A Movie Role Based Video Dataset for Autism Spectrum Disorder Micro-Expression Detection
Autism Spectrum Disorder (ASD) is a developmental neurodevelopmental disorder in which patients often suffer from social difficulties, language and communication disorders. Cinemetrics refers to the quantitative analysis and measurement of movies using a variety of metrics. Micro-expressions are very brief, imperceptible changes in facial expression, and the micro-expressions of people with autism spectrum disorders are often a true reflection of their internal state. Therefore, the use of the Cinemetrics method for extracting micro-expressions from ASD patients in movies and analyzing them can help doctors in the early diagnosis of ASD patients. In this paper, a dataset of micro-expressions of ASD patients in movies is established, and 2,944 micro-expression images from 972 shots in ten films are collected and labelled with micro-expressions. Five target detection algorithms are used to detect them. Finally, our database is proved to be effective by the detection results.
Authors
- Ning Xu (ORCID: https://orcid.org/0000-0001-8336-5926)
- Haolong Li
- Yuxuan Li (ORCID: https://orcid.org/0000-0002-7749-6476)
- Chen Li (ORCID: https://orcid.org/0000-0002-0260-2182)
- Chenxi Liu (ORCID: https://orcid.org/0000-0003-3613-1662)
- Jiarong Liu (ORCID: https://orcid.org/0000-0002-4603-4340)
- Yutong Gu
- Hongwei Lei (ORCID: https://orcid.org/0000-0002-4985-9858)
Institutions
- Liaoning Shihua University (CN)
- Northeastern University (CN)
Publication Details
- Published
- 2024-09-13
- DOI
- https://doi.org/10.1109/mvipit65697.2024.00011
- Primary Topic
- Autism Spectrum Disorder Research
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China