Gabor Wavelet Subband Correlation Structure Face Recognition Method
A related structure, wavelet subband technology, applied in the field of face recognition, to achieve the effect of high accuracy, fast response speed and good robustness
Inactive Publication Date: 2019-10-22
YIBIN UNIV
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
After decomposing the face picture with Gabor wavelet, several decomposition subbands will be obtained
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
[0043] The present invention will be further described below in conjunction with accompanying drawing. figure 1 Shown is the method flow process of the present invention, and concrete implementation steps are as follows:
[0044] Step (1) Preprocess the face image. During face acquisition, the accuracy of recognition will be affected due to illumination changes and image noise. Preprocessing mainly removes lighting effects and image noise. Let the input face image be I, and carry out the following three aspects of preprocessing in turn:
[0045] (1a) First, use Gamma correction to eliminate part of the illumination. Gamma correction is to replace the index value of the pixel with its own value, expressed as: I=I λ , λ is the correction factor, which is 0.2 here.
[0046] (1b) Second, the image is enhanced with histogram regularization. A histogram of a frontal image without the influence of light and noise is selected as a reference, and the current face picture is standa...
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 using Gabor wavelet subband correlation structure. Gabor wavelet can decompose the face image into several subbands in different scales and directions. Since Gabor wavelet is a redundant transformation, there is a strong correlation between the subbands it decomposes. A Gabor wavelet subband can be regarded as the observation data of a random variable, so the correlation structure of multidimensional random variables can be used to represent the subband correlation of Gabor wavelet. Copula is a tool for creating multidimensional statistical models, and its role is to describe the correlation structure between random variables. The present invention uses Gaussian copula to describe the correlation structure between Gabor wavelet subbands, and this correlation structure has a good ability to distinguish human faces. The operation steps are: first preprocess the face, then perform Gabor wavelet decomposition and Gaussian copula to extract the relevant feature matrix, and finally use the relevant feature matrix for face recognition. This method has strong robustness to illumination changes and image noise, high recognition rate and wide application prospect.
Description
technical field [0001] The invention relates to face recognition technology, in particular to a face recognition technology based on Gabor wavelet decomposition sub-band correlation structural features. Background technique [0002] Human face is an important biological characteristic. Like fingerprints, facial features play an important role in identifying people. The current mainstream face recognition technologies include: local descriptor technology, Gabor feature technology and the recent deep learning technology. The complexity of deep learning is extremely high, and the calculation is time-consuming, which is not suitable for ordinary occasions. LBP (Local Binary Pattern) is a widely used local descriptor. However, the descriptors of LBP and its extended version are easily disturbed by noise in the image, such as face photos with large illumination changes and face photos with strong noise taken at night. The Gabor feature method uses the energy (average value) of...
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 Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/40G06T5/40
CPCG06T5/40G06T2207/20192G06T2207/30201G06T2207/20064G06V40/16G06V10/30
Inventor 李朝荣
Owner YIBIN UNIV