Unlock instant, AI-driven research and patent intelligence for your innovation.

Face recognition algorithm based on fused HOG (Histogram of Oriented Gradient) features and deep belief network

A deep belief network, face recognition technology, applied in character and pattern recognition, computing, computer parts and other directions, can solve the problems of lack of facial expression and occlusion robustness, to improve recognition accuracy, improve accuracy and The effect of automation and improved recognition rate

Inactive Publication Date: 2018-03-20
INNER MONGOLIA UNIV OF SCI & TECH
View PDF9 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, it lacks robustness to light, facial expressions and occlusions

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Face recognition algorithm based on fused HOG (Histogram of Oriented Gradient) features and deep belief network
  • Face recognition algorithm based on fused HOG (Histogram of Oriented Gradient) features and deep belief network
  • Face recognition algorithm based on fused HOG (Histogram of Oriented Gradient) features and deep belief network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0045] See attached figure 2 , the present embodiment selects the images in three face databases of ORL, YALE, and CAS-PEAL as recognition objects, carries out face recognition according to the algorithm of the present invention, and compares the recognition results with the recognition effects of several other traditional face recognition methods A comparison is made, as shown in Table 1 to Table 3. In the objective evaluation index comparison table shown in Table 1 to Table 3, in the comparison of several indicators such as recognition rate, feature dimension, and feature extraction time, although the algorithm of the present invention is slightly inferior to that in feature dimension and feature extraction. Other traditional algorithms, but the recognition rate far exceeds other traditional algorithms, thus verifying the effectiveness and feasibility of the algorithm described in the present invention.

[0046] Table 1 Recognition rate and parameters of different algorith...

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

No PUM Login to View More

Abstract

The invention discloses a face recognition algorithm based on fused HOG (Histogram of Oriented Gradient) features and a deep belief network. The face recognition algorithm combines fused HOG featuresand the deep belief network, selects the fused HOG features to serve as input of the DBN (Deep Belief Network) so as to help the DBN to know distribution of image features and improve the representation capacity of the DBN; and human intervention is reduced by using features extracted by training of the DBN, and automatic face recognition is realized. The algorithm comprises the steps of dividinga source image into cells, calculating fused features of the image, training the DBN, learning high-level features and abstract features of the image, and realizing image classification and recognition. The face recognition algorithm extracts global and local fused HOG features of the image for recognition by using the characteristics that HOG features are insensitive to direction and light, global features can extract overall features of a face contour and local features can well adapt to local variations of the face. Meanwhile, the face recognition algorithm effectively improves the face recognition accuracy by using the deep learning ability of the DBN.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a face recognition algorithm based on fusion of HOG features and a deep belief network. Background technique [0002] The distribution of facial features is very complex and non-linear. The expression of the face, the person's posture, age, location, and lighting conditions and coverage all affect the face recognition effect to varying degrees. Effective face feature extraction and description is the key to improve the accuracy of face recognition. At present, the main feature extraction methods divide face features into two categories: global features and local features. Global features can represent complete structural information, such as facial contours, skin color, and the overall nature of facial features. To extract these features, global feature-based methods construct a linear subspace of the training set, to which other images can be re-expressed...

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
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/48G06K9/46G06K9/62
CPCG06V40/169G06V40/171G06V10/50G06V10/473G06V10/46G06F18/253G06F18/214
Inventor 张宝华李腾郝逸夫赵艳峰侯海鹏高子翔郭佩瑜
Owner INNER MONGOLIA UNIV OF SCI & TECH