Interpretable and Interactive Deep Multiple Instance Learning for Dental Caries Classification in Bitewing X-rays
This addresses the problem of improving interpretability and accuracy in dental caries classification for clinicians, though it is incremental as it builds on existing multiple instance learning methods.
The paper tackles dental caries detection in bitewing X-rays by proposing an interpretable deep multiple instance learning architecture that outputs heatmaps from weak image-level labels and can be guided by segmentation labels, achieving competitive performance on a dataset of ~38k bitewings (~316k teeth with significant improvements when using external segmentation guidance.
We propose a simple and efficient image classification architecture based on deep multiple instance learning, and apply it to the challenging task of caries detection in dental radiographs. Technically, our approach contributes in two ways: First, it outputs a heatmap of local patch classification probabilities despite being trained with weak image-level labels. Second, it is amenable to learning from segmentation labels to guide training. In contrast to existing methods, the human user can faithfully interpret predictions and interact with the model to decide which regions to attend to. Experiments are conducted on a large clinical dataset of $\sim$38k bitewings ($\sim$316k teeth), where we achieve competitive performance compared to various baselines. When guided by an external caries segmentation model, a significant improvement in classification and localization performance is observed.