Visualizing Color-wise Saliency of Black-Box Image Classification Models
Visualizing Color-wise Saliency of Black-Box Image Classification Models
复制标题
DOI：
10.1007/978-3-030-69535-4_12
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Yuhki Hatakeyama;Hiroki Sakuma;Yoshinori Konishi;Kohei Suenaga
中科院分区：
文献类型：
--
作者：
Yuhki Hatakeyama;Hiroki Sakuma;Yoshinori Konishi;Kohei Suenaga
基于机器学习的图像分类正在被广泛使用。然而，由包括深度学习在内的高级方法给出的分类结果通常很难解释。这个可解释性问题是在安全关键系统中部署训练模型的主要障碍之一。已经提出了几种技术来解决这个问题;其中之一是RISE，它通过热图解释分类结果，称为显着图，解释每个像素的重要性。我们提出了MC-RISE（多色RISE），这是一个增强的RISE考虑到颜色信息的解释。我们的方法不仅显示了给定图像中每个像素的显着性，就像原始的RISE一样，而且还显示了每个像素的颜色分量的重要性;具有颜色信息的显着性图特别适用于颜色信息重要的领域（例如，交通标志识别）。我们实现了MC-RISE，并使用两个数据集（GTSRB和ImageNet）对其进行了评估，以证明我们的方法与现有的图像分类结果解释技术相比的有效性。
Image classification based on machine learning is being commonly used. However, a classification result given by an advanced method, including deep learning, is often hard to interpret. This problem of interpretability is one of the major obstacles in deploying a trained model in safety-critical systems. Several techniques have been proposed to address this problem; one of which is RISE, which explains a classification result by a heatmap, called a saliency map, that explains the significance of each pixel. We propose MC-RISE (Multi-Color RISE), which is an enhancement of RISE to take color information into account in an explanation. Our method not only shows the saliency of each pixel in a given image as the original RISE does, but the significance of color components of each pixel; a saliency map with color information is useful especially in the domain where the color information matters (eg, traffic-sign recognition). We implemented MC-RISE and evaluate them using two datasets (GTSRB and ImageNet) to demonstrate the effectiveness of our methods in comparison with existing techniques for interpreting image classification results.