Figure 12

Preprocessing step and segmentation of non-pathological retinal fundus image obtained from DRIVE dataset. The smoothed image using the trace-based approach[39](a), the image generated by subtracting the original green channel image from the smoothed one (b), the sharpened image[39](c), the estimated intensity inhomogeneity image obtained by using the median filter (d), the corrected image based on our approach (e), segmented image using the proposed structure-based level set segmentation method (f), segmented result using only green channel of the given image based on the proposed structure-based level set segmentation method without preprocessing (g), and segmented result using only green channel of the histogram-equalized given image based on the proposed structure-based level set segmentation method without preprocessing after 61 iterations (h).