide a candidate face area so as to realize high face detection precision and less misdetection.
Description
Technical field
The embodiment of the present invention relates to the technical field of Face datection, particularly relates to a kind of method and the device that process facial image.
Background technology
Face datection, as sport technique segment the most basic and important, has very important effect in face survey region, is not only the basic steps of follow-up face recognition technology, also developed into a research field independently had wide application prospects.Current Face datection starts to be widely applied to many fields such as brand-new man-machine interface, content-based retrieval, Digital Video Processing, vision monitoring.
Utilize separately a certain complexion model to detect face in existing human face detection tech scheme, or utilize separately Haar-like feature classifiers to detect face.Accuracy of detection is low and false drop rate is high to utilize separately a certain complexion model detection face easily to cause.
Summary of the invention
The object of the embodiment of the present invention is to propose a kind of method and the device that process facial image, is intended to solve the problem that accuracy of detection is low and false drop rate is high utilizing separately a kind of complexion model detection face to cause.
For reaching this object, the embodiment of the present invention by the following technical solutions:
First aspect, a kind of method processing facial image, described method comprises:
By pending image from RGB color space conversion to hsv color space;
The area of skin color of described pending image is obtained according to the threshold value of H component in described hsv color space;
Be colour of skin likelihood image by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image;
The face candidate image that area of skin color is separated with non-area of skin color is obtained according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image;
Mark the area of skin color in described face candidate image.
Preferably, the described threshold value according to H component in described hsv color space obtains the area of skin color of described pending image, comprising:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
Preferably, described is colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image, comprising:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
Preferably, the described area of skin color according to the colour of skin likelihood image after binary segmentation and described pending image obtains the face candidate image that area of skin color is separated with non-area of skin color, comprising:
Obtain the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, get rid of non-area of skin color;
According to mathematical morphological operation, elimination hole is carried out to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtain the face candidate image that area of skin color is separated with non-area of skin color.
Preferably, the area of skin color in the described face candidate image of described mark, comprising:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the area of skin color in face candidate image described in precise marking.
Second aspect, a kind of device processing facial image, described device comprises:
Modular converter, for by pending image from RGB color space conversion to hsv color space;
First acquisition module, for obtaining the area of skin color of described pending image according to the threshold value of H component in described hsv color space;
Segmentation module, for being colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image;
Second acquisition module, for obtaining the face candidate image that area of skin color is separated with non-area of skin color according to the colour of skin likelihood image after binary segmentation and the area of skin color of described pending image;
Mark module, for marking the area of skin color in described face candidate image.
Preferably, described first acquisition module, for:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
Preferably, described segmentation module, for:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
Preferably, described second acquisition module, comprising:
Rejected unit, for obtaining the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, gets rid of non-area of skin color;
Acquiring unit, for carrying out elimination hole according to mathematical morphological operation to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtains the face candidate image that area of skin color is separated with non-area of skin color.
Preferably, described mark module, for:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the human face region of the area of skin color of the described pending image after precise marking obtains.
The embodiment of the present invention is passed through pending image from RGB color space conversion to hsv color space; The area of skin color of described pending image is obtained according to the threshold value of H component in described hsv color space; Be colour of skin likelihood image by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image; The face candidate image that area of skin color is separated with non-area of skin color is obtained according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image; Mark the area of skin color in described face candidate image.The present invention utilizes colourity in hsv color space, brightness, saturation degree can be separated that can to realize colour of skin point lessly overlapping with non-colour of skin point well, improves the precision detected; Utilize H-SGM complexion model to get rid of the interference of non-area of skin color and most of background area, be partitioned into the human face region of candidate, thus the effect such as the precision realizing Face datection is high, error detection is low.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of the method for embodiment of the present invention process facial image;
Fig. 2 is the high-level schematic functional block diagram of the embodiment of the present invention second acquisition module;
Fig. 3 is the high-level schematic functional block diagram of the device of embodiment of the present invention process facial image.
Embodiment
Below in conjunction with drawings and Examples, the embodiment of the present invention is described in further detail.Be understandable that, specific embodiment described herein is only for explaining the embodiment of the present invention, but not the restriction to the embodiment of the present invention.It also should be noted that, for convenience of description, illustrate only the part relevant to the embodiment of the present invention in accompanying drawing but not entire infrastructure.
Embodiment one
It is the schematic flow sheet of the method for embodiment of the present invention process facial image with reference to figure 1, Fig. 1.
