Interpretable and Interactive Deep Multiple Instance Learning for Dental Caries Classification in Bitewing X-rays
Authors:
Benjamin Bergner,
Csaba Rohrer,
Aiham Taleb,
Martha Duchrau,
Guilherme De Leon,
Jonas Almeida Rodrigues,
Falk Schwendicke,
Joachim Krois,
Christoph Lippert
Abstract:
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…
▽ More
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.
△ Less
Submitted 26 September, 2023; v1 submitted 17 December, 2021;
originally announced December 2021.
