MICCAI-iseg2017挑战赛小结与婴儿脑组织分割总结
按照數據集進行劃分:
關于自動分割工具,嬰兒腦MR圖像來自單個時間點,其中縱向數據集不可用,因此必須開發不針對縱向數據集的分割工具,目前提出了一些機器學習方法,但這些方法的效果并不令人滿意。
iseg-2017:
競賽結果一共評估了兩次,第一次評估中的TOP-1為MSL_SKKU,第二次評估中降為了第二名,但由于第二次評估的時間較晚,因此已發表的論文大多與第一次評估中的冠軍作對比,在以下的總結中,把第一次評估的第一名作為挑戰賽的冠軍。
第二次評估結果入下:
第二次評估結果MICCAI iseg-2017挑戰賽結果:? 第一名為MSL_SKKU
DICE、MHD、ASD?
DICE結果排名如下:
三項分割標簽各自的DICE排名?
MHD排名結果如下:?
3項分割標簽各自的MHD排名?
ASD排名結果如下:
3項分割標簽各自的ASD排名?
?
前五名的結果為:
top-5?
與數據集相關的論文:
Bui T D, Shin J, Moon T. 3d densely convolution networks for volumetric segmentation[J]. arXiv preprint arXiv:1709.03199, 2017.?
?---10引用? ?采用DenseNet? ? 挑戰賽中排名第一? ?隊伍名:MSL-SKKU
取小數點后3位的情況下9項指標6項第一
Dolz J, Desrosiers C, Wang L, et al. Deep CNN ensembles and suggestive annotations for infant brain MRI segmentation[J]. arXiv preprint arXiv:1712.05319, 2017.? ? ? ?
? ---6引用? ?采用SemiDenseNet? ?在某些指標中排名第一或者第二,結果如下:
?
Dolz J, Ayed I B, Yuan J, et al. Isointense infant brain segmentation with a hyper-dense connected convolutional neural network[C]//Biomedical Imaging (ISBI 2018), 2018 IEEE 15th International Symposium on. IEEE, 2018: 616-620.?
??---3引用? ?采用DenseNet? ?9個指標中6個排名前三
該文章基于另外兩篇文章:
?Konstantinos Kamnitsas et al., "Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation",?Medical image analysis, vol. 36, pp. 61-78, 2017.?
Jose Dolz et al., "3D fully convolutional networks for subcortical segmentation in MRI: A large-scale study",?NeuroImage, 2017.
?
Dolz J, Gopinath K, Yuan J, et al. HyperDense-Net: A hyper-densely connected CNN for multi-modal image segmentation[J]. arXiv preprint arXiv:1804.02967, 2018.
?
iseg-2017?
取小數點后兩位的情況下,都為9項指標7項第一
對于MRBrainS2013:
MRBrainS2013提交時間為18.02.16,提交時為第一名,目前排名第6,對比結果如下:
| HyperDenseNet(top-6) | 0.8633 | 1.34 | 6.19 | 0.8946 | 1.78 | 6.03 | 0.8342 | 2.26 | 7.31 |
| XMU_SmartDSP2(top-1) | 0.865 | 1.29 | 5.75 | 0.899 | 1.73 | 5.47 | 84.8 | 1.84 | 6.83 |
Dolz J, Ayed I B, Yuan J, et al. HyperDense-Net: A hyper-densely connected CNN for multi-modal image semantic segmentation[J]. arXiv preprint arXiv:1710.05956, 2017.
? ---未了解(上一篇的效果較好)
Fonov V, Doyle A, Evans A C, et al. NeuroMTL iSEG challenge methods[J]. bioRxiv, 2018: 278465.
? ---排名較靠前? ?隊伍名:NeuroMTL
Sanroma G, Benkarim O M, Piella G, et al. Learning to combine complementary segmentation methods for fetal and 6-month infant brain MRI segmentation[J]. Computerized Medical Imaging and Graphics, 2018, 69: 52-59.???
? ---非頂會,效果不好但拿不到數據集
Zeng G, Zheng G. Multi-stream 3D FCN with multi-scale deep supervision for multi-modality isointense infant brain MR image segmentation[C]//Biomedical Imaging (ISBI 2018), 2018 IEEE 15th International Symposium on. IEEE, 2018: 136-140.??
