CN112613421B - Dimension reduction feature analysis and comparison method for face picture - Google Patents
Dimension reduction feature analysis and comparison method for face picture Download PDFInfo
- Publication number
- CN112613421B CN112613421B CN202011570403.1A CN202011570403A CN112613421B CN 112613421 B CN112613421 B CN 112613421B CN 202011570403 A CN202011570403 A CN 202011570403A CN 112613421 B CN112613421 B CN 112613421B
- Authority
- CN
- China
- Prior art keywords
- face
- features
- feature
- dimension reduction
- extracting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The method for analyzing and comparing the dimension reduction characteristics of the face picture adopts multi-dimensional face characteristic extraction to ensure that the acquired characteristics are known and reliable, PCA dimension reduction is carried out on each extracted face characteristic template after the multi-dimensional characteristics are obtained, the repeatability of the characteristics of a multi-dimensional part is avoided, after the dimension reduction characteristics are obtained, the characteristic points incompatible with the whole face characteristics are fused or abandoned to achieve integration of various dimension reduction and mean value characteristics, and finally the method is based on the mean value characteristics of the five-angle face templates, so that the problem of low face similarity comparison probability is solved, and the description and coding of the face characteristics are more convenient.
Description
Technical Field
The invention relates to an analysis method of picture characteristics, in particular to a dimension reduction characteristic analysis comparison method of a face picture.
Background
In the prior art, the features of the human face are recognized through machine learning, then the features of the human face to be compared are extracted, and then feature values are reserved for comparison. The human face has more features, so that the problem is that the feature values are more, each feature needs to be processed, the calculation amount is large, and the algorithm cannot be directly operated on the embedded equipment in real time, so that the algorithm can only run on a large-scale computer but cannot be applied to the embedded equipment.
Disclosure of Invention
In order to solve the problems, the invention provides a method for analyzing and comparing the dimension-reducing characteristics of the face image, which can effectively reduce the characteristic value, reduce the operation amount and ensure the accuracy in comparison and identification.
In order to achieve the purpose, the method for analyzing and comparing the dimension reduction characteristics of the face picture comprises the following steps:
a) Collecting multi-dimensional face images including a front face, a left face, a right face, an overlooking face and an upwelling face, storing the images in a digital mode, and then extracting the features of the collected images which are decomposed into the outlines of eyes, a nose, a mouth and cheeks;
b) Extracting the features of the front face, and storing the features of the front face;
c) Extracting features of a left face, and performing interval dimension reduction on the features of a right face to extract features so as to obtain more left face features;
d) Extracting the features of the right face, and performing interval dimension reduction on the features of the left face to extract the features so as to obtain more right face features;
e) Extracting the characteristics of the overlook face, extracting the characteristics of the lower face, and performing interval dimension reduction on the upper face to extract the characteristics so as to obtain more lower face characteristics;
f) Extracting the characteristics of the upward-looking face, extracting the characteristics of the upper face, and performing interval dimension reduction on the lower face to extract the characteristics to obtain more upper face characteristics;
g) After the five human face features from the step b to the step f are obtained, comparing the amplitude change of the feature value, and if the amplitude change of the feature value obtained in a certain step is overlarge, discarding the human face feature at the current angle;
h) Carrying out weighted addition on the feature values obtained by the processing in the step g in a multi-dimensional space to obtain the feature sum of the faces at the angles, and calculating the mean feature points of all the obtained feature points at present;
i) And then comparing each obtained feature point with the calculated average value point, if the distance between the current face feature point and the average value feature point is far, considering that the quality of the current feature point is poor, and discarding the current face feature point, and if the distance between the current face feature point and the average value face feature point is close, considering that the quality of the feature point of the current face is high, and keeping the feature point of the current face.
j) And finally, obtaining the mean characteristic of the face by adopting a method of weighting and averaging, and storing the mean characteristic into a face library.
The method for analyzing and comparing the dimension reduction characteristics of the face picture adopts multi-dimensional face characteristic extraction to ensure that the acquired characteristics are known and reliable, PCA dimension reduction is carried out on each extracted face characteristic template after the multi-dimensional characteristics are obtained, the repeatability of the characteristics of a multi-dimensional part is avoided, after the dimension reduction characteristics are obtained, the characteristic points incompatible with the whole face characteristics are fused or abandoned to achieve integration of various dimension reduction and mean value characteristics, and finally the method is based on the mean value characteristics of the five-angle face templates, so that the problem of low face similarity comparison probability is solved, and the description and coding of the face characteristics are more convenient.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the embodiments, structures, features and effects according to the present invention will be given with reference to the preferred embodiments.