In embodiment one, the method for described process facial image comprises:
Step 101, by pending image from RGB color space conversion to hsv color space;
Concrete, by the coloured image of camera collection from RGB color space conversion to hsv color space.In RGB color space, three color components all comprise monochrome information and there is correlativity, poor to luminance adaptation.In hsv color space, colourity, brightness, saturation degree can be separated well, and lessly overlapping with non-colour of skin point.
Step 102, obtains the area of skin color of described pending image according to the threshold value of H component in described hsv color space;
Preferably, the described threshold value according to H component in described hsv color space obtains the area of skin color of described pending image, comprising:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
Concrete, separate picture, at three components in hsv color space, is selected H component as skin color segmentation feature, is obtained the image I removing non-area of skin color
1.H component meets threshold range H ∈ [a, b], then think area of skin color, otherwise is non-area of skin color.
H component is better to the area of skin color extent of polymerization of image, and the mode experiencing color with people is also consistent, can reflect the color attributes of target more accurately, can reduce the susceptibility that image changes brightness.
A, b are empirical value, carry out statistics can obtain skin distribution region to H component.
Step 103, is colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image;
Preferably, described is colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image, comprising:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
Concrete, utilize single Gauss's complexion model coloured image to be converted into colour of skin likelihood image, binary segmentation is carried out to colour of skin likelihood image, obtains image I
2.
Step 104, obtains according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image the face candidate image that area of skin color is separated with non-area of skin color;
Preferably, the described area of skin color according to the colour of skin likelihood image after binary segmentation and described pending image obtains the face candidate image that area of skin color is separated with non-area of skin color, comprising:
Obtain the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, get rid of non-area of skin color;
According to mathematical morphological operation, elimination hole is carried out to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtain the face candidate image that area of skin color is separated with non-area of skin color.
Concrete, by the image I obtained above-mentioned steps with computing, mathematical morphological operation etc.
1with I
2process.
Concrete, by with computing, extract the common ground of two kinds of complexion model testing results, remove the region of result difference, get rid of non-area of skin color; Utilize mathematical morphological operation to carry out the operation such as elimination hole, few pixel connected region elimination to image, obtain the face candidate image of the background separation such as skin area and the non-colour of skin.
Above-mentioned colour of skin pre-treatment step can called after H-SGM complexion model.
Step 105, marks the area of skin color in described face candidate image.
Preferably, the area of skin color in the described face candidate image of described mark, comprising:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the area of skin color in face candidate image described in precise marking.
Concrete, candidate region is inputed to multiple dimensioned piece of local binary patterns feature (textural characteristics conventional in a kind of image procossing) cascade classifier, whether accurate detection exists human face region and marks face.Carry out with cascade classifier scanning the region that may there is face, without the need to traveling through all subwindows of entire image, decreasing search volume, improve the effect of Face datection, improve detection speed, also reduce false drop rate and loss.
The embodiment of the present invention is passed through pending image from RGB color space conversion to hsv color space; The area of skin color of described pending image is obtained according to the threshold value of H component in described hsv color space; Be colour of skin likelihood image by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image; The face candidate image that area of skin color is separated with non-area of skin color is obtained according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image; Mark the area of skin color in described face candidate image.The present invention utilizes colourity in hsv color space, brightness, saturation degree can be separated that can to realize colour of skin point lessly overlapping with non-colour of skin point well, improves the precision detected; Utilize H-SGM complexion model to get rid of the interference of non-area of skin color and most of background area, be partitioned into the human face region of candidate, thus the effect such as the precision realizing Face datection is high, error detection is low.
Embodiment two
It is the high-level schematic functional block diagram of the device of embodiment of the present invention process facial image with reference to figure 2, Fig. 2.
In embodiment two, the device of described process facial image comprises:
Modular converter 201, for by pending image from RGB color space conversion to hsv color space;
Concrete, by the coloured image of camera collection from RGB color space conversion to hsv color space.In RGB color space, three color components all comprise monochrome information and there is correlativity, poor to luminance adaptation.In hsv color space, colourity, brightness, saturation degree can be separated well, and lessly overlapping with non-colour of skin point.
First acquisition module 202, for obtaining the area of skin color of described pending image according to the threshold value of H component in described hsv color space;
Preferably, described first acquisition module 202, for:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
Concrete, separate picture, at three components in hsv color space, is selected H component as skin color segmentation feature, is obtained the image I removing non-area of skin color
1.H component meets threshold range H ∈ [a, b], then think area of skin color, otherwise is non-area of skin color.
