EEG Based Emotion recognition machine learning
Budget: $30 – $250 CAD
Topic: Electroencephalogram (EEG)-based emotion recognition
Source:
• Data: Use SEED-IV EEG dataset containing samples for 4 different emotions. You can also use the
following URL to access the dataset - https://bcmi.sjtu.edu.cn/home/seed/index.html
• Public sample code and publications on the same topic can be found on the following website:
leader board with papers and code. You can also use the following URL to access them -
https://paperswithcode.com/dataset/seed-1
Deliverables:
1- A better machine learning or deep learning model(s) compared to the existing solutions in the
publications. If your machine learning model has better performance than the reported results in
the research articles, the project will be awarded a maximum of 5 bonus points with respect to
the percentage of improvement.
2- Each group will share a folder on Google drive to show their implementation (code and executed
results). The folder must contain a very detailed documented code. A code without
documentation WILL NOT be accepted. The code and intermediate reports (see next page)
must also be uploaded to the D2L.
Please refer to “coding-and-report-writing-tips.pdf”, for proper coding and documentation.
3- A well-written formal project report with appropriate flowcharts, block diagrams, data
visualizations, experimental results, tables, figures, graphs, etc. (the report is to be written in IEEE
- Manuscript Templates for Conference Proceedings)
Infrastructure:
• A desktop or a laptop with quad-core, W10, and 8GB memory, will be sufficient for this project.
• It is recommended to use the Google CoLab cloud environment.
Prerequisite Programming Skills:
python
--- Need the project within 3 days and will pay after I review the project.
Source:
• Data: Use SEED-IV EEG dataset containing samples for 4 different emotions. You can also use the
following URL to access the dataset - https://bcmi.sjtu.edu.cn/home/seed/index.html
• Public sample code and publications on the same topic can be found on the following website:
leader board with papers and code. You can also use the following URL to access them -
https://paperswithcode.com/dataset/seed-1
Deliverables:
1- A better machine learning or deep learning model(s) compared to the existing solutions in the
publications. If your machine learning model has better performance than the reported results in
the research articles, the project will be awarded a maximum of 5 bonus points with respect to
the percentage of improvement.
2- Each group will share a folder on Google drive to show their implementation (code and executed
results). The folder must contain a very detailed documented code. A code without
documentation WILL NOT be accepted. The code and intermediate reports (see next page)
must also be uploaded to the D2L.
Please refer to “coding-and-report-writing-tips.pdf”, for proper coding and documentation.
3- A well-written formal project report with appropriate flowcharts, block diagrams, data
visualizations, experimental results, tables, figures, graphs, etc. (the report is to be written in IEEE
- Manuscript Templates for Conference Proceedings)
Infrastructure:
• A desktop or a laptop with quad-core, W10, and 8GB memory, will be sufficient for this project.
• It is recommended to use the Google CoLab cloud environment.
Prerequisite Programming Skills:
python
--- Need the project within 3 days and will pay after I review the project.