Face recognition method for wearing mask
A face recognition and recognition method technology, applied in the direction of kernel method, neural learning method, character and pattern recognition, etc., can solve the problems of face recognition of people who are not easy to wear masks
Pending Publication Date: 2020-10-30
SHENYANG POLYTECHNIC UNIV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] Purpose of the invention: The method of this application proposes a face recognition training method for wearing mask
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0036] Below in conjunction with accompanying drawing, the present invention will be further described:
[0037] like figure 1 As shown, a face recognition method wearing a mask, the recognition method is divided into a pre-training part and a technical application part; the pre-training part is carried out according to the following steps:
[0038] s1: Put the face pictures of the same person wearing a mask and not wearing a mask with uniform pixel size into folders named after the person;
[0039] s2: Express each pixel value of the face image in grayscale, correct and normalize the image space, and perform black colorization on the part covered by the mask to prevent unnecessary feature data from being extracted;
[0040] s3: Extract the HOG feature value of the upper half of the face wearing a mask; the HOG feature value includes the features of the eye part, eyebrow part, nose bridge part and face edge part, and save it as a feature code;
[0041] s4: Extract the HOG fe...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention relates to a face recognition method for wearing a mask. The method is divided into an early training part and a technical application part. The early training part is divided into the following steps: S1, collecting front face pictures of the same person wearing or not wearing a mask after unifying pixel sizes; S2, performing graying expression on each pixel value of the face picture, performing correction to normalize the space of the picture, and performing black homochromatism on the mask shielding part; S3, extracting the HOG feature value of the upper half face of the facewearing the mask, wherein the HOG feature value comprises features of an eye part, an eyebrow part, a nose bridge part and a face edge part and is stored as a feature code; S4, extracting an HOG feature value of a face without a mask, and storing as a feature code; S5, processing the two groups of HOG feature values to train a machine learning model; S6, carrying out evaluation, adjustment and repeated training by using untrained pictures. And more effective features capable of distinguishing different human faces on the human face can be extracted through calculation.
Description
technical field [0001] The invention belongs to the technical field of face recognition methods, and in particular relates to a face recognition method for wearing a mask. Background technique [0002] As an attribute of human beings with strong self-stability and individual differences, biometric features have become the most ideal basis for automatic identity verification, and face recognition is one of the applications. When the current face recognition technology is used on a face covered by a mask, its recognition rate is greatly reduced or even not recognized. For example, the face recognition model trained by the Haar cascade classifier has a recognition rate of more than 90% for normal faces, but it is basically unable to recognize the faces covered by masks; most of the existing methods are extracted from the whole face. Features, including face shape, mouth, and chin, so most of the effective features cannot be extracted in the case of masks, resulting in inaccura...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06K9/00G06N3/08G06N20/10
CPCG06N20/10G06N3/08G06V40/161G06V40/168G06V40/172
Inventor 任建刘令文张莺露刘津成史璐媛
Owner SHENYANG POLYTECHNIC UNIV