segmentation of optic cup Hong Kang,Kai Wang,Song Guo,Yingqi Gao, - PowerPoint PPT Presentation

Pixel quantification for robust segmentation of optic cup Hong Kang,Kai Wang,Song Guo,Yingqi Gao, Ning Li,Jinyuan Weng,Xiaoxing Li,Tao Li 上工医信 Beijing Shanggong HIS. Technology Co., Ltd 1/10

Pipeline binary Component Analysis Pixel Quantification DeepLab v3+ Classification Fundus image binary Component Analysis Figure 1. An overview of our proposed pipeline 2/10

Pipeline (Pixel Quantification) Pixel quantification is used to reduce the sensitivity of the segmentation model to color. Eq (1) Figure 2. Fundus images from training set and validation set 𝑦 is an image from training set, 𝑠 is a hyper-parameter. 3/10 Figure 3. Fundus image after pixel quantification

Pipeline (Segmentation model) Patch: 900*900 Data augmentation: Rotation, flipping, scaling Hyper-parameters: learning rate 0.0005 max iteration 100000 momentum 0.9 weight decay 0.00004 Figure 4. An overview of DeepLab v3+ 4/10

Pipeline (Component Analysis) False positive reduction: • Binary probability map through a fixed threshold (hyper-parameter) • Find all connected component in the binary map • Reserve the largest connected component (a) (b) Figure 5. (a) binary map (b) processed binary map 5/10

Pipeline (Classification) Glaucoma classification: • Calculate vertical CDR (V-CDR) • V-CDR is treated as the probability of getting glaucoma 𝑤𝑑𝑒𝑠 = 𝑠𝑑 Eq (2) 𝑠𝑒 Figure 6. An example abot the calculation of vcdr. 6/10

Result Ablation study on pixel quantification • Pixel quantification can lead to 0.029 improvement in OC segmentation • Pixel quantification can improve model robustness to make it suitable for different fundus cameras. Table 1. Segmentation results on validation set 7/10

Result Segmentation Leaderboard 8/10

Result Overall Leaderboard 9/10

Conclusion Our solution   A pipeline for optic cup, optic disc segmentation and glaucoma classification  A robust segmentation model suitable for different fundus cameras  Feature work  Interpretable deep learning model for glaucoma screening 10/10

Thanks! Q&A 上工医信 Beijing Shanggong HIS. Technology Co., Ltd