NuI-Go: Recursive Non-Local Encoder-Decoder Network for Retinal Image Non-Uniform Illumination Removal

Chongyi Li 1   Huazhu Fu 2   Runmin Cong 3   Zechao Li 4   Qianqian Xu 5 
1 Nanyang Technological University, Singapore
2 Inception Institute of Artificial Intelligence (IIAI), Abu Dhabi, UAE
3 Beijing Jiaotong University, Beijing, China
4 Nanjing University of Science and Technology, China
5 Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China

Abstract


Retinal images have been widely used by clinicians for early diagnosis of ocular diseases. However, the quality of retinal images is often clinically unsatisfactory due to eye lesions and imperfect imaging process. One of the most challenging quality degradation issues in retinal images is non-uniform which hinders the pathological information and further impairs the diagnosis of ophthalmologists and computer-aided analysis. To address this issue, we propose a non-uniform illumination removal network for retinal image, called NuI-Go, which consists of three Recursive Non-local Encoder-Decoder Residual Blocks (NEDRBs) for enhancing the degraded retinal images in a progressive manner. Each NEDRB contains a feature encoder module that captures the hierarchical feature representations, a non-local context module that models the context information, and a feature decoder module that recovers the details and spatial dimension. Additionally, the symmetric skip-connections between the encoder module and the decoder module provide long-range information compensation and reuse. Extensive experiments demonstrate that the proposed method can effectively remove the non-uniform illumination on retinal images while well preserving the image details and color. We further demonstrate the advantages of the proposed method for improving the accuracy of retinal vessel segmentation.

    Framework

    The network architecture of NuI-Go.. Thue NuI-Go includes three recursive NEDRBs, where each NEDRB contains a feature encoder module, a non-local context module, and a feature decoder module.

Highlights

  1. This work is the first attempt to provide a deep learning solution for the problem of retinal image non-uniform illumination removal, which can effectively correct the illumination of a retinal image while well preserving the original details and color appearances.

  2. We propose a simple yet effective non-local encoder-decoder network to progressively enhance degraded retinal images, which integrates local and non-local information simultaneously.

  3. Benefiting from novel network architecture and physical model-based non-uniform illumination synthesis strategy, our method achieves impressive performance on retinal image non-uniform illumination removal.

Results

1. Non-uniform Illumination Synthesis of Retinal Image


2. Stage-wise Results of NuI-Go


3. Visual Comparisons


4. Retinal Vessel Segmentation Results


Materials



Paper


Code and Model

Citation

@Article{liNuI-GO2020,
          author = {Li, Chongyi and Fu, Huazhu and Cong, Runmin and Li, Zechao and Xu, Qianqian},
          title = {NuI-Go: Recursive Non-Local Encoder-Decoder Network for Retinal Image Non-Uniform Illumination Removal},
          journal = {ACM MM},
          pape={},
          year = {2020}
          }

Contact

If you have any questions, please contact Chongyi Li at lichongyi25@gmail.com.