? ---排名第三 隊伍名:Bern_IPMI? 采用FCN?
Wang Z, Zou N, Shen D, et al. Global Deep Learning Methods for Multimodality Isointense Infant Brain Image Segmentation[J]. arXiv preprint arXiv:1812.04103, 2018.
-----預印本,目前最新效果最好的的論文(WM與GM為最佳性能,CSF具有可比性),提交時間為2018.12.10,未參與挑戰賽排名,實驗結果為:
| ? | WM | GM | CSF |
| top-1 | 0.901 | 0.919 | 0.958 |
| 論文 | 0.9044 | 0.9219 | 0.9557 |
Li T, Zhou F, Zhu Z, et al. A label-fusion-aided convolutional neural network for isointense infant brain tissue segmentation[C]//Biomedical Imaging (ISBI 2018), 2018 IEEE 15th International Symposium on. IEEE, 2018: 692-695.?
使用的數據集為 iseg-2017,使用的網絡為FCNN,實驗結果接近第一名,實驗結果如下:
CSF指標相同Kumar S, Conjeti S, Roy A G, et al. InfiNet: Fully convolutional networks for infant brain MRI segmentation[C]//Biomedical Imaging (ISBI 2018), 2018 IEEE 15th International Symposium on. IEEE, 2018: 145-148.?
數據集來源于BCP項目,該數據集作為iseg-2017挑戰賽數據集的一部分(總數,訓練測試集數一樣),結果比挑戰賽TOP-1低2%,但參數減半。
3D-DenseNet為挑戰賽第一名的結果推薦的文章為:但3篇文章都為預印本
1.Bui T D, Shin J, Moon T. 3d densely convolution networks for volumetric segmentation[J]. arXiv preprint arXiv:1709.03199, 2017.?
?---10引用? ?采用DenseNet? ? 挑戰賽中排名第一? ?隊伍名:MSL-SKKU
取小數點后3位的情況下9項指標6項第一
2.Dolz J, Gopinath K, Yuan J, et al. HyperDense-Net: A hyper-densely connected CNN for multi-modal image segmentation[J]. arXiv preprint arXiv:1804.02967, 2018.
--取小數點后2為的情況下都為9項指標7項第一
3.Wang Z, Zou N, Shen D, et al. Global Deep Learning Methods for Multimodality Isointense Infant Brain Image Segmentation[J]. arXiv preprint arXiv:1812.04103, 2018.
-----預印本,目前最新效果最好的的論文(WM與GM為最佳性能,CSF具有可比性),提交時間為2018.12.10,未參與挑戰賽排名
?
NDAR數據集
Wang L, Li G, Shi F, et al. Volume-based analysis of 6-month-old infant brain MRI for autism biomarker identification and early diagnosis[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, Cham, 2018: 411-419.
采用U-Net,數據集來源于NDAR,包含18名受試者,分割結果很好------未找到數據集
分割結果?
neobrains(類別較多)--相關文章較少
第一名排名結果如下(在大多數指標中排名第一)提交時間為2018.09
在40周矯正年齡時獲得軸向掃描
| ? | CoGM | UMW | CSF | Vent | CB | BS | BGT | MWM | UWM+MWM | CSF+Vent |
| DC | 0.89 | 0.94 | 0.85 | 0.90 | 0.95 | 0.86 | 0.94 | 0.53(2) | 0.93(3) | 0.85 |
| MSD | 0.1 | 0.09 | 0.16 | 0.14 | 0.26 | 0.24 | 0.30 | 0.60 | 0.10 | 0.16 |
| HD | 18.79 | 10.57 | 8.66 | 13.76 | 16.40 | 7.13 | 19.84 | 9.19 | 7.83 | 8.27 |
在30周矯正年齡時獲得冠狀位掃描
| ? | CoGM | UWM | CSF | Vent | CB | BS | BGT | MWM | UMW+MWM | SCF+Vent |
| DC | 0.79 | 0.95 | 0.89 | 0.87 | 0.91 | 0.84 | 0.89 | - | 0.95 | 0.91 |
| MSD | 0.14 | 0.14 | 0.13 | 0.33 | 0.30 | 0.37 | 0.42 | - | 0.14 | 0.10 |
| HD | 13.81 | 9.16 | 6.43 | 13.69 | 7.09 | 7.54 | 7.98 | - | 9.99 | 4,96 |
在40周矯正年齡時獲得冠狀位掃描
| ? | CoGM | UWM | CSF | Vent | CB | BS | BGT | MWM | UWM+MWM | CSF+Vent |
