Adaptive image enhancement for retinal blood vessel segmentation

An adaptive image enhancement method was proposed in which the background image was estimated and subtracted from the original image. The original image was modeled as the composite of blood vessels and background under the influence of non-uniform illumination function of the imaging system. Background subtraction of retinal blood vessels (BSRBV) was found to be a robust algorithm that provided better input images for subsequent image segmentation processing. The resulting image was found free from background interference and suitable for subsequent segmentation.