Explorations in Texture Learning
This work addresses interpretability in CNNs for researchers, though it is incremental as it builds on existing texture analysis methods.
The paper tackled the problem of identifying and analyzing the textures learned by object classification models, revealing three categories of texture-object associations: strong and expected, strong and unexpected, and expected but not present.
In this work, we investigate \textit{texture learning}: the identification of textures learned by object classification models, and the extent to which they rely on these textures. We build texture-object associations that uncover new insights about the relationships between texture and object classes in CNNs and find three classes of results: associations that are strong and expected, strong and not expected, and expected but not present. Our analysis demonstrates that investigations in texture learning enable new methods for interpretability and have the potential to uncover unexpected biases.