ABSTRACT
In this work, we proposed and validated a novel transfer learning based method for automatic cornea segmentation from ocular staining images. An encoder-decoder structure was used, containing an Xception feature extractor, Atrous Spatial Pyramid Pooling and a simple but effective decoder. The proposed method successfully solved the problem of training a large network with limited data and capturing contextual information with a relatively simple network by fine-tuning a large network that had been pretrained on large-scale common object datasets. A total of 712 ocular staining images were used in our experiments. Evaluations were conducted both quantitatively and qualitatively through 5-fold cross-validation. Compared with several other deep learning based segmentation techniques, including U-Net, Fully Convolutional Network (FCN), and DeepLab, the proposed method was found to exhibit superior performance, with a Dice score of 95.82%, sensitivity of 95.37%, and accuracy of 97.63%. We also evaluated the method with different transfer learning strategies and found that fine-tuning the whole network worked the best in terms of segmentation accuracy.
- Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. 2018. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European conference on computer vision (ECCV). 801–818.Google ScholarDigital Library
- François Chollet. 2017. Xception: Deep learning with depthwise separable convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition. 1251–1258.Google ScholarCross Ref
- Elisabeth J Cohen, Peter R Laibson, Juan J Arentsen, and Carol S Clemons. 1987. Corneal ulcers associated with cosmetic extended wear soft contact lenses. Ophthalmology 94, 2 (1987), 109–114.Google ScholarCross Ref
- Alexander de Brebisson and Giovanni Montana. 2015. Deep neural networks for anatomical brain segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 20–28.Google ScholarCross Ref
- Lijie Deng, Junyan Lyu, Haixiang Huang, Yuqing Deng, Jin Yuan, and Xiaoying Tang. 2020. The SUSTech-SYSU dataset for automatically segmenting and classifying corneal ulcers. Scientific Data 7, 1 (2020), 1–7.Google ScholarCross Ref
- Pedro F Felzenszwalb, Ross B Girshick, David McAllester, and Deva Ramanan. 2009. Object detection with discriminatively trained part-based models. IEEE transactions on pattern analysis and machine intelligence 32, 9(2009), 1627–1645.Google Scholar
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition. 770–778.Google ScholarCross Ref
- Daniel S Kermany, Michael Goldbaum, Wenjia Cai, Carolina CS Valentim, Huiying Liang, Sally L Baxter, Alex McKeown, Ge Yang, Xiaokang Wu, Fangbing Yan, 2018. Identifying medical diagnoses and treatable diseases by image-based deep learning. Cell 172, 5 (2018), 1122–1131.Google ScholarCross Ref
- Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. 2014. Microsoft coco: Common objects in context. In European conference on computer vision. Springer, 740–755.Google ScholarCross Ref
- Jonathan Long, Evan Shelhamer, and Trevor Darrell. 2015. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition. 3431–3440.Google ScholarCross Ref
- Sinno Jialin Pan and Qiang Yang. 2009. A survey on transfer learning. IEEE Transactions on knowledge and data engineering 22, 10(2009), 1345–1359.Google ScholarDigital Library
- Olaf Ronneberger, Philipp Fischer, and Thomas Brox. 2015. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention. Springer, 234–241.Google ScholarCross Ref
- Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, 2015. Imagenet large scale visual recognition challenge. International journal of computer vision 115, 3 (2015), 211–252.Google ScholarDigital Library
- Xian Nan Tang, Ari Ethan Berman, Raymond Alan Swanson, and Midori Anne Yenari. 2010. Digitally quantifying cerebral hemorrhage using Photoshop® and Image J. Journal of neuroscience methods 190, 2 (2010), 240–243.Google ScholarCross Ref
- Yue Zhang, Jiong Wu, Chen Wanli, Chen Yifan, and Xiaoying Tang. 2019. Prostate Segmentation Using Z-Net. In 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019). IEEE, 11–14.Google Scholar
Recommendations
Image clustering segmentation based on SLIC superpixel and transfer learning
Traditional fuzzy C-means clustering algorithm has poor noise immunity and clustering results in image segmentation. To overcome this problem, a novel image clustering algorithm based on SLIC superpixel and transfer learning is proposed in this paper. ...
Segmentation of dermoscopy image using adversarial networks
Skin lesion image segmentation task has many difficulties due to the hair occlusion or low contrast. Traditional methods based image quality are powerless when processing these kinds of skin images. In this paper, a segmentation architecture based ...
Deep Convolutional Neural Networks for Brain Tumor Segmentation: Boosting Performance Using Deep Transfer Learning: Preliminary Results
Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain InjuriesAbstractBrain tumor segmentation through MRI images analysis is one of the most challenging issues in medical field. Among these issues, Glioblastomas (GBM) invade the surrounding tissue rather than displacing it, causing unclear boundaries, furthermore, ...
Comments