ALADIN: All Layer Adaptive Instance Normalization for Fine-grained Style Similarity
This work addresses the challenge of defining and labeling artistic style for visual search, benefiting applications in digital art and content retrieval, though it is incremental as it builds on existing normalization techniques.
The paper tackles the problem of fine-grained style similarity for digital artworks by introducing ALADIN, a novel architecture that sets a new state-of-the-art accuracy for style-based visual search, achieving results on both coarse and fine-grained datasets, including a contributed dataset of 2.62 million images.
We present ALADIN (All Layer AdaIN); a novel architecture for searching images based on the similarity of their artistic style. Representation learning is critical to visual search, where distance in the learned search embedding reflects image similarity. Learning an embedding that discriminates fine-grained variations in style is hard, due to the difficulty of defining and labelling style. ALADIN takes a weakly supervised approach to learning a representation for fine-grained style similarity of digital artworks, leveraging BAM-FG, a novel large-scale dataset of user generated content groupings gathered from the web. ALADIN sets a new state of the art accuracy for style-based visual search over both coarse labelled style data (BAM) and BAM-FG; a new 2.62 million image dataset of 310,000 fine-grained style groupings also contributed by this work.