Assamese speech emotion recognition datasets

KuralHub lists 2 speech emotion recognition datasets for Assamese. Each entry below gives the emotion labels, size, recording setup, licence, where to get it and how to cite it.

Language family: Indo-European (Indo-Aryan)
ISO code: as
Spoken mainly in: India (Northeast India: Assam, Arunachal Pradesh, Meghalaya, Nagaland)
Datasets listed: 2
In benchmark: No

Datasets at a glance

Dataset	Type	Samples	Speakers	Recording	Access	Licence
Assamese Emotional Speech Corpus	Audio	800	20	Acted (controlled settings)	On request	CC BY 4.0
Assamese Emotional Speech Dataset	Audio	1750	25	Acted (controlled settings)	On request

Assamese Emotional Speech Corpus

Type: Audio
Emotion labels: Angry, Happy, Sad, Neutral
Size: samples: approximately 800, speakers: 20
Speakers: 20
Recording: Acted (controlled settings)
Access: On request
Licence: CC BY 4.0
Paper: Emotion Recognition from Assamese Speech using MFCC Features and GMM Classifier
Contact: parismita@dibru.ac.in

The Assamese Emotional Speech Corpus comprises recordings from native Assamese speakers, capturing four distinct emotional states: angry, happy, sad, and neutral. This dataset was utilized to investigate emotion recognition using Mel-frequency cepstral coefficients (MFCC) for feature extraction and Gaussian Mixture Model (GMM) classifiers. The research aimed to enhance the accuracy of emotion recognition systems for Assamese speech.

Note: Access details are not specified in the available sources. Interested researchers may need to contact the authors of the paper for access.

Citation

@article{gogoi2023feature,
  title={Feature Extraction of Assamese Speech Based One Motion Analysis},
  author={Gogoi, P and Sharma, D and Bordoloi, R and Sarma, S and Goswami, A},
  journal={Indian Journal of Science and Technology},
  volume={16},
  pages={6--14},
  year={2023}
}

Source: datasets/Assamese/AESC.md · Suggest an edit

Assamese Emotional Speech Dataset

Type: Audio
Emotion labels: Calm, Neutral, Anger, Sad, Happy, Surprise, Disgust
Size: samples: 1750, speakers: 25
Speakers: 25
Recording: Acted (controlled settings)
Access: On request
Licence: Not stated
Paper: Emotion Recognition in Standard Spoken Assamese Language Using Support Vector Machine and Ensemble Model
Contact: nupur.choudhury@dbuniversity.ac.in

The Assamese Emotional Speech Dataset comprises recordings from native Assamese speakers, capturing seven distinct emotional states: calm, neutral, anger, sad, happy, surprise, and disgust. This dataset was utilized to investigate emotion recognition using Support Vector Machine (SVM) and Ensemble Models, including Random Forest and Rotation Forest classifiers. The research aimed to enhance the accuracy of emotion recognition systems for Assamese speech, achieving up to 94.40% accuracy with the Rotation Forest Classifier when multiple features were combined.

Note: Access details are not specified in the available sources. Interested researchers may need to contact the authors of the paper for access.

Citation

@article{choudhury2021emotion,
  title={Emotion recognition in standard spoken Assamese language using support vector machine and ensemble model},
  author={Choudhury, Nupur and Sharma, Uzzal},
  journal={Indian J. Computer Sci. Eng},
  volume={12},
  pages={148--158},
  year={2021}
}

Source: datasets/Assamese/AESD.md · Suggest an edit

Know another Assamese speech emotion dataset, or a newer version of one listed here? Suggest it or open a pull request.

This page is generated from datasets/Assamese/ in the KuralHub repository. If you use a dataset, cite its authors. If KuralHub helped you find it, please also cite the KuralHub paper.