Anatomical Positional Embeddings
This work addresses the need for efficient anatomical positional embeddings in medical imaging, enabling improved downstream applications like organ localization and image cropping, though it is incremental as it builds on existing models.
The authors tackled the problem of generating 3D anatomical positional embeddings for medical image voxels, resulting in a model that efficiently produces voxel-wise embeddings for whole volumetric images and demonstrates superior performance in tasks like anatomical landmark retrieval and weakly-supervised organ localization, with practical applications such as cropping CT images with 0.99 recall and reducing image volume by 10-100 times.
We propose a self-supervised model producing 3D anatomical positional embeddings (APE) of individual medical image voxels. APE encodes voxels' anatomical closeness, i.e., voxels of the same organ or nearby organs always have closer positional embeddings than the voxels of more distant body parts. In contrast to the existing models of anatomical positional embeddings, our method is able to efficiently produce a map of voxel-wise embeddings for a whole volumetric input image, which makes it an optimal choice for different downstream applications. We train our APE model on 8400 publicly available CT images of abdomen and chest regions. We demonstrate its superior performance compared with the existing models on anatomical landmark retrieval and weakly-supervised few-shot localization of 13 abdominal organs. As a practical application, we show how to cheaply train APE to crop raw CT images to different anatomical regions of interest with 0.99 recall, while reducing the image volume by 10-100 times. The code and the pre-trained APE model are available at https://github.com/mishgon/ape .