Exemplar Diffusion: Improving Medical Object Detection with Opportunistic Labels
This work addresses the challenge of enhancing object detection accuracy in medical imaging, particularly for datasets with clear spatial structure, by enabling non-expert annotation and quantifying predictive uncertainty, though it appears incremental as it builds on existing diffusion methods.
The paper tackles the problem of improving object detection in medical images by leveraging existing labels at inference, called exemplars, and demonstrates that the method yields an across-the-board increase in average precision and recall while being robust to exemplar quality.
We present a framework to take advantage of existing labels at inference, called \textit{exemplars}, in order to improve the performance of object detection in medical images. The method, \textit{exemplar diffusion}, leverages existing diffusion methods for object detection to enable a training-free approach to adding information of known bounding boxes at test time. We demonstrate that for medical image datasets with clear spatial structure, the method yields an across-the-board increase in average precision and recall, and a robustness to exemplar quality, enabling non-expert annotation. Moreover, we demonstrate how our method may also be used to quantify predictive uncertainty in diffusion detection methods. Source code and data splits openly available online: https://github.com/waahlstrand/ExemplarDiffusion