← Back to Catalog

🇧🇷UFPR-SR-Plates

Super-Resolution + Recognition · Brazil · 2025

Staff Only
Images
100K (10K tracks x 10 frames)
Frame Type
Multi
Chars
36
Classes (+CTC blank)
37
Annotation
5 LR + 5 HR sequential frames + plate text

Description

100K LP images in 10K tracks from real-world surveillance. Each track: 5 LR frames (vehicle far) + 5 HR frames (vehicle close). Sequential capture for multi-frame super-resolution.

Resolution: LR (far from camera) + HR (close to camera), real surveillance

License: Academic Staff Only

Access

Citation

BibTeX
@article{silva2025ufpr_sr_plates,
  title={UFPR-SR-Plates: A Benchmark for Real-World License Plate Super-Resolution},
  author={Silva, Valfrido and Others},
  journal={Neurocomputing},
  year={2025}
}

Related Datasets