Extended Cohn-Kanade Dataset
Project Overview:
Objective
We wanted to make the Extended Cohn-Kanade dataset better by adding more types of facial expressions and higher-quality videos. This enhancement would help train machine learning models to recognize human emotions more accurately.
Scope
We added new video clips to the dataset showing different facial expressions from various groups of people.
Sources
- We used the original CK+ dataset, recorded extra videos from over 150 people, and included other publicly available datasets to increase the variety of emotions.
Data Collection Metrics
- New Sequences Recorded: 720
- Subjects Involved: 150
- Total Frames Processed: 25,000
Annotation Process
Stages
We carefully labeled each frame of the videos with specific emotions, improved how we identified facial features, and noted when emotions started and peaked.
Annotation Metrics
- Frames Annotated: 25,000
- Emotional States Identified: 32,000
- Facial Landmarks Annotated: 22,000
Quality Assurance
Stages
Annotation Check: Our team of experts reviews the annotations to make sure they’re accurate.
Data Cleanup: We remove any recordings that aren’t useful or aren’t good quality to keep the dataset reliable.
Data Privacy: We make sure we follow the rules about keeping data safe and private.
QA Metrics
- Facial Expression Recognition Accuracy: 97%
- Consistency in Annotation: 99%
- Data Privacy Compliance: 100%
Conclusion
Improving the CK+ dataset has made it much better for training AI systems. This helps researchers and developers who are working on recognizing facial expressions. Our work has widened the dataset’s uses and set a higher standard for the quality of data used in training emotion recognition models.
Quality Data Creation
Guaranteed TAT
ISO 9001:2015, ISO/IEC 27001:2013 Certified
HIPAA Compliance
GDPR Compliance
Compliance and Security
Let's Discuss your Data collection Requirement With Us
To get a detailed estimation of requirements please reach us.