A new metric and texture-identification method show that ImageNet classifiers rely heavily on specific textures, but the claim that texture bias explains natural adversarial examples is largely a consequence of how textures are identified.
Explorations in Texture Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
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.
citation-role summary
other 1
citation-polarity summary
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1roles
other 1polarities
unclear 1representative citing papers
citing papers explorer
-
Err on the Side of Texture: Texture Bias on Real Data
A new metric and texture-identification method show that ImageNet classifiers rely heavily on specific textures, but the claim that texture bias explains natural adversarial examples is largely a consequence of how textures are identified.