This is the authors' abstract. We don't add key points for this paper.
Currently, the mainstream research on micro-expression recognition mostly employs image sequences decoded from videos for analysis. Due to the subtle nature of facial micro-expression movements, high-resolution images are required to meet accuracy standards, thereby posing computational demands. Therefore, the paper substitutes motion vector information extracted from video encoding data for RGB images as micro-expression motion features. This substitution reduces storage requirements and alleviates the storage pressure caused by high-resolution image sequences. Based on this, the paper proposes a micro-expression recognition framework to validate our approach. Ultimately, an evaluation is conducted on the CASME II dataset. Experimental results demonstrate that, compared to existing methods, the micro-expression motion features obtained by this approach have reduced parameter count by one-third, achieving an average recognition accuracy of 63%.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Xie et al. (2023) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: