A Method for Face Recognition Using Novel Density Clustering
A density clustering and face recognition technology, applied in character and pattern recognition, instruments, computing and other directions, can solve problems such as inability to recognize well, and achieve the effect of improving accuracy
Active Publication Date: 2018-11-13
HUAQIAO UNIVERSITY
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Due to the complex distribution of face image data, commonly used clustering methods cannot identify complex irregular shape classifications well.
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 moreExamples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0023] A kind of method that adopts novel density clustering of the present invention to carry out face recognition, comprises following several steps:
[0024] Step 1. Read the face image:
[0025] The face image used is a grayscale image with a size of M×N, each pixel is used as a feature point, and K face images are read in to obtain the image feature matrix A i(M×N) , where i=1,2,...,K; for the convenience of calculation, the image feature matrix A i(M×N) Convert to feature vector f i(1×MN) , where f i (1:N)=A i (1,1:N), f i ((N+1):2N)=A i (2,(1:N)), and so on, for the feature vector f i Do 0-1 normalization for each dimension of where j=1,2,…,MN, then update the feature vector f i ;
[0026] Step 2. Calculate the distance matrix:
[0027] The feature vector set f is used as the input of the data set P to be clustered, the number of face feature vector points is denoted as s, and the ith face feature vector point in the data set P to be clustered is denoted as p...
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 present invention proposes a method for face recognition using novel density clustering, reads the face image, converts the image matrix into a feature vector, uses the feature vector set f as the input of the data set P to be clustered, and calculates the The distance between every two face feature vector points, calculate the mean value of all points, find the point closest to the mean point in P as the density center point, calculate the density of all points, iteratively find the density center point, and the final convergence point As the density convergence point, the density convergence point is itself and the point whose density value is greater than t is used as the local density center point, and all the points in the set LPS are clustered according to the nearest neighbor algorithm, and the category label is marked, and all other points are The category of the non-local density center point p is initialized to -1, which is classified into the same category as its convergence center point, and the remaining points with the label of -1 are marked as outliers, and finally the clustering results are output. The present invention has accuracy High, the advantage of being able to recognize complex data improves the accuracy of face recognition.
Description
technical field [0001] The invention relates to a method for face recognition using novel density clustering. Background technique [0002] In recent years, face recognition has become a hot computer technology research field. As a kind of biometric technology, face recognition technology combines image processing, computer graphics, pattern recognition and other research fields. Clustering method is one of the most important components of face recognition technology. Due to the complex distribution of face image data, commonly used clustering methods cannot identify complex irregular shape classifications well. How to apply accurate, robust and efficient clustering methods to face recognition is an urgent problem to be solved. Contents of the invention [0003] The purpose of the present invention is to propose a method for face recognition using a new type of density clustering. Compared with other centroid-based clustering methods, it has the advantages of high accur...
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): G06K9/00G06K9/62
CPCG06V40/169G06V40/172G06F18/23211
Inventor 陈叶旺汤盛宇
Owner HUAQIAO UNIVERSITY