Visualizing Color-wise Saliency of Black-Box Image Classification Models
This work addresses interpretability for deploying models in safety-critical systems, such as traffic-sign recognition, but is incremental as it builds on existing RISE techniques.
The authors tackled the interpretability problem in black-box image classification models by enhancing RISE to incorporate color information, resulting in MC-RISE, which demonstrated effectiveness on datasets like GTSRB and ImageNet.
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, which 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 (e.g., 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.