A Third Generation Facial Spontaneous Micro-Expression Database with Depth Information and High Ecological Validity

Copyright: Institute of Psychology, Chinese Academy of Siences

Part A: 100 subjects, 13 videos per subject, with labeled micro-expressions and macro-expression

Part B: 116 subjects, 13 videos per subject, unlabeled

Part C: 31 subjects, one 8 min video per subject, with labeled micro-expressions and macro-expression

J. Li et al., "CAS(ME)$^{3}$: A Third Generation Facial Spontaneous Micro-Expression Database with Depth Information and High Ecological Validity," in IEEE Transactions on Pattern Analysis and Machine Intelligence, doi: 10.1109/TPAMI.2022.3174895.

http://casme.psych.ac.cn/casme/e4

Micro-expression (ME) is a significant non-verbal communication clue that reveals one person's genuine emotional state. The development of micro-expression analysis (MEA) has just gained attention in the last decade. However, the small sample size problem constrains the use of deep learning on MEA. Besides, ME samples distribute in six different databases, leading to database bias. Moreover, the ME database development is complicated. Addressing this issue, we introduce a large-scale spontaneous ME database: CAS(ME)$^{3}$. The contribution is summarized as follows:

• CAS(ME)$^{3}$ offers around 80 hours of videos with over 8,000,000 frames, including over 1,000 manually labeled MEs and over 3,400 manually labeled macro-expressions. Such a large sample size allows effective MEA method validation while avoiding database bias.

• Inspired by psychological experiments, CAS(ME)$^{3}$ provides the depth information as an additional modality unprecedentedly, contributing to multi-modal MEA.

• For the first time, CAS(ME)$^{3}$ elicits ME with high ecological validity using the mock crime paradigm, along with physiological and voice signals, contributing to practical MEA.

• Besides, CAS(ME)$^{3}$ provides 1,508 unlabeled videos with more than 4,000,000 frames, i.e., a data platform for unsupervised MEA methods.

• Finally, we demonstrate the effectiveness of depth information by the proposed depth flow algorithm and RGB-D information.

This folder mainly contains some exception information for the samples in the database. Users are welcome to contribute here, i.e., upload issues encountered in using the database.

For example, casme3_partA_error_list_ZhouJu.xlsx is the anomaly information of some micro-expression samples in Part A provided by Zhou Ju.

This folder includes the supplementary files such as the self report of the subjects per emotional stimulus video.

We will further improve the annotation of the database, add relevant materials.

If you have any concerns or suggestions, such as what information you would like us to provide, please do not hesitate to let us know.

Questions and comments can be sent to:

Jingting Li(lijt@psych.ac.cn) or Su-Jing Wang(wangsujing@psych.ac.cn)