Cartoon Explanations of Image Classifiers
Cartoon Explanations of Image Classifiers
复制标题
DOI：
10.1007/978-3-031-19775-8_26
复制
发表时间：
2021-10
期刊：
影响因子：
--
通讯作者：
S. Kolek;Duc Anh Nguyen;R. Levie;Joan Bruna;Gitta Kutyniok
中科院分区：
文献类型：
--
作者：
S. Kolek;Duc Anh Nguyen;R. Levie;Joan Bruna;Gitta Kutyniok
我们提出了CartoonX（卡通解释），一种新的模型不可知的解释方法，专门针对图像分类器和率失真解释（RDE）框架的基础上。自然图像是大致逐段平滑的信号，也称为卡通图像，并且在小波域中往往是稀疏的。CartoonX是第一个利用这一点的解释方法，它要求其解释在小波域中是稀疏的，从而提取图像的相关分段平滑部分，而不是相关的像素稀疏区域。我们证明，CartoonX可以揭示新的有价值的解释信息，特别是错误分类。此外，我们表明，CartoonX实现了更低的失真与更少的系数比国家的最先进的方法。
We presentCartoonX(Cartoon Explanation), a novel model-agnostic explanation method tailored towards image classifiers and based on the rate-distortion explanation (RDE) framework. Natural images are roughly piece-wise smooth signals—also called cartoon-like images—and tend to be sparse in the wavelet domain. CartoonX is the first explanation method to exploit this by requiring its explanations to be sparse in the wavelet domain, thus extracting therelevant piece-wise smoothpart of an image instead of relevant pixel-sparse regions. We demonstrate that CartoonX can reveal novel valuable explanatory information, particularly for misclassifications. Moreover, we show that CartoonX achieves a lower distortion with fewer coefficients than state-of-the-art methods.