Main Article Content
Abstract
The performance of various filtering techniques has been analyzed and presented in this paper by targeting speckle noise in retinal images. Most of the state-of-the-art filters are trying to reduce the speckle noise effect from the images without considering the relevant information of blood vessel of retinal images. They also filter the significant features in the targeted images needed for the diagnosis purpose. In order to overcome the limitation of over filtering, the anisotropic diffusion filter is proposed to remove the speckle noise from the retinal image. The performance of the exiting filters and the proposed filter to remove the speckle noise in retinal images is evaluated in terms of the signal-to-noise ratio (SNR), Root Mean Square Error (RMSE) and Standard Deviation (SD) and compared. The experimental results show that the anisotropic diffusion filter is best suitable for removing speckle noise present in the retinal images.