Improving deep neural network generalization and robustness to background bias via layer-wise relevance propagation optimization
This addresses shortcut learning in image classification, particularly for medical applications like COVID-19 and tuberculosis detection, though it is incremental as it builds on existing LRP methods.
The paper tackled the problem of background bias causing poor generalization in deep neural networks by optimizing Layer-wise Relevance Propagation heatmaps, resulting in superior robustness and significantly better generalization performance on external test databases compared to eight state-of-the-art models.
Features in images' backgrounds can spuriously correlate with the images' classes, representing background bias. They can influence the classifier's decisions, causing shortcut learning (Clever Hans effect). The phenomenon generates deep neural networks (DNNs) that perform well on standard evaluation datasets but generalize poorly to real-world data. Layer-wise Relevance Propagation (LRP) explains DNNs' decisions. Here, we show that the optimization of LRP heatmaps can minimize the background bias influence on deep classifiers, hindering shortcut learning. By not increasing run-time computational cost, the approach is light and fast. Furthermore, it applies to virtually any classification architecture. After injecting synthetic bias in images' backgrounds, we compared our approach (dubbed ISNet) to eight state-of-the-art DNNs, quantitatively demonstrating its superior robustness to background bias. Mixed datasets are common for COVID-19 and tuberculosis classification with chest X-rays, fostering background bias. By focusing on the lungs, the ISNet reduced shortcut learning. Thus, its generalization performance on external (out-of-distribution) test databases significantly surpassed all implemented benchmark models.