Color correction of tissue section images by histogram transfer according to control tissues

Aim: In this study, we aimed to present a color correction method by histogram transfer de-pening upon control tissue image (CTI) differences and to evaluate its performance. Methods: Images from colon and placenta sections stained by anti-CD34 were used as CTI and/or sample tissue images (STI). Thirty-six slides but one (standard stained slide with im-ages sCTI and sSTI), were stained for different durations and dilutions than the standard pro-cedures. For H&E (20 slides) and Van Gieson (20 slides) stainings, colon and liver tissues were used. Digital images taken by a CCD camera connected to a light microscope without normalization was stored at a computer. Software was prepared in order to find the histogram difference between two CTIs and transfer the difference to the STI for achieving a corrected STI (corSTI). sSTI (one image) and STI and corSTI (for each images) were semiquantitatively scored by two observers in blind fashion and the STI and corSTI scores were compared with sSTI score. Also total optic density (TOD), median optic density (MOD) and intensity of images were calculated by software. Results: Total number of images was 474. The wKappa was 0,59 for two observers (MEDCALC) for semiquantitative analysis. The STI semiquantitative score was same as the sSTI in 23,5% of the STI but this was 76,35% for corSTI. TOD and intensity values of CD34 stained placenta images and TOD and MOD values of H&E stained colonic mucosa images increased when these images were corrected. Conclusion: It seems the novel approach of this study ?histogram transfer depending upon CTIs? may be a valuable tool for color correction of tissue section images. Key Words: Color normalization, histogram transfer, color correction according to control tissue image