Research on FRIT of Color Fundus Image Based on Prior Knowledge

To carry out image post-processing efficiently, it is necessary representing image with statistics principle in order to increase researchers´ prior knowledge to image. In the essay, a representation about special color fundus images is proposed by extracting corresponding image color space´s component grayscale. And then, the qualitative and quantitative statistical analysis to component images is put forward. The results show that the G-component image and the I-component image, considering the visual effect, are more appropriate for post-processing because of higher contrast and lower noise, which should meet the application requirements of the vessel segmentation. And from the statistical average, the grayscale are ~146; the mean square error (MSE) pixel-based is lower (<;0.26%). Both PSNRs are higher than ~74dB. All those are suitable for the vessels segmentation. Through the novel FRIT-Wiener algorithm for noise filtering, and edge detection with classic algorithm, the results indicate that the G-component image and the I-component image can be suitable for the processing demand needed while the design pressure of processing algorithm should be down obviously.