| DC | 0.79 | 0.91 | 0.82 | 0.82 | 0.89 | 0.68 | 0.86 | 0.04 | 0.91 | 0.83 |
| MSD | 0.21 | 0.18 | 0.30 | 0.41 | 0.65 | 1.13 | 0.98 | 7.07 | 0.22 | 0.28 |
| HD | 27.20 | 24.00 | 12.86 | 15.42 | 26.18 | 15.15 | 25.91 | 27.92 | 13.75 | 12.52 |
使用了該數據集的部分文章:
- Sanroma et al., In: Machine Learning in Medical Imaging (MICCAI), 2016, 27–35
- Moeskops et al., IEEE Transactions on Medical Imaging, 2016, 35(5):1252–1261
- Beare et al., Frontiers in Neuroinformatics, 2016, 10:12
- Moeskops et al., PLOS ONE, 2015, 10(7):e0131552
- Moeskops et al., NeuroImage, 2015, 118:628–641
- Cherel et al., In: SPIE Medical Imaging, 2015, 9413:941311
- Wang et al., NeuroImage, 2015, 108:160–172
- Wang et al., In: Medical Computer Vision (MICCAI), 2014, 8848:22–33
- Chita et al., In: SPIE Medical Imaging, 2013, 8669:86693
- Moeskops et al., In: SPIE Medical Imaging, 2013, 8670:867011
?
?
?
其他
1.Nie D, Wang L, Adeli E, et al. 3-D fully convolutional networks for multimodal isointense infant brain image segmentation[J]. IEEE Transactions on Cybernetics, 2018.? ? 1區
2.Bernal J, Kushibar K, Cabezas M, et al. Quantitative analysis of patch-based fully convolutional neural networks for tissue segmentation on brain magnetic resonance imaging[J]. arXiv preprint arXiv:1801.06457, 2018.
?3.Chen J, Zhang H, Nie D, et al. Automatic accurate infant cerebellar tissue segmentation with densely connected convolutional network[C]//International Workshop on Machine Learning in Medical Imaging. Springer, Cham, 2018: 233-240.
MICCAI挑戰賽底部文章:
[1]. Li Wang, Yaozong Gao, Feng Shi, Gang Li, John H. Gilmore, Weili Lin, Dinggang Shen.?LINKS: Learning-based multi-source IntegratioN frameworK for Segmentation of infant brain images, Neuroimage, 108, 160-172, 2015.
[2]. Li Wang, Feng Shi, Yaozong Gao, Gang Li, John H. Gilmore, Weili Lin, Dinggang Shen.?Integration of Sparse Multi-modality Representation and Anatomical Constraint for Isointense Infant Brain MR Image Segmentation, Neuroimage, 89, 152-164, 2014.
[3]. Wenlu Zhang, Rongjian Li, Houtao Deng, Li Wang, Weili Lin, Shuiwang Ji, Dinggang Shen.?Deep Convolutional Neural Networks for Multi-Modality Isointense Infant Brain Image Segmentation, Neuroimage, 108, 214–224, 2015.
綜述文章:
Li G, Wang L, Yap P T, et al. Computational neuroanatomy of baby brains: A review[J]. Neuroimage, 2018.????
? ---引用9 ???綜述文章
A review on automatic fetal and neonatal brain MRI segmentation
? ---綜述文章
Convolutional Neural Networks for Rapid and Simultaneous Brain Extraction and Tissue Segmentation? ?書籍
?
?
?
?
?
?
?
?
?
?
?
?
?
?
吃雞加我:360258904
總結
以上是生活随笔為你收集整理的MICCAI-iseg2017挑战赛小结与婴儿脑组织分割总结的全部內容,希望文章能夠幫你解決所遇到的問題。
- 上一篇: 有人负责,才有质量:写给在集市中迷失的一
- 下一篇: Java综合项目----开发团队分配管理