Cartoon Explanations of Image Classifiers

We present CartoonX (Cartoon Explanation), a novel model-agnostic explanation method tailored towards image classifiers and based on the rate-distortion (RDE) framework. Natural images are roughly piece-wise smooth signals—also called cartoon-like images—and tend to be sparse in wavelet domain. is first exploit this by requiring its explanations domain, thus extracting relevant part of an instead pixel-sparse regions. demonstrate that can reveal valuable explanatory information, particularly for misclassifications. Moreover, we show achieves lower distortion with fewer coefficients than state-of-the-art methods.