H component is better to the area of skin color extent of polymerization of image, and the mode experiencing color with people is also consistent, can reflect the color attributes of target more accurately, can reduce the susceptibility that image changes brightness.
A, b are empirical value, carry out statistics can obtain skin distribution region to H component.
Segmentation module 203, for being colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image;
Preferably, described segmentation module 203, for:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
Concrete, utilize single Gauss's complexion model coloured image to be converted into colour of skin likelihood image, binary segmentation is carried out to colour of skin likelihood image, obtains image I
2.
Second acquisition module 204, for obtaining the face candidate image that area of skin color is separated with non-area of skin color according to the colour of skin likelihood image after binary segmentation and the area of skin color of described pending image;
Preferably, with reference to figure 3, Fig. 3 be the high-level schematic functional block diagram of the embodiment of the present invention second acquisition module 204.
Described second acquisition module 204 comprises:
Rejected unit 301, for obtaining the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, gets rid of non-area of skin color;
Acquiring unit 302, for carrying out elimination hole according to mathematical morphological operation to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtains the face candidate image that area of skin color is separated with non-area of skin color.
Concrete, by the image I obtained above-mentioned steps with computing, mathematical morphological operation etc.
1with I
2process.
Concrete, by with computing, extract the common ground of two kinds of complexion model testing results, remove the region of result difference, get rid of non-area of skin color; Utilize mathematical morphological operation to carry out the operation such as elimination hole, few pixel connected region elimination to image, obtain the face candidate image of the background separation such as skin area and the non-colour of skin.
Above-mentioned colour of skin pre-treatment step can called after H-SGM complexion model.
Mark module 205, for marking the area of skin color in described face candidate image.
Preferably, described mark module 205, for:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the area of skin color in face candidate image described in precise marking.
Concrete, candidate region is inputed to multiple dimensioned piece of local binary patterns feature (textural characteristics conventional in a kind of image procossing) cascade classifier, whether accurate detection exists human face region and marks face.Carry out with cascade classifier scanning the region that may there is face, without the need to traveling through all subwindows of entire image, decreasing search volume, improve the effect of Face datection, improve detection speed, also reduce false drop rate and loss.
The embodiment of the present invention is passed through pending image from RGB color space conversion to hsv color space; The area of skin color of described pending image is obtained according to the threshold value of H component in described hsv color space; Be colour of skin likelihood image by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image; The face candidate image that area of skin color is separated with non-area of skin color is obtained according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image; Mark the area of skin color in described face candidate image.The present invention utilizes colourity in hsv color space, brightness, saturation degree can be separated that can to realize colour of skin point lessly overlapping with non-colour of skin point well, improves the precision detected; Utilize H-SGM complexion model to get rid of the interference of non-area of skin color and most of background area, be partitioned into the human face region of candidate, thus the effect such as the precision realizing Face datection is high, error detection is low.
Below the know-why of the embodiment of the present invention is described in conjunction with specific embodiments.These describe the principle just in order to explain the embodiment of the present invention, and can not be interpreted as the restriction to embodiment of the present invention protection domain by any way.Based on explanation herein, those skilled in the art does not need to pay other embodiment that performing creative labour can associate the embodiment of the present invention, these modes all by fall into the embodiment of the present invention protection domain within.
Claims (10)
1. process a method for facial image, it is characterized in that, described method comprises:
By pending image from RGB color space conversion to hsv color space;
The area of skin color of described pending image is obtained according to the threshold value of H component in described hsv color space;
Be colour of skin likelihood image by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image;
The face candidate image that area of skin color is separated with non-area of skin color is obtained according to the area of skin color of the colour of skin likelihood image after binary segmentation and described pending image;
Mark the area of skin color in described face candidate image.
2. method according to claim 1, is characterized in that, the described threshold value according to H component in described hsv color space obtains the area of skin color of described pending image, comprising:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
3. method according to claim 1, is characterized in that, described is colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image, comprising:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
4. method according to claim 1, is characterized in that, the described area of skin color according to the colour of skin likelihood image after binary segmentation and described pending image obtains the face candidate image that area of skin color is separated with non-area of skin color, comprising:
Obtain the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, get rid of non-area of skin color;
According to mathematical morphological operation, elimination hole is carried out to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtain the face candidate image that area of skin color is separated with non-area of skin color.
