Deep Active Latent Surfaces for Medical Geometries
This work addresses shape reconstruction challenges in medical imaging, offering a solution that could improve accuracy in clinical applications, though it appears incremental as it builds on existing latent representation methods.
The paper tackles the problem of reconstructing 3D shapes from noisy or incomplete medical data by proposing a hybrid deep-learning approach that uses separate latent vectors at each mesh vertex, balancing flexibility and generalization to avoid overfitting. The method is demonstrated on multiple medical image processing tasks, showing effectiveness in shape reconstruction.
Shape priors have long been known to be effective when reconstructing 3D shapes from noisy or incomplete data. When using a deep-learning based shape representation, this often involves learning a latent representation, which can be either in the form of a single global vector or of multiple local ones. The latter allows more flexibility but is prone to overfitting. In this paper, we advocate a hybrid approach representing shapes in terms of 3D meshes with a separate latent vector at each vertex. During training the latent vectors are constrained to have the same value, which avoids overfitting. For inference, the latent vectors are updated independently while imposing spatial regularization constraints. We show that this gives us both flexibility and generalization capabilities, which we demonstrate on several medical image processing tasks.