Blank Canvas Dataset: Art-Free Image Segments with Captions

Art-Free-SAM contains filtered image segment IDs from the original SA-1B dataset. The dataset pairs these segments with captions sourced from SAM-LLaVA-Captions10M, organized in a hierarchical folder structure. The dataset was authored by rhfeiyang and last updated on Hugging Face in December 2024.

Use Cases

• Train or evaluate image segmentation models based on the filtered segment IDs.
• Develop vision-language models using the paired image segments and text captions.
• Fine-tune models for generating descriptive captions for specific image regions.
• Conduct research on removing artistic or synthetic elements from visual training data.

Strengths

• Builds upon the established SA-1B (Segment Anything 1 Billion) dataset.
• Integrates captions from the SAM-LLaVA-Captions10M dataset.
• Provides a specific 'art-free' filter on the original data.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count, column definitions, and file formats are unknown.
• License information is not provided.

Provenance

Source

Filtered from the SA-1B dataset; captions from SAM-LLaVA-Captions10M.

Collection Method

Filtering and pairing of existing dataset components.

Time Range

null

Freshness

Last updated 2024-12-04 06:28:57; freshness should be verified.

Geography

null