DragonPaint: Rule based bootstrapping for small data with an application to cartoon coloring
This addresses the problem of data scarcity in deep learning for practitioners in art, design, and animation, though it appears incremental as it builds on existing image-to-image translation methods.
The paper tackles deep learning's need for large labeled datasets by proposing a rule-based strategy for extreme augmentation of small datasets, applying it to automate cel-style cartoon coloring with limited training data and achieving performance comparable to models trained on much larger datasets.
In this paper, we confront the problem of deep learning's big labeled data requirements, offer a rule based strategy for extreme augmentation of small data sets and apply that strategy with the image to image translation model by Isola et al. (2016) to automate cel style cartoon coloring with very limited training data. While our experimental results using geometric rules and transformations demonstrate the performance of our methods on an image translation task with industry applications in art, design and animation, we also propose the use of rules on partial data sets as a generalizable small data strategy, potentially applicable across data types and domains.