RAP-Net: Coarse-to-Fine Multi-Organ Segmentation with Single Random Anatomical Prior

TL;DR

This paper introduces RAP-Net, a coarse-to-fine multi-organ segmentation method that uses a single refined model with anatomical priors, achieving superior accuracy over state-of-the-art approaches in abdominal organ segmentation.

Contribution

The paper presents a novel single-model coarse-to-fine segmentation pipeline that incorporates anatomical priors, reducing complexity and improving accuracy in multi-organ segmentation.

Findings

Single refined model outperforms multiple organ-specific models

Achieved an average Dice score of 84.58% on 13 organs

Statistically significant improvement over previous methods

Abstract

Performing coarse-to-fine abdominal multi-organ segmentation facilitates to extract high-resolution segmentation minimizing the lost of spatial contextual information. However, current coarse-to-refine approaches require a significant number of models to perform single organ refine segmentation corresponding to the extracted organ region of interest (ROI). We propose a coarse-to-fine pipeline, which starts from the extraction of the global prior context of multiple organs from 3D volumes using a low-resolution coarse network, followed by a fine phase that uses a single refined model to segment all abdominal organs instead of multiple organ corresponding models. We combine the anatomical prior with corresponding extracted patches to preserve the anatomical locations and boundary information for performing high-resolution segmentation across all organs in a single model. To train and evaluate our method, a clinical research cohort consisting of 100 patient volumes with 13 organs well-annotated is used. We tested our algorithms with 4-fold cross-validation and computed the Dice score for evaluating the segmentation performance of the 13 organs. Our proposed method using single auto-context outperforms the state-of-the-art on 13 models with an average Dice score 84.58% versus 81.69% (p<0.0001).