Micro-Expression Recognition and Analysis

Problem It Solves

In video, image, or behavioral analysis workflows, textual descriptions and surface-level expressions are often not enough: which facial regions show brief changes, whether an emotion appears only for a moment, and whether spoken expression matches micro-expressions are hard to judge consistently by hand. This skill focuses on structured analysis of facial material, turning subjective “reading expressions” into a reviewable report.

How the Skill Works

• Inputs: supports images, videos, local files, and web URLs; common video formats are mp4, avi, and mov, with a 10MB limit.
• Analysis: the internal script processes the input with types such as comprehensive, basic, micro, and trust, producing basic emotions, complex emotions, emotion fluctuation, expression consistency, emotion authenticity, and key micro-expression cues.
• History reports: when triggered, the history list is read from the cloud API instead of local records or long-term memory, and results are presented as a Markdown table with report links.
• Identity handling: user identity is linked internally by the system; users are not asked for usernames, phone numbers, or other identity values, and internal identity values are not exposed in output.

Boundaries and Notes

Micro-expression analysis is affected by video clarity, lighting, facial occlusion, and expression intensity. It is suitable for human-computer interaction, behavioral observation, or reference in psychological assessment workflows, but it cannot replace professional lie detection, diagnosis, or counseling. For best results, use frontal, clear, unobstructed footage with even lighting; for consequential decisions, combine human review and fuller context.