A New Feature Extraction and Recognition Method for Microexpression Based on Local Non-negative Matrix Factorization

A New Feature Extraction and Recognition Method for Microexpression Based on Local Non-negative Matrix Factorization
Title:
A New Feature Extraction and Recognition Method for Microexpression Based on Local Non-negative Matrix Factorization
Other Titles:
Frontiers in Neurorobotics
Publication Date:
16 November 2020
Citation:
Gao J, Chen H, Zhang X, Guo J and Liang W (2020) A New Feature Extraction and Recognition Method for Microexpression Based on Local Non-negative Matrix Factorization. Front. Neurorobot. 14:579338. doi: 10.3389/fnbot.2020.579338
Abstract:
Micro-expression is usually characterized by short duration and small action range, and the existing general expression recognition algorithms do not work well for micro-expression. As a feature extraction method, non-negative matrix factorization can decompose the original data into different components, which has been successfully applied to facial recognition. In this paper, local non-negative matrix factorization is explored to decompose micro-expression into some facial muscle actions, and extract features for recognition based on apex frame. However, the existing micro-expression datasets fall short of samples to train a classifier with good generalizability. The macro-to-micro algorithm based on singular value decomposition can augment the number of micro-expressions, but it cannot meet non-negative properties of feature vectors. To address these problems, we propose an improved macro-to-micro algorithm to augment micro-expression samples by manipulating the macro-expression data based on local non-negative matrix factorization. Finally, several experiments are conducted to verify the effectiveness of the proposed scheme which results show that it has a higher recognition accuracy for micro-expression compared with the related algorithms based on CK+/CASME2/SAMM datasets.
License type:
http://creativecommons.org/licenses/by/4.0/
Funding Info:
This research is supported by National Natural Science Foundation (NNSF) of China under Grant 61803103 and China Scholarship Council (CSC) under Grant 201908440537.
Description:
ISSN:
1662-5218
Files uploaded:
File Size Format Action
There are no attached files.