An application of cascaded 3D fully convolutional networks for medical image segmentation

By Holger R. Roth, H. Oda, Xiangrong Zhou, N. Shimizu, Ying Yang, Y. Hayashi, M. Oda, Michitaka Fujiwara, Kazunari Misawa, Kensaku Mori. Comput. Medical Imaging Graph., 2018

295 citations recorded by the source corpus.

Recent advances in 3D fully convolutional networks (FCN) have made it feasible to produce dense voxel-wise predictions of volumetric images. In this work, we show that a multi-class 3D FCN trained on manually labeled CT scans of several anatomical structures (ranging from the large organs to thin vessels) can achieve competitive segmentation results, while avoiding the need for handcrafting features or training class-specific models. To this end, we propose a two-stage, coarse-to-fine approach that will first use a 3D FCN to roughly define a candidate region, which will then be used as input to a second 3D FCN. This reduces the number of voxels the second FCN has to classify to ∼10% and allows it to focus on more detailed segmentation of the organs and vessels. We utilize training and validation sets consisting of 331 clinical CT images and test our models on a completely unseen data collection acquired at a different hospital that includes 150 CT scans, targeting three anatomical organs (liver, spleen, and pancreas). In challenging organs such as the pancreas, our cascaded approach improves the mean Dice score from 68.5 to 82.2%, achieving the highest reported average score on this dataset. We compare with a 2D FCN method on a separate dataset of 240 CT scans with 18 classes and achieve a significantly higher performance in small organs and vessels. Furthermore, we explore fine-tuning our models to different datasets. Our experiments illustrate the promise and robustness of current 3D FCN based semantic segmentation of medical images, achieving state-of-the-art results.1.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/84b3ccf6f01145da6aedeab16b40f01004de2536

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation (2022, background): “Besides, to resolve the challenges of small target organs, a series of works [30, 21, 4] adopt the cascaded structures, where the networks are designed in a coarse-to-fine manner.”
A review of deep learning in medical imaging: Imaging traits, technology trends, case studies with progress highlights, and future promises (2020, methodology): “Several groups have used data sets such as TCIA CT pancreas to improve pancreas segmentation with Dice coefﬁcients reaching the mid 80 percentile [181], [182], [183], [184].”
AbdomenCT-1K: Is Abdominal Organ Segmentation a Solved Problem? (2020, background): “strategies were proposed [52], [57], where the first stage networks played different roles.”