5. the method according to Claims 1-4 any one, is characterized in that, the area of skin color in the described face candidate image of described mark, comprising:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the area of skin color in face candidate image described in precise marking.
6. process a device for facial image, it is characterized in that, described device comprises:
Modular converter, for by pending image from RGB color space conversion to hsv color space;
First acquisition module, for obtaining the area of skin color of described pending image according to the threshold value of H component in described hsv color space;
Segmentation module, for being colour of skin likelihood image by described hsv color space transforming, and carries out binary segmentation to described colour of skin likelihood image;
Second acquisition module, for obtaining the face candidate image that area of skin color is separated with non-area of skin color according to the colour of skin likelihood image after binary segmentation and the area of skin color of described pending image;
Mark module, for marking the area of skin color in described face candidate image.
7. device according to claim 6, is characterized in that, described first acquisition module, for:
If described H component meets threshold range H ∈ [a, b], then corresponding with described H component pending image is area of skin color.
8. device according to claim 6, is characterized in that, described segmentation module, for:
Be colour of skin likelihood image by single Gauss's complexion model by described hsv color space transforming, and binary segmentation is carried out to described colour of skin likelihood image.
9. device according to claim 6, is characterized in that, described second acquisition module comprises:
Rejected unit, for obtaining the common ground of the area of skin color of the colour of skin likelihood image after described binary segmentation and described pending image, gets rid of non-area of skin color;
Acquiring unit, for carrying out elimination hole according to mathematical morphological operation to the image after the non-area of skin color of eliminating, few pixel connected region eliminates operation, obtains the face candidate image that area of skin color is separated with non-area of skin color.
10. the device according to claim 6 to 9 any one, is characterized in that, described mark module, for:
By multiple dimensioned piecemeal local binary aspect of model cascade classifier, the area of skin color in face candidate image described in precise marking.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510724126.8A CN105426816A (en) | 2015-10-29 | 2015-10-29 | Method and device of processing face images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510724126.8A CN105426816A (en) | 2015-10-29 | 2015-10-29 | Method and device of processing face images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105426816A true CN105426816A (en) | 2016-03-23 |
Family
ID=55505016
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510724126.8A Pending CN105426816A (en) | 2015-10-29 | 2015-10-29 | Method and device of processing face images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105426816A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845115A (en) * | 2017-09-13 | 2018-03-27 | 捷开通讯(深圳)有限公司 | Intelligent mobile terminal and its eye care method, the device with store function |
| CN108701355A (en) * | 2016-03-25 | 2018-10-23 | 英特尔公司 | GPU optimizes and the skin possibility predication based on single Gauss online |
| WO2020052525A1 (en) * | 2018-09-10 | 2020-03-19 | 深圳市前海安测信息技术有限公司 | Facial glossiness classification device and method, and computer storage medium |
| CN111046763A (en) * | 2019-11-29 | 2020-04-21 | 广州久邦世纪科技有限公司 | Portrait cartoon method and device |
| WO2020199475A1 (en) * | 2019-04-03 | 2020-10-08 | 平安科技（深圳）有限公司 | Facial recognition method and apparatus, computer device and storage medium |
| CN111815651A (en) * | 2020-07-08 | 2020-10-23 | 深圳市梦网视讯有限公司 | Method, system and equipment for segmenting face and body skin color area |
| CN113076801A (en) * | 2021-03-04 | 2021-07-06 | 广州铁路职业技术学院（广州铁路机械学校） | Train on-road state intelligent linkage detection system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289680A (en) * | 2011-09-02 | 2011-12-21 | 北京新媒传信科技有限公司 | Cutting method and device for skin color area in image |
| CN102324036A (en) * | 2011-09-02 | 2012-01-18 | 北京新媒传信科技有限公司 | Obtain the method and apparatus of face complexion area in the image |
-
2015
