Autoencoders for Multi-Label Prostate MR Segmentation
arXiv:1806.08216v26 citations
Originality Synthesis-oriented
AI Analysis
This addresses segmentation accuracy for prostate MR imaging, but appears incremental as it applies an existing autoencoder method to a specific medical domain.
The paper tackled multi-label prostate MR segmentation by training an autoencoder to learn a low-dimensional representation of the segmentation, resulting in some positive improvements.
Organ image segmentation can be improved by implementing prior knowledge about the anatomy. One way of doing this is by training an autoencoder to learn a lowdimensional representation of the segmentation. In this paper, this is applied in multi-label prostate MR segmentation, with some positive results.