Mapillary street-level images labeled for quality assessment. The dataset likely contains images categorized as 'good' or 'bad' for training models. The author and organization are unknown.
Use Cases
- Train image quality classifiers based on 'good' vs. 'bad' labels.
- Benchmark computer vision models for street-level scene understanding.
- Filter low-quality images from geospatial datasets.
- Develop preprocessing pipelines for autonomous vehicle or mapping systems.
Strengths
- Focuses on a specific task of image quality assessment.
- Likely contains labeled data suitable for supervised learning.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.