Synthetic RAW augmentations create continuous low-light samples matching sensor noise, enabling fine-grained evaluation of person detection performance where real data is sparse.
Practical signal-dependent noise parameter estimation from a single noisy image.IEEE Transactions on Image Process- ing, 23(10):4361–4371, 2014
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Making the Discrete Continuous: Synthetic RAW Augmentations for Fine-Grained Evaluation of Person Detection Performance in Low Light
Synthetic RAW augmentations create continuous low-light samples matching sensor noise, enabling fine-grained evaluation of person detection performance where real data is sparse.