- 2015-10-29 CN CN201510724126.8A patent/CN105426816A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289680A (en) * | 2011-09-02 | 2011-12-21 | 北京新媒传信科技有限公司 | Cutting method and device for skin color area in image |
| CN102324036A (en) * | 2011-09-02 | 2012-01-18 | 北京新媒传信科技有限公司 | Obtain the method and apparatus of face complexion area in the image |
Non-Patent Citations (6)
| Title |
|---|
| 丁娜娜: "面向任务的仿人机器人室内定位导航技术", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 叶剑华: "基于LBP和Fisherfaces的多模态人脸识别", 《计算机工程》 * |
| 李伟生 等: "基于多肤色空间及Adaboost算法的人脸检测方法", 《计算机应用研究》 * |
| 杨春友: "基于肤色分割和PCA的支持向量新颖检测的人脸检测研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 段新涛 等: "基于肤色和图像似然度的人脸检测", 《中国科技信息》 * |
| 谢淑丽: "基于双肤色模型及Adaboost算法的人脸检测", 《莆田学院学报》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108701355A (en) * | 2016-03-25 | 2018-10-23 | 英特尔公司 | GPU optimizes and the skin possibility predication based on single Gauss online |
| CN108701355B (en) * | 2016-03-25 | 2023-07-28 | 英特尔公司 | GPU optimization and online single Gaussian-based skin likelihood estimation |
| CN107845115A (en) * | 2017-09-13 | 2018-03-27 | 捷开通讯(深圳)有限公司 | Intelligent mobile terminal and its eye care method, the device with store function |
| WO2020052525A1 (en) * | 2018-09-10 | 2020-03-19 | 深圳市前海安测信息技术有限公司 | Facial glossiness classification device and method, and computer storage medium |
| WO2020199475A1 (en) * | 2019-04-03 | 2020-10-08 | 平安科技（深圳）有限公司 | Facial recognition method and apparatus, computer device and storage medium |
| CN111046763A (en) * | 2019-11-29 | 2020-04-21 | 广州久邦世纪科技有限公司 | Portrait cartoon method and device |
| CN111046763B (en) * | 2019-11-29 | 2024-03-29 | 广州久邦世纪科技有限公司 | Portrait cartoon method and device |
| CN111815651A (en) * | 2020-07-08 | 2020-10-23 | 深圳市梦网视讯有限公司 | Method, system and equipment for segmenting face and body skin color area |
| CN111815651B (en) * | 2020-07-08 | 2024-01-30 | 深圳市梦网视讯有限公司 | Method, system and equipment for segmenting human face and body skin color region |
| CN113076801A (en) * | 2021-03-04 | 2021-07-06 | 广州铁路职业技术学院（广州铁路机械学校） | Train on-road state intelligent linkage detection system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105426816A (en) | Method and device of processing face images | |
| Yuan et al. | Robust traffic sign recognition based on color global and local oriented edge magnitude patterns | |
| CN102194108B (en) | Smile face expression recognition method based on clustering linear discriminant analysis of feature selection | |
| CN103020992B (en) | A kind of video image conspicuousness detection method based on motion color-associations | |
| CN105160691A (en) | Color histogram based vehicle body color identification method | |
| CN105844242A (en) | Method for detecting skin color in image | |
| CN104123529A (en) | Human hand detection method and system thereof | |
| CN104951742A (en) | Detection method and system for sensitive video | |
| CN104463138A (en) | Text positioning method and system based on visual structure attribute | |
| CN104657724A (en) | Method for detecting pedestrians in traffic videos | |
| Li et al. | Automatic passenger counting system for bus based on RGB-D video | |
| Zhang | Vehicle target detection methods based on color fusion deformable part model | |
| Qin et al. | Traffic sign segmentation and recognition in scene images | |
| CN104866826A (en) | Static gesture language identification method based on KNN algorithm and pixel ratio gradient features | |
| Hu et al. | Fast face detection based on skin color segmentation using single chrominance Cr | |
| Das et al. | Human face detection in color images using HSV color histogram and WLD | |
| Youlian et al. | Face detection method using template feature and skin color feature in rgb color space | |
| CN109635679B (en) | Real-time target paper positioning and loop line identification method | |
| Liang et al. | Applying Image Processing Technology to Face Recognition. | |
| CN104077566B (en) | Bayonet socket picture method for detecting human face based on color difference | |
| Le et al. | Pedestrian lane detection in unstructured environments for assistive navigation | |
| Wu et al. | Real-time 2D hands detection and tracking for sign language recognition | |
| CN107977604B (en) | Hand detection method based on improved aggregation channel characteristics | |
| Wu et al. | Traffic lights detection and recognition algorithm based on multi-feature fusion | |
| KR102139932B1 (en) | A Method of Detecting Character Data through a Adaboost Learning Method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160323 |