An unsupervised cross-database micro-expression recognition method
A recognition method and micro-expression technology, applied in the field of unsupervised cross-database micro-expression recognition, can solve the problems of large differences in the characteristics of the source domain and the target domain, limited sample recognition effect in the target domain, etc., to improve the accuracy and reduce manual annotation. effect of work
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0108] An unsupervised cross-database micro-expression recognition method, such as figure 1 shown, including:
[0109] 1) Prepare training samples, the training samples include macro-expression samples from a macro-expression database and micro-expression samples from a micro-expression database; obtain macro-expression samples, and extract features from the macro-expression to form a macro-expression data matrix where d 1 Represents the dimension of macro-expression feature, N 1 Represents the number of macro-expression samples; obtains micro-expression samples, and extracts features from micro-expressions to form a micro-expression data matrix where d 2 Represents the micro-expression feature dimension, N 2 Represents the number of micro-expression samples;
[0110] 2) After the data alignment of the macro-expression data matrix and the micro-expression data matrix through feature selection and data standardization, the macro-expression data matrix and the micro-expre...
Embodiment 2
[0114] A kind of unsupervised cross-database micro-expression recognition method according to embodiment 1, its difference is:
[0115] In step 1), the feature extracted for the macro-expression is the LBP feature. When extracting the LBP feature of the macro-expression, the facial part block method of the macro-expression adopts the block method of the multi-scale LBP-TOP feature to extract the feature of the micro-expression as a multi-scale feature. LBP-TOP features and MDMO features. The block method of multi-scale LBP-TOP features divides the face area into four grid types of 1×1, 2×2, 4×4 and 8×8, a total of 85 local sub-regions, for each face sub-region The regions are subjected to LBP feature extraction and cascaded.
[0116] In step 2), feature selection is performed on the macro-expression data matrix and the micro-expression data matrix respectively by principal component analysis (PCA), and the macro-expression data matrix and the micro-expression data matrix are ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More