CASME II: An Improved Spontaneous Micro-Expression Database and the Baseline Evaluation

Wen‐Jing Yan, Xiaobai Li, Sujing Wang, Guoying Zhao, Yong‐Jin Liu, Yu‐Hsin Chen, Xiaolan Fu

PLoS ONE · 2014 · 963 citations · 24 references

DOIFull text

Open access

Concepts

TL;DR

A robust automatic micro‑expression recognition system would have broad applications in national safety, police interrogation, and clinical diagnosis, yet high‑quality databases are currently unavailable. We constructed the CASME II database, recording nearly 3,000 facial movements at 200 fps and 280×340 resolution, selecting 247 micro‑expressions with labeled action units and emotions, and evaluated baseline performance using LBP‑TOP features with an SVM classifier under leave‑one‑subject‑out cross‑validation. The best 5‑class classification accuracy achieved was 63.41 %.

Abstract

A robust automatic micro-expression recognition system would have broad applications in national safety, police interrogation, and clinical diagnosis. Developing such a system requires high quality databases with sufficient training samples which are currently not available. We reviewed the previously developed micro-expression databases and built an improved one (CASME II), with higher temporal resolution (200 fps) and spatial resolution (about 280×340 pixels on facial area). We elicited participants' facial expressions in a well-controlled laboratory environment and proper illumination (such as removing light flickering). Among nearly 3000 facial movements, 247 micro-expressions were selected for the database with action units (AUs) and emotions labeled. For baseline evaluation, LBP-TOP and SVM were employed respectively for feature extraction and classifier with the leave-one-subject-out cross-validation method. The best performance is 63.41% for 5-class classification.

References

24