Instance Segmentation of Anatomical Structures in Chest Radiographs

Automatic and accurate segmentation of anatomical structures in chest radiographs is fundamental and essential for computer-aided diagnosis system. We introduce Mask R-CNN for instance segmentation of lung fields, heart and clavicles. This method efficiently detects different structures and generates accurate segmentation mask for each instance. To the best of our knowledge, we are the first to implement instance segmentation of these three anatomical structures in chest radiographs. We have done extensive experiments on a common benchmark dataset. Results show that the best of our models achieves the state-of-the-art segmentation performance on image resolution of 512 × 512. The Dice and Ω similarity are 0.976 and 0.953 for lung fields, 0.949 and 0.904 for heart, 0.920 and 0.852 for clavicles. And the average contour distance outperforms human observer on both lungs and heart with image resolution of 256 × 256. In addition, it takes only 0.16 and 0.12 seconds per image for the above two resolutions during inference, which is comparable to or even better than current methods.

[1]  Thomas Theelen,et al.  Robust total retina thickness segmentation in optical coherence tomography images using convolutional neural networks. , 2017, Biomedical optics express.

[2]  Iasonas Kokkinos,et al.  DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Ajay Mittal,et al.  LF-SegNet: A Fully Convolutional Encoder–Decoder Network for Segmenting Lung Fields from Chest Radiographs , 2018, Wirel. Pers. Commun..

[4]  Yaozong Gao,et al.  Hierarchical Lung Field Segmentation With Joint Shape and Appearance Sparse Learning , 2014, IEEE Transactions on Medical Imaging.

[5]  Bram van Ginneken,et al.  Segmentation of anatomical structures in chest radiographs using supervised methods: a comparative study on a public database , 2006, Medical Image Anal..

[6]  Clement J. McDonald,et al.  Automatic Tuberculosis Screening Using Chest Radiographs , 2014, IEEE Transactions on Medical Imaging.

[7]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Clement J. McDonald,et al.  Lung Segmentation in Chest Radiographs Using Anatomical Atlases With Nonrigid Registration , 2014, IEEE Transactions on Medical Imaging.

[9]  Shanq-Jang Ruan,et al.  Low order adaptive region growing for lung segmentation on plain chest radiographs , 2018, Neurocomputing.

[10]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[11]  Werapon Chiracharit,et al.  Automatic lung segmentation in chest radiographs using shadow filter and multilevel thresholding , 2016, 2016 International Computer Science and Engineering Conference (ICSEC).

[12]  Bram van Ginneken,et al.  Clavicle segmentation in chest radiographs , 2012, Medical Image Anal..

[13]  Roberto Cipolla,et al.  SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Chirag Agarwal,et al.  Accurate segmentation of lung fields on chest radiographs using deep convolutional networks , 2017, Medical Imaging.

[15]  Hao Chen,et al.  DCAN: Deep Contour-Aware Networks for Accurate Gland Segmentation , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Yuan Yu,et al.  TensorFlow: A system for large-scale machine learning , 2016, OSDI.

[17]  Kaiming He,et al.  Feature Pyramid Networks for Object Detection , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  K. Doi,et al.  Development of a digital image database for chest radiographs with and without a lung nodule: receiver operating characteristic analysis of radiologists' detection of pulmonary nodules. , 2000, AJR. American journal of roentgenology.

[20]  Tao Xu,et al.  An edge-region force guided active shape approach for automatic lung field detection in chest radiographs , 2012, Comput. Medical Imaging Graph..

[21]  Dinggang Shen,et al.  Segmenting Lung Fields in Serial Chest Radiographs Using Both Population-Based and Patient-Specific Shape Statistics , 2008, IEEE Transactions on Medical Imaging.

[22]  Maria Wimmer,et al.  Fully Convolutional Architectures for Multiclass Segmentation in Chest Radiographs , 2017, IEEE Transactions on Medical Imaging.

[23]  Alexandros Karargyris,et al.  Detecting tuberculosis in radiographs using combined lung masks , 2012, 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[24]  Eric P. Xing,et al.  SCAN: Structure Correcting Adversarial Network for Organ Segmentation in Chest X-Rays , 2017, DLMIA/ML-CDS@MICCAI.

[25]  Sangheum Hwang,et al.  Accurate Lung Segmentation via Network-Wise Training of Convolutional Networks , 2017, DLMIA/ML-CDS@MICCAI.

[26]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[27]  Pietro Perona,et al.  Microsoft COCO: Common Objects in Context , 2014, ECCV.

[28]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[29]  Dieter Fox,et al.  PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes , 2017, Robotics: Science and Systems.