Quick Draw Game Sketches Across 345 Categories

Loaded with 50 million drawings contributed by players of the game Quick, Draw!. It covers 345 distinct categories of sketched objects.

Use Cases

• Train a sketch recognition model on 345 categories of user-drawn images.
• Analyze drawing patterns and common errors within specific object categories.
• Generate new sketches using generative models trained on the 50 million drawing samples.

Strengths

• Contains 50 million individual drawings, providing a large-scale data resource.
• Covers a wide variety of 345 distinct object categories for classification tasks.
• Data is sourced from a global user base via an interactive game, capturing diverse drawing styles.

Limitations

• The dataset consists of simple, time-limited sketches which may not represent detailed artistic renderings.
• Drawings are created in a game context, which may introduce intentional or unintentional label noise.
• Specific metadata per drawing, such as user demographics or precise timestamps, is not detailed in the provided input.

Provenance

Source

Players of the game Quick, Draw!.

Collection Method

Crowdsourced contributions from an online drawing game.

Freshness

Last updated on 2024-09-10.

Geography

Global, with a tag indicating US region involvement.