PAP Smear Image Enhancement Using Diffusion Stop Function Based Clahe Algorithm

Image pre-processing is a very crucial step in cervical cancer(CC) detection from Papanicolaou test images or Pap-smear(PS) image analysis. The slides prepared from pap-test will be given as input to an automated or semi automated system. The system will then analyze the images and will classify as cancerous or non-cancerous. Various researches have been done in the automatic pap-smear analysis and classification. A new approach is presented in this study for PS image enhancement(IE). The images collected from the Mendeley database are enhanced using the DSF-CLAHE(diffusion stop function-contrast limited adaptive histogram equalization) method. DSF is a process in which the edges of images are preserved while smoothening the inside pixels. When compared to some of the existing approaches such as histogram equalization and median filtering, or CLAHE combined with other filters such as adaptive median filter and guided filter, CLAHE when combined with DSF gives better results. The pre-processed results are evaluated with performance matrices MSE(mean squared error), PSNR(peak signal-to-noise ratio), SSIM(structural similarity index measure), and AMBE(absolute mean brightness error) which shows that the proposed method surpass some of the currently available methods.