Quantitation of vital bleaching by computer analysis of photographic images

Background. The authors investigated the use of computer processing of phot ographic images to monitor changes in tooth brightness after nightguard vit al bleaching, or NGVB. Methods. Photographs of shade guides and clinical cases (patients' teeth) w ere taken on 35-millimeter film with electronic flash illumination and proc essed commercially. A slide scanner was used to digitize images as red, gre en and blue, or RGB, files, with constant brightness, contrast and linearit y settings; the images were then analyzed with commercial software. Relevan t image components (that is, teeth or shade guide tabs) were separated, and histograms of various numerical color descriptors were generated for each image component. Results. Analysis of shade tab images showed that the mean pixel intensity for the RGB blue channel, or MPIb, was the most satisfactory brightness des criptor, with clear sequential MPIb increments from lighter to darker shade s in each series of colors (A through D) and close correlation with the man ufacturer's brightness scale (r = .83). Mathematical analysis of MPIb data for shade tabs in the same image yielded a brightness index that was reprod ucible and correlated well with the manufacturer's brightness scale. Sequen tial measurements of this index in three subjects whose teeth were bleached with carbamide peroxide for 14 days correlated well with assessments made by visual shade guide comparisons. Conclusions. The authors conclude that computer analysis of digitized photo graphic images with internal color controls provides an index of tooth brig htness that is reproducible from image to image. Clinical Implications. A brightness index derived from computer analysis of digitized photographic images may be useful for monitoring the effectivene ss of NGVB.