PornographyDatabase: Video, Segments, and Frames from 2013

Curated for the 2013 paper 'Pooling in image representation: The visual codeword point of view', this repository facilitates requests for video, segments, and frames. It was created by Sandra Avila to support research into visual codeword points and image representation.

Use Cases

• Training image classification models using the provided frames
• Analyzing temporal features in video segments for content recognition
• Evaluating pooling methods in image representation as described in the source paper

Strengths

• Includes multi-level visual data including videos, segments, and frames
• Linked to peer-reviewed research on visual codeword representation
• Released under a permissive MIT license

Limitations

• Data is not directly hosted in the repository and requires a request process
• Metadata and column definitions are not publicly documented
• Temporal staleness as the data was originally compiled in 2013

Provenance

Source

Pooling in image representation: The visual codeword point of view (2013)

Time Range

2013

Freshness

The repository was updated in 2024, though the underlying data originates from 2013.