Details
Real-world Academic Emotion Recognition Dataset (RAER) is a dataset collected in naturalistic learning environments, comprising approximately 2,700 video clips from about 140 students. Each clip was independently annotated by around 10 annotators using two sets of academic emotion labels with different levels of granularity, in order to enhance annotation consistency and reliability. RAER has the following key features:
- Diverse natural learning environments, including classrooms, libraries, laboratories, and dormitories;
- Varied learning contexts, covering both classroom sessions and individual study;
- Rich contextual semantic information, such as common academic behaviors like using a phone, reading, writing, using a computer, or fidgeting with fingers;
- Each video is labeled with one of five academic emotion categories (Neutrality, Enjoyment, Confusion, Fatigue, and Distraction).
To enable objective evaluation of subsequent research, the dataset is divided into a training set and a test set, with the training set being approximately four times larger than the test set. The distribution of emotion categories is kept nearly identical between the two subsets.
Sample Images
Terms and Conditions of Use
By downloading or using the RAER dataset, you agree to comply with the following terms and conditions:
1. Non-Commercial Use Only
The RAER dataset is provided strictly for non-commercial research purposes. Any commercial use, including but not limited to commercial product development, service deployment, or monetization, is strictly prohibited.
2. Ownership and Responsibility
All videos and images in the RAER dataset were collected from consenting volunteer participants. These materials do not belong to the Zhejiang Gongshang University. The University assumes no legal responsibility for the content, interpretation, or usage of the images or videos in the dataset.
3. Restrictions on Use
You agree not to:
- Reproduce, duplicate, or copy the dataset for purposes other than internal research.
- Sell, trade, resell, or exploit any part of the RAER dataset or its derivatives for commercial purposes.
- Publish, distribute, or share the dataset or any substantial portion of it with third parties outside your organization.
4. Internal Use Within an Organization
Copies of the dataset may be made solely for internal use within a single site of the same organization. Redistribution outside of the organization or across multiple sites is not allowed.
5. Publication Requirements
If you use the RAER dataset in your research and publish any results, you must:
- Acknowledge the use of the RAER dataset in your publication.
- Cite the associated RAER dataset paper as specified on the dataset's official release page.
6. Termination of Access
The authors reserve the right to terminate your access to the RAER dataset at any time, for any reason, including but not limited to suspected misuse or violation of these terms.
7. Modification and Updates
These terms and conditions are subject to change without prior notice. It is your responsibility to stay updated with the latest terms by referring to the official dataset release page.
How to get the Dataset
This dataset is publicly available. It is free for professors and researcher scientists affiliated to a University.
Permission to use but not reproduce or distribute the RAER dataset is granted to all researchers given that the following steps are properly followed:
Send an e-mail to Luming Zhao before downloading the dataset. You will need a password to access the files of this dataset. Your Email MUST be set from a valid University account and MUST include the following text:
Content Preview
To be added
Citation
Contact
Please contact Luming Zhao and Wenwu Yang for questions about the dataset.