Example 1.
The method for analyzing and comparing the dimension reduction features of the face picture, which is described in the embodiment of the invention, comprises the following steps of:
a) Collecting multi-dimensional face images including a front face, a left face, a right face, an overlooking face and an upwelling face, storing the images in a digital mode, and then extracting the features of the collected images which are decomposed into the outlines of eyes, a nose, a mouth and cheeks;
b) Extracting the features of the front face, and storing the features of the front face;
c) Extracting the features of the left face, and performing interval dimension reduction on the features of the right face to extract the features so as to obtain more left face features;
d) Extracting the features of the right face, and performing interval dimension reduction on the features of the left face to extract the features so as to obtain more features of the right face;
e) Extracting the characteristics of the overlook face, extracting the characteristics of the lower face, and performing interval dimension reduction on the upper face to extract the characteristics so as to obtain more lower face characteristics;
f) Extracting the characteristics of the upward-looking face, extracting the characteristics of the upper face, and performing interval dimension reduction on the lower face to extract the characteristics to obtain more upper face characteristics;
g) After the five human face features from the step b to the step f are obtained, comparing the amplitude change of the feature value, and if the amplitude change of the feature value obtained in a certain step is overlarge, discarding the human face feature at the current angle; corresponding face images are collected again;
h) Performing weighted addition on the characteristic values obtained by the processing in the step g in a multi-dimensional space to obtain the characteristic sum of the faces with the angles, and calculating the mean characteristic points of all the currently obtained characteristic points;
i) And then comparing each obtained feature point with the calculated average value point, if the distance between the current face feature point and the average value feature point is far, considering that the quality of the current feature point is poor, and discarding the current face feature point, and if the distance between the current face feature point and the average value face feature point is close, considering that the quality of the feature point of the current face is high, and keeping the feature point of the current face.
j) And finally, obtaining the mean characteristic of the face by adopting a method of weighting and averaging, and storing the mean characteristic into a face library.
The interval dimension reduction extraction is to divide the corresponding part of the picture into different regions and then extract features intermittently, namely to abandon feature extraction in the part of the regions.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (1)
1. A face image dimension reduction feature analysis comparison method is characterized by comprising the following steps:
a) Collecting multi-dimensional face images including a front face, a left face, a right face, an overlooking face and an upwelling face, storing the images in a digital mode, and then extracting the features of the collected images which are decomposed into the outlines of eyes, a nose, a mouth and cheeks;
b) Extracting the features of the front face, and storing the features of the front face;
c) Extracting the features of the left face, and performing interval dimension reduction on the features of the right face to extract the features so as to obtain more left face features;
d) Extracting the features of the right face, and performing interval dimension reduction on the features of the left face to extract the features so as to obtain more right face features;
e) Extracting features of the overlook face, extracting the face features of the lower part, and performing interval dimension reduction extraction on the upper face to obtain more lower face features;
f) Extracting the characteristics of the upward-looking face, extracting the characteristics of the upper face, and performing interval dimension reduction on the lower face to extract the characteristics to obtain more upper face characteristics;
g) After the five human face features from the step b to the step f are obtained, comparing the amplitude change of the feature value, and if the amplitude change of the feature value obtained in a certain step is overlarge, discarding the human face feature at the current angle;
h) Carrying out weighted addition on the feature values obtained by the processing in the step g in a multi-dimensional space to obtain the feature sum of the faces at the angles, and calculating the mean feature points of all the obtained feature points at present;
