Four Severity Levels for Grading the Tortuosity of a Retinal Fundus Image

Abstract

Hypertensive retinopathy severity classification is proportionally related to tortuosity severity grading. No tortuosity severity scale enables a computer-aided system to classify the tortuosity severity of a retinal image. This work aimed to introduce a machine learning model that can identify the severity of a retinal image automatically and hence contribute to developing a hypertensive retinopathy or diabetic retinopathy automated grading system. First, the tortuosity is quantified using fourteen tortuosity measurement formulas for the retinal images of the AV-Classification dataset to create the tortuosity feature set. Secondly, a manual labeling is performed and reviewed by two ophthalmologists to construct a tortuosity severity ground truth grading for each image in the AV classification dataset. Finally, the feature set is used to train and validate the machine learning models (J48 decision tree, ensemble rotation forest, and distributed random forest). The best performance learned model is used as the tortuosity severity classifier to identify the tortuosity severity (normal, mild, moderate, and severe) for any given retinal image. The distributed random forest model has reported the highest accuracy (99.4%) compared to the J48 Decision tree model and the rotation forest model with minimal least root mean square error (0.0000192) and the least mean average error (0.0000182). The proposed tortuosity severity grading matched the ophthalmologist's judgment. Moreover, detecting the tortuosity severity of the retinal vessels', optimizing vessel segmentation, the vessel segment extraction, and the created feature set have increased the accuracy of the automatic tortuosity severity detection model.

Keywords: blood vessels; decision support system; decision tree; diagnosis; distributed random forest; inflection count metric; retinal images; skeletonization; tortuosity.

Figure 1

The AV-classification dataset used as input to this research work.

Figure 2

The process of measuring the tortuosity severity levels.

Figure 3

Illustration of the tortuosity calculation steps: (A) the fundus image, (B) vessels extraction, (C) intersection points identification, (D) skeletonization, (E) vessel segments of fragments segmentation, (F) tortuosity calculation for the fourteen metrics.

Figure 4

Illustration for arc and chord length in: (a) tortuosity index (TI), (b) tortuosity density, (c) sum of angles metric (SOAM), and (d) Inflection count metric (ICM).

Figure 5

ERD diagram of the image level and segment level feature sets.

Figure 6

Tortuosity morphometric analysis and quantification form.

Figure 7

Illustration of the distributed random forest method used with feature set created from the AV-classification dataset.

Figure 8

The updated RVM Dataset.

Figure 9

A retinal images sample at every tortuosity grade: (A) normal, (B) mild, (C) moderate, and (D) severe.

Figure 10

A comparison graph of the loss reported from using the DRF model of 50 Trees versus a 105 trees DRF model loss, in terms of (MSE, RMSE, MAE, RMSLE, and mean residual deviance) loss measures.

Figure 11

The model training and validation scoring results of RMSE (a) vs. number of trees (b) vs. number of EPOCS (c) and MAE vs. number of EPOCS.