July 2019
Volume 60, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2019
Fully automated artificial intelligence (AI) pipeline for feature-based segmentation and classification of diabetic retinopathy in fundus photographs
Author Affiliations & Notes
  • yue wu
    UW, Seattle, Washington, United States
  • Fenghua Wang
    Eye Institute Reading Center, Shanghai Jiaotong University, Shanghai, China
    Ophthalmology, Shanghai General Hospital, Shanghai, China
  • Sa Xiao
    UW, Seattle, Washington, United States
  • Yuka Kihara
    UW, Seattle, Washington, United States
  • Ted Spaide
    UW, Seattle, Washington, United States
  • Cecilia S Lee
    UW, Seattle, Washington, United States
  • Aaron Y Lee
    UW, Seattle, Washington, United States
    eScience, University of Washington, Seattle, Washington, United States
  • Footnotes
    Commercial Relationships   yue wu, None; Fenghua Wang, None; Sa Xiao, None; Yuka Kihara, None; Ted Spaide, None; Cecilia Lee, None; Aaron Lee, Carl-Zeiss Meditec Inc (F), Novartis Pharmaceuticals (F), Topcon Corporation (S)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science July 2019, Vol.60, 2205. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      yue wu, Fenghua Wang, Sa Xiao, Yuka Kihara, Ted Spaide, Cecilia S Lee, Aaron Y Lee; Fully automated artificial intelligence (AI) pipeline for feature-based segmentation and classification of diabetic retinopathy in fundus photographs. Invest. Ophthalmol. Vis. Sci. 2019;60(9):2205.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : Diabetic retinopathy (DR) is a leading cause of blindness worldwide. Early diagnosis and reliable detection of progression are critical for better prognosis. Currently, DR is classified by severities but evaluating clinical features directly may improve the accuracy of classifications and provide important insights in DR pathology or progression. We developed a deep learning pipeline that not only classifies DR severity, but also quantifies all major DR features.

Methods : 55 fundus images were manually segmented by reading center experts and used to train a deep segmentation network that locates all major diabetic retinopathy features, namely microaneurysms, hemorrhages, exudates, cottonwool spots, neovascularization, intraretinal microvascular abnormality and venous beading, at high resolution. To investigate the usefulness of the segmentations, this network was used to segment DR features in Kaggle Diabetic Retinopathy competition (EyePACS). We then trained three different classification networks: a traditional convolutional neural network, and two pyramidal networks with fractional max pooling (a shallower and a deeper model), to predict DR severity on the Kaggle data with and without the segmentation masks.

Results : Representative segmentations from the validation set are shown in Figure 2. Higher classification accuracy were achieved with masks (69.2%, 66.2%, 70.1%) than without (66.9%, 63.2%, 68.9%) for the 3 networks. Random forest ensembling improved accuracies to 83% vs 78% for with and without respectively.

Conclusions : We present a new AI pipeline that improves the accuracy of DR classification by incorporating DR feature segmentation. The pipeline yields segmentations that can help clinicians identify and locate clinically relevant DR features, thereby enabling a new generation of feature-based studies of DR diagnosis and progression rather than the traditional severity scheme, leading to novel hypotheses and analyses.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

 

Schematic of AI pipeline for fully automated DR feature segmentation and severity classification.

Schematic of AI pipeline for fully automated DR feature segmentation and severity classification.

 

Representative feature segmentation. A,D,G. fundus image; B,E,H. Expert segmented; C,F,I. AI segmented. Colors: red, microaneurysm; green, exudate; yellow, hemorrhage; blue, cottonwool spot; purple, neovascularization. Predicted class probabilities for NO DR, Mild, Moderate, Severe NPDR & PDR are shown in last column.

Representative feature segmentation. A,D,G. fundus image; B,E,H. Expert segmented; C,F,I. AI segmented. Colors: red, microaneurysm; green, exudate; yellow, hemorrhage; blue, cottonwool spot; purple, neovascularization. Predicted class probabilities for NO DR, Mild, Moderate, Severe NPDR & PDR are shown in last column.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×