Takaaki Sugino1, Yutaro Suzuki1, Taichi Kin2, Nobuhito Saito2, Shinya Onogi1, Toshihiro Kawase1, Kensaku Mori3,4,5, Yoshikazu Nakajima6. 1. Department of Biomedical Information, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan. 2. Department of Neurosurgery, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan. 3. Graduate School of Informatics, Nagoya University, Nagoya, Japan. 4. Information Technology Center, Nagoya University, Nagoya, Japan. 5. Research Center for Medical Bigdata, National Institute of Informatics, Tokyo, Japan. 6. Department of Biomedical Information, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan. nakajima.bmi@tmd.ac.jp.
Abstract
PURPOSE: In recent years, fully convolutional networks (FCNs) have been applied to various medical image segmentation tasks. However, it is difficult to generate a large amount of high-quality annotation data to train FCNs for medical image segmentation. Thus, it is desired to achieve high segmentation performances even from incomplete training data. We aim to evaluate performance of FCNs to clean noises and interpolate labels from noisy and sparsely given label images. METHODS: To evaluate the label cleaning and propagation performance of FCNs, we used 2D and 3D FCNs to perform volumetric brain segmentation from magnetic resonance image volumes, based on network training on incomplete training datasets from noisy and sparse annotation. RESULTS: The experimental results using pseudo-incomplete training data showed that both 2D and 3D FCNs could provide improved segmentation results from the incomplete training data, especially by using three orthogonal annotation images for network training. CONCLUSION: This paper presented a validation for label cleaning and propagation based on FCNs. FCNs might have the potential to achieve improved segmentation performances even from sparse annotation data including possible noises by manual annotation, which can be an important clue to more efficient annotation.
PURPOSE: In recent years, fully convolutional networks (FCNs) have been applied to various medical image segmentation tasks. However, it is difficult to generate a large amount of high-quality annotation data to train FCNs for medical image segmentation. Thus, it is desired to achieve high segmentation performances even from incomplete training data. We aim to evaluate performance of FCNs to clean noises and interpolate labels from noisy and sparsely given label images. METHODS: To evaluate the label cleaning and propagation performance of FCNs, we used 2D and 3D FCNs to perform volumetric brain segmentation from magnetic resonance image volumes, based on network training on incomplete training datasets from noisy and sparse annotation. RESULTS: The experimental results using pseudo-incomplete training data showed that both 2D and 3D FCNs could provide improved segmentation results from the incomplete training data, especially by using three orthogonal annotation images for network training. CONCLUSION: This paper presented a validation for label cleaning and propagation based on FCNs. FCNs might have the potential to achieve improved segmentation performances even from sparse annotation data including possible noises by manual annotation, which can be an important clue to more efficient annotation.