Brain Tumors

项目名称:脑肿瘤分割
编程语言:Keras/Python


Introduction:

Task 1: Segmentation of gliomas in pre-operative MRI scans.
       The participants are called to address this task by using the provided clinically-acquired training data to develop their method and produce segmentation labels of the different glioma sub-regions. The sub-regions considered for evaluation are: 1) the “enhancing tumor” (ET), 2) the “tumor core” (TC), and 3) the “whole tumor” (WT) [see figure below]. The ET is described by areas that show hyper-intensity in T1Gd when compared to T1, but also when compared to “healthy” white matter in T1Gd. The TC describes the bulk of the tumor, which is what is typically resected. The TC entails the ET, as well as the necrotic (fluid-filled) and the non-enhancing (solid) parts of the tumor. The appearance of the necrotic (NCR) and the non-enhancing (NET) tumor core is typically hypo-intense in T1-Gd when compared to T1. The WT describes the complete extent of the disease, as it entails the TC and the peritumoral edema (ED), which is typically depicted by hyper-intense signal in FLAIR.
       The labels in the provided data are: 1 for NCR & NET2 for ED4 for ET, and 0 for everything else.
       The participants are called to upload their segmentation labels into CBICA’s Image Processing Portal for evaluation.

Task 2: Prediction of patient overall survival (OS) from pre-operative scans.
       Once the participants produce their segmentation labels in the pre-operative scans, they will be called to use these labels in combination with the provided multimodal MRI data to extract imaging/radiomic features that they consider appropriate, and analyze them through machine learning algorithms, in an attempt to predict patient OS. The participants do not need to be limited to volumetric parameters, but can also consider intensity, morphologic, histogram-based, and textural features, as well as spatial information, and glioma diffusion properties extracted from glioma growth models.
       Note that participants are expected to provide predicted survival status only for subjects with resection status of GTR (i.e., Gross Total Resection).
      The participants are called to upload a .csv file with the subject ids and the predicted survival values into CBICA’s Image Processing Portal for evaluation.

Feel free to send any communication related to the BraTS challenge to brats2018@cbica.upenn.edu


Process:


Model:

paper:Multi-level Activation for Segmentation of Hierarchically-nested Classes on 3D-Unet


Tricks:

MethodsName
NormalizationBN / LN / IN / SN
DroupoutS dropout / dropblock / Targeted dropout
Reduce dim1*1 kernel
AugmentSTN / Deformable ConvNets
ContextSPP / ParseNet / PSP / HDC / ASPP / FPA
Updeconvolution / bilinear / DUC / GAU
Crfdense crf
Domainconnected domain
Early stopyes
Lossdice loss / jaccard loss / focal loss
Regularizationl1_l2 / W constraint
ActivationReLU / PreLU / LeakyReLU / ELU

结果:

配置:显卡Nvidia 1060, 图片使用112*112*112
训练:Epoch 175/500 loss: -0.7303 – jaccard: 0.6066 – val_loss: -0.6615 – val_jaccard: 0.5421
Scores:
1.WT:0.8886     
2.TC:0.8049     
3.ET:0.7772
learderboard 2019 :leardboard

Continue reading Brain Tumors