i) Then comparing each obtained feature point with the average value point obtained by calculation, if the distance between the current face feature point and the average value feature point is far, considering that the quality of the current feature point is poor, and discarding the current face feature point, and if the distance between the current face feature point and the average value face feature point is close, considering that the quality of the feature point of the current face is high, keeping the feature point of the current face;
j) And finally, obtaining the average value characteristics of the human face by adopting a method of weighting and averaging, and storing the average value characteristics into a human face library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011570403.1A CN112613421B (en) | 2020-12-26 | 2020-12-26 | Dimension reduction feature analysis and comparison method for face picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011570403.1A CN112613421B (en) | 2020-12-26 | 2020-12-26 | Dimension reduction feature analysis and comparison method for face picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112613421A CN112613421A (en) | 2021-04-06 |
| CN112613421B true CN112613421B (en) | 2023-04-14 |
Family
ID=75247966
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011570403.1A Active CN112613421B (en) | 2020-12-26 | 2020-12-26 | Dimension reduction feature analysis and comparison method for face picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112613421B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104637246A (en) * | 2015-02-02 | 2015-05-20 | 合肥工业大学 | Driver multi-behavior early warning system and danger evaluation method |
| CN107480658A (en) * | 2017-09-19 | 2017-12-15 | 苏州大学 | Face identification device and method based on multi-angle video |
| CN108108760A (en) * | 2017-12-19 | 2018-06-01 | 山东大学 | A kind of fast human face recognition |
-
2020
- 2020-12-26 CN CN202011570403.1A patent/CN112613421B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104637246A (en) * | 2015-02-02 | 2015-05-20 | 合肥工业大学 | Driver multi-behavior early warning system and danger evaluation method |
| CN107480658A (en) * | 2017-09-19 | 2017-12-15 | 苏州大学 | Face identification device and method based on multi-angle video |
| CN108108760A (en) * | 2017-12-19 | 2018-06-01 | 山东大学 | A kind of fast human face recognition |
Non-Patent Citations (1)
| Title |
|---|
| 基于多子空间直和特征融合的人脸识别算法;叶继华等;《数据采集与处理》;20160115(第01期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112613421A (en) | 2021-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020108362A1 (en) | Body posture detection method, apparatus and device, and storage medium | |
| CN111241975B (en) | Face recognition detection method and system based on mobile terminal edge calculation | |
| CN110991315A (en) | Method for detecting wearing state of safety helmet in real time based on deep learning | |
| CN105678213B (en) | Dual-mode mask person event automatic detection method based on video feature statistics | |
| CN111723687A (en) | Human body action recognition method and device based on neural network | |
| WO2013075295A1 (en) | Clothing identification method and system for low-resolution video | |
| CN110096945B (en) | Indoor monitoring video key frame real-time extraction method based on machine learning | |
| JP2021517281A (en) | Multi-gesture fine division method for smart home scenes | |
| CN115661187A (en) | Image enhancement method for Chinese medicinal preparation analysis | |
| CN109254654B (en) | Driving fatigue feature extraction method combining PCA and PCANet | |
| CN111160107A (en) | Dynamic region detection method based on feature matching | |
| CN117746076A (en) | Equipment image matching method based on machine vision | |
| CN112613421B (en) | Dimension reduction feature analysis and comparison method for face picture | |
| CN112258532A (en) | Method for positioning and segmenting corpus callosum in ultrasonic image | |
| CN110866470A (en) | Face anti-counterfeiting detection method based on random image characteristics | |
| Işikdoğan et al. | Automatic recognition of Turkish fingerspelling | |
| CN115909398A (en) | Cross-domain pedestrian re-identification method based on feature enhancement | |
| CN109977890B (en) | Action recognition method and recognition system thereof | |
| CN108830166B (en) | Real-time bus passenger flow volume statistical method | |
| CN111898473A (en) | Driver state real-time monitoring method based on deep learning | |
| CN114972433B (en) | Image processing method, apparatus, device, storage medium, and program product | |
| CN111046745A (en) | Multi-view clustering method based on soft block diagonal | |
| CN112733705B (en) | 3D intelligent analysis system based on human face | |
| CN111191550B (en) | Visual perception device and method based on automatic dynamic adjustment of image sharpness | |
| CN113591564B (en) | Scene abnormal state detection method |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Hongkuan Inventor after: Xu Quan Inventor after: Shi Liucheng Inventor after: Hu Quan Inventor after: Lou Yongliang Inventor after: Jin Haoxuan Inventor after: Wang Hongwu Inventor after: Xu Xin Inventor before: Xu Quan Inventor before: Jin Haoxuan Inventor before: Zhang Hongkuan Inventor before: Shi Liucheng Inventor before: Wang Hongwu Inventor before: Hu Quan Inventor before: Xu Xin |