中国科学院心理研究所机构知识库

摘要：微表情对应着人类的真实情绪。通过分析微表情，结合上下文语境，有望揭示人类的真实意图。微表情分析包括微表情检测和微表情识别两个部分，在反恐、刑事审讯、安检等领域有重要的应用价值。微表情的自动分析能够通过计算机自动完成表情的检测和识别，因此具有极大的推广价值。随着深度学习在各行各业绽放异彩，微表情分析借助深度学习的工具也取得了新的进展。本报告回顾深度学习的背景、基本概念，介绍微表情分析在深度学习领域的相关进展，也总结微表情自动分析的未来方向。

Abstract：Micro-expression can reveal the real emotion of human. Through the analysis of micro-expression and context, the real intent of human may be uncovered. The analysis of micro-expression includes micro-expression spotting and recognition, and finds applications in the field of lie detection and homeland security. The automatic micro-expression analysis system can spot and recognize the micro-expression with aid of computer, and may be deployed in wide range of application environment. With the help of the deep learning, the automatic analysis of micro-expression has made progress. This presentation will review the background and concept of deep learning, introduce the progress of analysis of micro-expression by using deep learning, and bring on the future prospect of automatic analysis of micro-expression.