Multimodal_Greek_Sign_Language_and_Lip_Reading (v1)

The Multimodal_Greek_Sign_Language_and_Lip_Reading (v1) is a comprehensive dataset designed for research and development in multimodal machine learning, speech recognition, vision recognition, sign language recognition, sign language translation and accessibility technologies.

Description

The dkourem/Multimodal_Greek_Sign_Language_and_Lip_Reading-v1 is a comprehensive dataset designed for research and development in multimodal machine learning, sign language recognition, speech recognition, and accessibility technologies. This dataset focuses on the Greek language and provides high-quality, synchronized data across three modalities:

Speech: High-resolution audio recordings of spoken Greek.
Lip Movement: Video recordings capturing detailed lip movements for lip-sync and speech alignment tasks.
Subtitles: Accurate text transcriptions and timestamps synchronized with audio and video.
Sign Language: High quality Video Greek Sign Language recordings capturing detailed sign language and facial expressions with synced subtitles and audio speech translation.

Suitable for Tasks Such As:

Automatic Speech Recognition (ASR)
Audio-Visual Speech Recognition (AVSR)
Text-to-Speech (TTS) and Speech-to-Text (STT) applications
Accessibility tools for the deaf and hard of hearing
Multimodal sentiment analysis and natural language processing
Sign Language recogntion and translations tasks (SLR,SLT).

Key Features:

Language: Greek (Hellenic) and Greek Sign Language
Format: Includes .wav for audio, .mp4 for video, and .srt/.json for subtitles and timestamps.
High Quality: Captured in a professional studio environment for maximum clarity and precision.
Accessibility: Designed with a focus on inclusivity and use in accessibility research.

Resposnsible codes

The resposnible codes used for the creation of this dataset is available at github https://github.com/dkourem/Multimodal_Greek_Sign_Language_and_Lip_Reading_v1/

Versioning

This is version 1.0 of the dataset. Future versions may include additional speakers, dialects, or extended annotations.

Citation

If you use this dataset, please cite it as: Dimitris Kouremenos, Klimis Ntalianis. (2024). dkourem/Multimodal_Greek_Sign_Language_and_Lip_Reading (v1). Zenodo. https://doi.org/[Insert DOI Here]

Acknowledgments

Special thanks to the contributors, linguists, and accessibility experts involved in the dataset creation process.

License and Usage Restrictions

The dkourem/Multimodal_Greek_Sign_Language_and_Lip_Reading (v1) is protected under the following terms:

Copyright

Usage Restrictions

Any use, reproduction, modification, or distribution of this dataset and the associated code requires prior written permission from the authors.
The dataset and code are intended for academic and research purposes only. Commercial use is strictly prohibited unless explicitly authorized and approved in writing by the authors.

Requesting Permission

For inquiries and permissions, please contact:

Dr. Dimitris Kouremenos: [dkourem[alpha]@gmail.com]
Prof. Klimis Ntalianis: [kntal[alpha]uniwa.gr]

video video