Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Automatic micro-expression recognition method based on Gabor features and edge orientation histogram (EOH) features

A recognition method and micro-expression technology, applied in character and pattern recognition, instruments, computer components, etc., can solve the problems of high false alarm rate, low expression strength, incapable of micro-expression and macro-expression video segmentation, etc., to improve Speed ​​and efficiency, the effect of improving recognition speed and accuracy

Active Publication Date: 2013-08-21
INST OF PSYCHOLOGY CHINESE ACADEMY OF SCI
View PDF1 Cites 49 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In addition, the method of inducing micro-expressions in this study is also very problematic: in this study, the micro-expressions were required to be imitated by the subjects, and the expression intensity should be as low as possible
The method proposed by Shreve in 2009 cannot segment videos that contain both micro-expressions and macro-expressions. In 2011, they first realized the segmentation of the two within a unified framework, but the results show that the method is not for micro-expressions. has a very low catch rate (50%) and a high false positive rate (50%)
In addition, the micro-expression data set collected by Shreve et al. also has a big problem. It is collected by asking the subjects to imitate
The most important thing is that the method of Shreve et al. is only a method for segmentation of expression videos, and cannot identify the category of expressions contained in the video.

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
  • Automatic micro-expression recognition method based on Gabor features and edge orientation histogram (EOH) features
  • Automatic micro-expression recognition method based on Gabor features and edge orientation histogram (EOH) features
  • Automatic micro-expression recognition method based on Gabor features and edge orientation histogram (EOH) features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0037] Embodiment 1. Facial Expression Recognition Combining Gabor Features and Gentleboost

[0038] To be able to analyze micro-expressions in videos, it is first necessary to be able to recognize common expressions in still images. Embodiment 1 Firstly, the facial expression recognition performance of the algorithm is evaluated on the facial expression database (also known as CK).

[0039] The database contains 6 basic expression videos of 100 college students (age range: 18-30 years old). Among them, 65% of the models are women, 15% of the models are black, and 3% of the models are Asian or Latino. The video shooting method is: the model performs the prescribed facial expressions according to the requirements, and the camera records the frontal facial expressions of the subjects with an analog S-video signal. Videos are finally stored as 640×480 8-bit grayscale images.

[0040] In Example 1, the neutral expressions in 374 videos of 97 models and one or two images of 6 ba...

Embodiment 2

[0070] Embodiment 2. Combining Gabor features and GentleSVM facial expression recognition

[0071] SVM is a general feed-forward neural network, which establishes a hyper-plane as a decision surface by minimizing structural risk, so that the margin between positive and negative examples is maximized. For expression recognition, the performance of SVM and Gentleboost is similar, and both have the highest performance in this field.

[0072] Gentleboost is used to select Gabor features, and SVM is trained on the new representation formed by feature selection to form the final classifier. In this study, this combination is called GentleSVM.

[0073] The data set used in the second embodiment is the same as that in the first embodiment. In order to test the generalization performance of the algorithm, a 10-fold cross-validation is used to evaluate the performance of the algorithm.

[0074] After the Gentleboost training is completed, the Gabor features used by the weak classifie...

Embodiment 2

[0078] Embodiment 2 compares the expression recognition performance of various GentleSVM algorithm combinations and the original SVM. The results are shown in Table 2 and Table 3.

[0079] Table 2 Expression recognition accuracy and training time of various GentleSVM algorithms

[0080]

[0081] Table 3 Recognition accuracy of various expressions

[0082]

[0083] As shown in Table 2, the accuracy rates of all GentleSVM algorithm combinations exceed the accuracy rates of the original SVM and the improved Gentleboost in Embodiment 1. In addition, all GentleSVM algorithm combinations completed 10 times of training within 20 seconds, which means that after completing Gentleboost training, the system can further improve performance with only a small cost of time. Among all GentleSVM algorithm combinations, the expression recognition accuracy of MI+DWT combination is the highest (92.66%). Combined with the results of Example 1, the above results show that the combination o...

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 provides an automatic expression recognition method. The method includes the following steps: step10, capturing face areas in frame images of a video and preprocessing the face areas, step20, extracting Gabor features and EOH features of images of the corresponding face areas, step30, integrating the corresponding features to acquire final superficial features of the target video and acquiring an expression tag sequence of each frame of video images through a classifier acquired by training, step 40, scanning the expression tag sequences, judging duration time of expressions and outputting expression classes according to the acquired micro- expressions.

Description

technical field [0001] The present invention relates to facial expression recognition and image recognition technologies, and more specifically, to an automatic micro-expression recognition method. Background technique [0002] Today's global political landscape is turbulent, and terrorist activities occur frequently in many places. Scientists and engineers around the world are working hard to find behavioral clues related to violence and extreme behavior, and try to develop technologies or methods that can detect these behaviors. [0003] Micro-expressions are closely related to the processing of human's inner emotional information. They cannot be forged and are not controlled by consciousness. They reflect the true emotions and intentions of human beings. Thus, micro-expressions become effective cues for detection of lies and dangerous intentions. The U.S. Department of Defense, the CIA, the Homeland Security Agency, and the Security Management Inspection Agency have eve...

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): G06K9/62G06K9/00
Inventor 吴奇申寻兵傅小兰
Owner INST OF PSYCHOLOGY CHINESE ACADEMY OF SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com