AutomotiveUI-Bench-4K: 4,208 Annotations for In-Vehicle Infotainment Interaction

AutomotiveUI-Bench-4K is a validation benchmark dataset containing 998 images and 4,208 annotations focused on user interaction with in-vehicle infotainment systems. The dataset was created by sparks-solutions and covers 15 automotive brands for model years 2018-2025. It was last updated on Hugging Face on May 14, -2025.

Use Cases

• Benchmark object detection models based on bounding box annotations for in-vehicle interfaces.
• Validate UI interaction recognition systems based on annotated photographs of IVI displays.
• Train models to identify automotive brand UI elements based on coverage of 15 OEMs.
• Analyze temporal trends in automotive UI design based on data from model years 2018-2025.

Strengths

• Contains 4,208 annotations across 998 images, providing a substantial validation set.
• Covers a specific temporal range, focusing on model years from 2018 to 2025.
• Includes data from 15 different automotive brands or OEMs, suggesting brand diversity.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Image source is primarily photographs, which may introduce variations in lighting and perspective compared to direct screenshots.
• Row count is unknown, which may limit suitability assessment for large-scale training.

Provenance

Source

sparks-solutions via Hugging Face.

Collection Method

Primarily photographs of IVI displays, with some direct screenshots (e.g., Android Auto).

Time Range

Model years 2018-2025.

Freshness

Last updated 2025-05-14 09:08:48; freshness should be verified.

Geography

Region is listed as 'us', but specific geographic coverage is not detailed.