Deep Retinal Image Segmentation
Under Geometrical Priors


In Collaboration with
Palo Alto Research Center (PARC)


Vessel segmentation of retinal images is a key diagnostic capability in ophthalmology. This problem faces several challenges including low contrast, variable vessel size and thickness, and presence of interfering pathology such as micro-aneurysms and hemorrhages. Early approaches addressing this problem employed hand-crafted filters to capture vessel structures, accompanied by morphological post-processing. More recently, deep learning techniques have been employed with significantly enhanced segmentation accuracy. We propose a novel domain enriched deep network that consists of two components: 1) a representation network that learns geometric features specific to retinal images, and 2) a custom designed computationally efficient residual task network that utilizes the features obtained from the representation layer to perform pixel-level segmentation. The representation and task networks are {\em jointly learned} for any given training set. To obtain physically meaningful and practically effective representation filters, we propose two new constraints that are inspired by expected prior structure on these filters: 1) orientation constraint that promotes geometric diversity of curvilinear features, and 2) a data adaptive noise regularizer that penalizes false positives. Multi-scale extensions are developed to enable accurate detection of thin vessels. Experiments performed on three challenging benchmark databases under a variety of training scenarios show that the proposed prior guided deep network outperforms state of the art alternatives as measured by common evaluation metrics, while being more economical in network size and inference time.

Orientation Diversity Regularizer

Representation Layer has the ability to extracts features at different orientation due to the proposed orientaion diversity regularizer.

Noise Regularizer

Representation Layer supresses the curvilinear non-vessel structures due to the noise regularizer.

Multi-Scale Representation Layer

Representation filters trained at multiple scales to capture diversity in vessel thicknesses.

Network Structure


Related Publications

  1. V. Cherukuri, BG V. Kumar, R. Bala and V. Monga, "Deep Retinal Image Segmentation with Regularization Under Geometric Priors", Accepted, IEEE Trans. On Image Processing, 2019. [arXiv]

  2. Cherukuri, Venkateswararao, et al. "Multi-Scale Regularized Deep Network for Retinal Vessel Segmentation." 2019 IEEE International Conference on Image Processing (ICIP). IEEE, 2019. [IEEE Xplore]

Selected References

  1. K.-K. Maninis et al., “Deep retinal image understanding,” in International Conf. on Medical Image Computing and Computer-Assisted Intervention. Springer, 2016, pp. 140–148. [paper].

  2. Liskowski, Paweł, and Krzysztof Krawiec. "Segmenting retinal blood vessels with deep neural networks." IEEE transactions on medical imaging 35.11 (2016): 2369-2380. [paper].

  3. Z. Yan, X. Yang, and K.-T. T. Cheng, “Joint segment-level and pixelwise losses for deep learning based retinal vessel segmentation,” IEEE Trans. on Biomedical Engineering, vol. 65, no. 9, 2018. [paper].

  4. X. Wang, X. Jiang, and J. Ren, “Blood vessel segmentation from fundus image by a cascade classification framework,” Pattern Recognition, vol. 88, pp. 331–341, 2019. [paper].

  5. Z. Fan, J. Lu et al., “A hierarchical image matting model for blood vessel segmentation in fundus images,” IEEE Trans. on Image Processing, vol. 28, no. 5, pp. 2367–2377, May 2019. [paper].


104 Electrical Engineering East,
University Park, PA 16802, USA

Lab Phone: