Real-Time Arabic Sign Language Sentences for YOLOv8

YOLOv8 dataset for real-time Arabic Sign Language (ArSL) sentences and words. The dataset is designed for real-time translation applications. The author and specific creation details are not provided.

Use Cases

• Train a YOLOv8 model to detect and classify Arabic sign language words from image frames.
• Develop a real-time sentence translation pipeline by chaining predictions of individual ArSL signs.
• Benchmark object detection performance on continuous, dynamic sign language gestures against static image datasets.
• Fine-tune a pre-trained YOLOv8 model for the specific vocabulary and signing style present in the ArSL data.

Strengths

• Dataset is formatted specifically for the YOLOv8 object detection framework, ensuring compatibility.
• Focuses on real-time sentences, implying data captures continuous signing rather than isolated signs.

Limitations

• The total number of images, annotations, and vocabulary size is unknown.
• The geographic origin and demographic diversity of the signers are not specified, which may limit model generalizability.
• The dataset's annotation quality, label accuracy, and potential class imbalance are not documented.

Provenance

Source

Kaggle

Collection Method

Presumably collected and annotated for training YOLOv8 models, but the specific gathering process is unknown.

Geography

Intended for Arabic Sign Language, but the specific countries or regions of origin are not specified.