Computational Intelligence and Neuroscience / 2016 / Article / Fig 4

Research Article

Learning-Based Visual Saliency Model for Detecting Diabetic Macular Edema in Retinal Image

Figure 4

Some ROC curves produced by 9 different models from the three examples in Figure 3. ROC of Ours was higher than other models from the 5% to 20% salient region; GBVS, Judd, and ITTI2 got higher ROC when the salient region was larger than 60%. The salient DME RoIs in a retinal image region are generally under 20%. As a result, the performance of our method is better than other models when the salient region is small.
(a)
(b)
(c)