A noise attack on license plate readers (Flock AI)
4 days ago
- #dataset-generation
- #adversarial-perturbation
- #license-plate
- A research tool for generating adversarially perturbed license plate overlays on vehicle images.
- Produces structured datasets with reproducibility, transparency, and ethical guardrails.
- Design Principle: user-first, safe by default, hackable by experts.
- Includes CLI and Python API for generating synthetic car backgrounds and license plate overlays.
- Supports custom configurations, perturbations, and reproducible results with random seeds.
- Generates datasets with metadata and results analysis.
- Follows a structured output format with images and labels.
- Includes development commands for formatting, linting, type-checking, and cleanup.
- Designed for research into adversarial robustness of OCR and ALPR systems.
- Comes with ethical guidelines and responsible use practices.