Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint
Fast and accurate segmentation of knee bone and cartilage on MRI images is becoming increasingly important in the orthopaedic area, as the segmentation is an essential prerequisite step to a patient-specific diagnosis, optimising implant design and preoperative and intraoperative planning. However,...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-05-01
|
| Series: | Frontiers in Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmed.2022.792900/full |
| _version_ | 1811250970448887808 |
|---|---|
| author | Hao Chen Na Zhao Tao Tan Yan Kang Chuanqi Sun Guoxi Xie Nico Verdonschot André Sprengers |
| author_facet | Hao Chen Na Zhao Tao Tan Yan Kang Chuanqi Sun Guoxi Xie Nico Verdonschot André Sprengers |
| author_sort | Hao Chen |
| collection | DOAJ |
| description | Fast and accurate segmentation of knee bone and cartilage on MRI images is becoming increasingly important in the orthopaedic area, as the segmentation is an essential prerequisite step to a patient-specific diagnosis, optimising implant design and preoperative and intraoperative planning. However, manual segmentation is time-intensive and subjected to inter- and intra-observer variations. Hence, in this study, a three-dimensional (3D) deep neural network using adversarial loss was proposed to automatically segment the knee bone in a resampled image volume in order to enlarge the contextual information and incorporate prior shape constraints. A restoration network was proposed to further improve the bone segmentation accuracy by restoring the bone segmentation back to the original resolution. A conventional U-Net-like network was used to segment the cartilage. The ultimate results were the combination of the bone and cartilage outcomes through post-processing. The quality of the proposed method was thoroughly assessed using various measures for the dataset from the Grand Challenge Segmentation of Knee Images 2010 (SKI10), together with a comparison with a baseline network U-Net. A fine-tuned U-Net-like network can achieve state-of-the-art results without any post-processing operations. This method achieved a total score higher than 76 in terms of the SKI10 validation dataset. This method showed to be robust to extract bone and cartilage masks from the MRI dataset, even for the pathological case. |
| first_indexed | 2024-04-12T16:12:41Z |
| format | Article |
| id | doaj.art-6c732c4d5d104559a3054a164fec0f43 |
| institution | Directory Open Access Journal |
| issn | 2296-858X |
| language | English |
| last_indexed | 2024-04-12T16:12:41Z |
| publishDate | 2022-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Medicine |
| spelling | doaj.art-6c732c4d5d104559a3054a164fec0f432022-12-22T03:25:50ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2022-05-01910.3389/fmed.2022.792900792900Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape ConstraintHao Chen0Na Zhao1Tao Tan2Yan Kang3Chuanqi Sun4Guoxi Xie5Nico Verdonschot6André Sprengers7Department of Biomechanical Engineering, University of Twente, Enschede, NetherlandsSchool of Instrument Science and Engineering, Southeast University, Nanjing, ChinaDepartment of Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, NetherlandsCollege of Health Science and Environmental Engineering, Shenzhen Technology University, Shenzhen, ChinaDepartment of Biomedical Engineering, The Sixth Affiliated Hospital, Guangzhou Medical University, Guangzhou, ChinaDepartment of Biomedical Engineering, The Sixth Affiliated Hospital, Guangzhou Medical University, Guangzhou, ChinaOrthopaedic Research Laboratory, Radboud University Medical Center, Nijmegen, NetherlandsDepartment of Biomedical Engineering and Physics, Amsterdam UMC, University of Amsterdam, Amsterdam, NetherlandsFast and accurate segmentation of knee bone and cartilage on MRI images is becoming increasingly important in the orthopaedic area, as the segmentation is an essential prerequisite step to a patient-specific diagnosis, optimising implant design and preoperative and intraoperative planning. However, manual segmentation is time-intensive and subjected to inter- and intra-observer variations. Hence, in this study, a three-dimensional (3D) deep neural network using adversarial loss was proposed to automatically segment the knee bone in a resampled image volume in order to enlarge the contextual information and incorporate prior shape constraints. A restoration network was proposed to further improve the bone segmentation accuracy by restoring the bone segmentation back to the original resolution. A conventional U-Net-like network was used to segment the cartilage. The ultimate results were the combination of the bone and cartilage outcomes through post-processing. The quality of the proposed method was thoroughly assessed using various measures for the dataset from the Grand Challenge Segmentation of Knee Images 2010 (SKI10), together with a comparison with a baseline network U-Net. A fine-tuned U-Net-like network can achieve state-of-the-art results without any post-processing operations. This method achieved a total score higher than 76 in terms of the SKI10 validation dataset. This method showed to be robust to extract bone and cartilage masks from the MRI dataset, even for the pathological case.https://www.frontiersin.org/articles/10.3389/fmed.2022.792900/fullcartilage segmentationbone segmentationMRIdeep learningCNN |
| spellingShingle | Hao Chen Na Zhao Tao Tan Yan Kang Chuanqi Sun Guoxi Xie Nico Verdonschot André Sprengers Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint Frontiers in Medicine cartilage segmentation bone segmentation MRI deep learning CNN |
| title | Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint |
| title_full | Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint |
| title_fullStr | Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint |
| title_full_unstemmed | Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint |
| title_short | Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint |
| title_sort | knee bone and cartilage segmentation based on a 3d deep neural network using adversarial loss for prior shape constraint |
| topic | cartilage segmentation bone segmentation MRI deep learning CNN |
| url | https://www.frontiersin.org/articles/10.3389/fmed.2022.792900/full |
| work_keys_str_mv | AT haochen kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT nazhao kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT taotan kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT yankang kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT chuanqisun kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT guoxixie kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT nicoverdonschot kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint AT andresprengers kneeboneandcartilagesegmentationbasedona3ddeepneuralnetworkusingadversariallossforpriorshapeconstraint |