Applying Augraphy-based document effects to digital training data improves handwritten localization mAP for 3 of 4 models, with gains up to 3.97 percentage points.
W.; Larson, S.; Maini, S.; and Boarman, J
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Document Key Information Localization Through Data Augmentation
Applying Augraphy-based document effects to digital training data improves handwritten localization mAP for 3 of 4 models, with gains up to 3.97 percentage points.