KuralHub is a comprehensive repository that reviews and benchmarks Speech Emotion Recognition (SER) datasets across multiple languages.
It provides detailed metadata, access links, and benchmarks using fine-tuned monolingual models for SER.

📄 Paper: Accepted at Interspeech 2026 (to appear) 🌐 Website: https://aaivu.github.io/KuralHub/

KuralHub/
│── datasets/             # Language-specific datasets
│   ├── english/
│   │   ├── README.md     # Overview of English SER datasets
│   │   ├── ravdess.md    # Dataset-specific details
│   ├── spanish/
│   │   ├── README.md
│   │   ├── dataset1.md
....

This survey covers 70+ languages (with 29 benchmarked), including open-source and restricted datasets.
If a language has no available dataset, it is marked accordingly.

We fine-tune pre-trained SER models on datasets individually and report their performance.

1. Browse Datasets: Navigate to datasets/ for language-specific SER datasets.
2. Download Datasets: Follow access links in each dataset file.
3. Run Benchmarks: Check benchmarks/ for model performance.

💡 Know of a missing dataset? Help us expand KuralHub!
📩 Submit a pull request or open an issue with new datasets.

📖 Contribution Guidelines

If you are using our research findings, please cite the following paper:

Citation details will be finalized once the paper is published.

@inproceedings{kuralhub2026,
  title     = {KuralHub: A Comprehensive Review of Speech Emotion Recognition Datasets},
  author    = {Thavarasa, Luxshan and Thevakumar, Jubeerathan and Sivatheepan, Thanikan and Thayasivam, Uthayasanker},
  booktitle = {Interspeech},
  year      = {2026},
  note      = {To appear}
}

All authors are with the Department of Computer Science & Engineering, University of Moratuwa, Sri Lanka.

We would like to thank Dr. Uthayasanker Thayasivam for his guidance as my supervisor, Braveenan Sritharan for his mentorship, and all the dataset owners for making their datasets available for us through open access or upon request. Your support has been invaluable.