Published 2024-11-01
“…We effectively extract the changing features as the input of the model, fusing channel attention and spatial attention into Vision Transformer to capture correlations between features in different dimensions, which enhances the accuracy of the
identification of
micro-expressions. In order to verify the effectiveness of the algorithm mentioned, we conduct experimental testing in the SAMM, CAS (ME) II, and SMIC datasets and compared the results with other former best algorithms. …”
Get full text
Article