선택적 커널 합성곱을 이용한 U-Net 기반 망막 혈관 분할 모델
SKU-Net: Improved U-Net using Selective Kernel Convolution for Retinal Vessel Segmentation
1
조회
0
다운로드
2
인용
0
참고문헌
키워드
초록
본 논문에서는 안저영상의 다중 스케일 정보를 다루기 위한 딥러닝 기반의 망막 혈관 분할 모델을 제안한다. 제안 모델은 이미지 분할 딥러닝 모델인 U-Net과 선택적 커널 합성곱을 통합한합성곱 신경망으로 안저영상에서 눈과 관련된 질병을 진단하는데 중요한 정보가 되는 망막 혈관의 다양한 모양과 크기를 갖는 특징 정보를 추출하고 분할한다. 제안 모델은 일반적인 합성곱과선택적 커널 합성곱으로 구성된다. 일반적인 합성곱 층은 같은 크기 커널 크기를 통해 정보를 추출하는 반면, 선택적 커널 합성곱은 다양한 커널 크기를 갖는 브랜치들에서 정보를 추출하고 이를 분할 주의집중을 통해 적응적으로 조정하여 결합한다. 제안 모델의 성능 평가를 위해 안저영상 데이터인 DRIVE와 CHASE DB1 데이터셋을 사용하였으며 제안 모델은 두 데이터셋에 대하여F1 점수 기준 82.91%, 81.71%의 성능을 보여 망막 혈관 분할에 효과적임을 확인하였다.
Abstract
In this paper, we propose a deep learning-based retinal vessel segmentation model for handling multi-scale information of fundus images. we integrate the selective kernel convolution into U-Net-based convolutional neural network. The proposed model extracts and segment features information with various shapes and sizes of retinal blood vessels, which is important information for diagnosing eye-related diseases from fundus images. The proposed model consists of standard convolutions and selective kernel convolutions. While the standard convolutional layer extracts information through the same size kernel size, The selective kernel convolution extracts information from branches with various kernel sizes and combines them by adaptively adjusting them through split-attention. To evaluate the performance of the proposed model, we used the DRIVE and CHASE DB1 datasets and the proposed model showed F1 score of 82.91% and 81.71% on both datasets respectively, confirming that the proposed model is effective in segmenting retinal blood vessels.
참고문헌 (24)
- D Marín, "A New Supervised Method for Blood Vessel Segmentation in Retinal Images by Using Gray-Level and Moment Invariants-Based Features", IEEE Transactions on Medical Imaging, 30(1), pp.146~158 (2011), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- M. U. Akram, "Blood vessel enhancement and segmentation using wavelet transform", 2009 International Conference on Digital Image Processing, pp.34~38 (2009), IEEE
- D. Onkaew, "Automatic Extraction of Retinal Vessels Based on Gradient Orientation Analysis", IEEE Eighth International Joint Conference on Computer Science and Software Engineering(JCSSE), pp.102~107 (2011)
- Alex Krizhevsky, "ImageNet classification with deep convolutional neural networks", Communications of the ACM, 60(6), pp.84~90 (2017), Association for Computing Machinery (ACM)[DOI]
- Jose Ignacio Orlando, "A Discriminatively Trained Fully Connected Conditional Random Field Model for Blood Vessel Segmentation in Fundus Images", IEEE Transactions on Biomedical Engineering, 64(1), pp.16~27 (2017), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- D. Maji, "Deep neural network and random forest hybrid architecture for learning to detect retinal vessels in fundus images", 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp.3029~3032 (2015)
- Shahab Aslani, "A new supervised retinal vessel segmentation method based on robust hybrid features", Biomedical Signal Processing and Control, 30, pp.1~12 (2016), Elsevier BV[DOI]
- J. Long, "Fully convolutional networks for semantic segmentation", Proceedings of the IEEE conference on computer vision and pattern recognition(CVPR), pp.3431~3440 (2015)
- O. Ronneberger, "U-net : Convolutional networks for biomedical image segmentation", International Conference on Medical image computing and computer-assisted intervention(MICCAI), 234-241, pp.234~241 (2015), Springer
- 손정우, "3D GPR 탐사 데이터와 심층 합성곱 신경망을 이용한 지하 관로 자동 탐색 시스템", 한국컴퓨터정보학회논문지, 26(2), pp.27~37 (2021), 한국컴퓨터정보학회[DOI]
- 김수민, "U-Net에 기반한 이미지 복원 기법을 이용한 콘크리트 균열 탐지 방안", 한국컴퓨터정보학회논문지, 25(10), pp.35~42 (2020), 한국컴퓨터정보학회[DOI]
- Zhengxin Zhang, "Road Extraction by Deep Residual U-Net", IEEE Geoscience and Remote Sensing Letters, 15(5), pp.749~753 (2018), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Md Zahangir Alom, "Recurrent residual U-Net for medical image segmentation", Journal of Medical Imaging, 6(01), pp.1~ (2019), SPIE-Intl Soc Optical Eng[DOI]
- Xiaomeng Li, "H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tumor Segmentation From CT Volumes", IEEE Transactions on Medical Imaging, 37(12), pp.2663~2674 (2018), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- J. Zhuang, "Laddernet: Multi-path networks based on u-net for medical image segmentation", arXiv preprint arXiv:1810.07810, ~ (2018)
- L. Li, "Iternet : Retinal image segmentation utilizing structural redundancy in vessel networks", IEEE Winter Conference on Applications of Computer Vision(WACV), pp.3656~3665 (2020)
- Qiangguo Jin, "DUNet: A deformable network for retinal vessel segmentation", Knowledge-Based Systems, 178, pp.149~162 (2019), Elsevier BV[DOI]
- G. Ghiasi, "Selective Kernel Networks", IEEE, 2019, 510~519
- K. He, "Deep residual learning for image recognition", IEEE Conference on Computer Vision and Pattern Recognition(CVPR), pp.770~778 (2016)
- X. Li, "Dropblock: A regularization method for convolutional networks", Neural Information Processing Systems (2018)
- S. Ioffe, "Batch normalization: Accelerating deep network training by reducing internal covariate shift", arXiv preprint arXiv:1502.03167, ~ (2015)
- T. Laibacher, "M2U-Net : Effective and efficient retinal vessel segmentation for resource-constrained environments", IEEE, 2019, 115~124
- C. Guo, "Residual Spatial Attention Network for Retinal Vessel Segmentation", International Conference on Neural Information Processing, ICONIP 2020, pp.509~519 (2020)
- T. DeVries, "Improved Regularization of Convolutional Neural Networks with Cutout", arXiv preprint arXiv:1708.04552, ~ (2017)