Patch-based fully convolutional neural network with skip connections for retinal blood vessel segmentation

@article{Feng2017PatchbasedFC,
  title={Patch-based fully convolutional neural network with skip connections for retinal blood vessel segmentation},
  author={Zhongwei Feng and Jie Yang and Lixiu Yao},
  journal={2017 IEEE International Conference on Image Processing (ICIP)},
  year={2017},
  pages={1742-1746}
}
Automated segmentation of retinal blood vessels plays an important role in the computer aided diagnosis of retinal diseases. The paper presents a new formulation of patch-based fully Convolutional Neural Networks (CNNs) that allows accurate segmentation of the retinal blood vessels. A major modification in this retinal blood vessel segmentation task is to improve and speed-up the patch-based fully CNN training by local entropy sampling and a skip CNN architecture with class-balancing loss. The… CONTINUE READING

6 Figures & Tables

Extracted Numerical Results

  • The proposed method is experimented on DRIVE dataset and achieves strong performance and significantly outperforms the-state-of-the-art for retinal blood vessel segmentation with 78.11% sensitivity, 98.39% specificity, 95.60% accuracy, 87.36% precision and 97.92% AUC score respectively.
  • As for F1 score, the proposed system equals 0.8183 compared with 0.7382 in DRIU.

Topics