OWT: A Foundational Organ-Wise Tokenization Framework for Medical Imaging
This addresses the need for interpretable and generalizable features in medical imaging, offering a foundational framework for semantically disentangled representation learning with broad scalability.
The paper tackles the problem of entangled semantic components in representation learning for medical imaging by proposing an Organ-Wise Tokenization (OWT) framework, which disentangles images into organ-specific token groups and achieves strong performance on tasks like image reconstruction and segmentation while enabling novel clinical capabilities such as organ-specific tumor identification without additional training.
Recent advances in representation learning often rely on holistic embeddings that entangle multiple semantic components, limiting interpretability and generalization. These issues are especially critical in medical imaging, where downstream tasks depend on anatomically interpretable features. To address these limitations, we propose an Organ-Wise Tokenization (OWT) framework with a Token Group-based Reconstruction (TGR) training paradigm. Unlike conventional approaches, OWT explicitly disentangles an image into separable token groups, each corresponding to a distinct organ or semantic entity. Our design ensures each token group encapsulates organ-specific information, boosting interpretability, generalization, and efficiency while enabling fine-grained control for targeted clinical applications. Experiments on CT and MRI datasets demonstrate OWT's power: it not only achieves strong performance on standard tasks like image reconstruction and segmentation, but also unlocks novel, high-impact clinical capabilities including organ-specific tumor identification, organ-level retrieval and semantic-level generation, without requiring any additional training. These findings underscore the potential of OWT as a foundational framework for semantically disentangled representation learning, offering broad scalability and a new perspective on how representations can be leveraged.