Zhaoye Zhou1, Gengyan Zhao2, Richard Kijowski2, Fang Liu2. 1. Department of Biomedical Engineering, University of Minnesota, Minneapolis, Minnesota. 2. Departments of Radiology, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin.
Abstract
PURPOSE: To describe and evaluate a new segmentation method using deep convolutional neural network (CNN), 3D fully connected conditional random field (CRF), and 3D simplex deformable modeling to improve the efficiency and accuracy of knee joint tissue segmentation. METHODS: A segmentation pipeline was built by combining a semantic segmentation CNN, 3D fully connected CRF, and 3D simplex deformable modeling. A convolutional encoder-decoder network was designed as the core of the segmentation method to perform high resolution pixel-wise multi-class tissue classification for 12 different joint structures. The 3D fully connected CRF was applied to regularize contextual relationship among voxels within the same tissue class and between different classes. The 3D simplex deformable modeling refined the output from 3D CRF to preserve the overall shape and maintain a desirable smooth surface for joint structures. The method was evaluated on 3D fast spin-echo (3D-FSE) MR image data sets. Quantitative morphological metrics were used to evaluate the accuracy and robustness of the method in comparison to the ground truth data. RESULTS: The proposed segmentation method provided good performance for segmenting all knee joint structures. There were 4 tissue types with high mean Dice coefficient above 0.9 including the femur, tibia, muscle, and other non-specified tissues. There were 7 tissue types with mean Dice coefficient between 0.8 and 0.9 including the femoral cartilage, tibial cartilage, patella, patellar cartilage, meniscus, quadriceps and patellar tendon, and infrapatellar fat pad. There was 1 tissue type with mean Dice coefficient between 0.7 and 0.8 for joint effusion and Baker's cyst. Most musculoskeletal tissues had a mean value of average symmetric surface distance below 1 mm. CONCLUSION: The combined CNN, 3D fully connected CRF, and 3D deformable modeling approach was well-suited for performing rapid and accurate comprehensive tissue segmentation of the knee joint. The deep learning-based segmentation method has promising potential applications in musculoskeletal imaging.
PURPOSE: To describe and evaluate a new segmentation method using deep convolutional neural network (CNN), 3D fully connected conditional random field (CRF), and 3D simplex deformable modeling to improve the efficiency and accuracy of knee joint tissue segmentation. METHODS: A segmentation pipeline was built by combining a semantic segmentation CNN, 3D fully connected CRF, and 3D simplex deformable modeling. A convolutional encoder-decoder network was designed as the core of the segmentation method to perform high resolution pixel-wise multi-class tissue classification for 12 different joint structures. The 3D fully connected CRF was applied to regularize contextual relationship among voxels within the same tissue class and between different classes. The 3D simplex deformable modeling refined the output from 3D CRF to preserve the overall shape and maintain a desirable smooth surface for joint structures. The method was evaluated on 3D fast spin-echo (3D-FSE) MR image data sets. Quantitative morphological metrics were used to evaluate the accuracy and robustness of the method in comparison to the ground truth data. RESULTS: The proposed segmentation method provided good performance for segmenting all knee joint structures. There were 4 tissue types with high mean Dice coefficient above 0.9 including the femur, tibia, muscle, and other non-specified tissues. There were 7 tissue types with mean Dice coefficient between 0.8 and 0.9 including the femoral cartilage, tibial cartilage, patella, patellar cartilage, meniscus, quadriceps and patellar tendon, and infrapatellar fat pad. There was 1 tissue type with mean Dice coefficient between 0.7 and 0.8 for joint effusion and Baker's cyst. Most musculoskeletal tissues had a mean value of average symmetric surface distance below 1 mm. CONCLUSION: The combined CNN, 3D fully connected CRF, and 3D deformable modeling approach was well-suited for performing rapid and accurate comprehensive tissue segmentation of the knee joint. The deep learning-based segmentation method has promising potential applications in musculoskeletal imaging.
Authors: Eva Steidle-Kloc; Adam G Culvenor; Jan Dörrenberg; Wolfgang Wirth; Anja Ruhdorfer; Felix Eckstein Journal: Arthritis Care Res (Hoboken) Date: 2018-03-11 Impact factor: 4.794
Authors: D T Felson; C E Chaisson; C L Hill; S M Totterman; M E Gale; K M Skinner; L Kazis; D R Gale Journal: Ann Intern Med Date: 2001-04-03 Impact factor: 25.391
Authors: M-P Hellio Le Graverand; R J Buck; B T Wyman; E Vignon; S A Mazzuca; K D Brandt; M Piperno; H C Charles; M Hudelmaier; D J Hunter; C Jackson; V Byers Kraus; T M Link; S Majumdar; P V Prasad; T J Schnitzer; A Vaz; W Wirth; F Eckstein Journal: Ann Rheum Dis Date: 2010-01 Impact factor: 19.103
Authors: S Reichenbach; M Yang; F Eckstein; J Niu; D J Hunter; C E McLennan; A Guermazi; F Roemer; M Hudelmaier; P Aliabadi; D T Felson Journal: Ann Rheum Dis Date: 2010-01 Impact factor: 19.103
Authors: Jean-Pierre Raynauld; Johanne Martel-Pelletier; Marie-Josée Berthiaume; Gilles Beaudoin; Denis Choquette; Boulos Haraoui; Hyman Tannenbaum; Joan M Meyer; John F Beary; Gary A Cline; Jean-Pierre Pelletier Journal: Arthritis Res Ther Date: 2005-12-30 Impact factor: 5.156
Authors: Jeffrey B Driban; Lori Price; Grace H Lo; Jincheng Pang; David J Hunter; Eric Miller; Robert J Ward; Charles B Eaton; John A Lynch; Timothy E McAlindon Journal: Arthritis Res Ther Date: 2013 Impact factor: 5.156
Authors: Wei Zha; Sean B Fain; Mark L Schiebler; Michael D Evans; Scott K Nagle; Fang Liu Journal: J Magn Reson Imaging Date: 2019-04-04 Impact factor: 4.813
Authors: Romil F Shah; Alejandro M Martinez; Valentina Pedoia; Sharmila Majumdar; Thomas P Vail; Stefano A Bini Journal: J Arthroplasty Date: 2019-07-24 Impact factor: 4.757
Authors: Michal Byra; Mei Wu; Xiaodong Zhang; Hyungseok Jang; Ya-Jun Ma; Eric Y Chang; Sameer Shah; Jiang Du Journal: Magn Reson Med Date: 2019-09-19 Impact factor: 4.668
Authors: Tomomi Nobashi; Claudia Zacharias; Jason K Ellis; Valentina Ferri; Mary Ellen Koran; Benjamin L Franc; Andrei Iagaru; Guido A Davidzon Journal: J Digit Imaging Date: 2020-04 Impact factor: 4.056