Craniofacial superimposition in its classic form requires a skull and a photograph of the suspected individual. This image of the person could be taken from a variety of sources and mediums, which are of differing quality and resolution. This can be problematic in cases where the quality of the images is so poor as to obstruct the accurate identification of the cephalometric landmarks. A solution to this problem could lie in the emerging new developments in the field of artificial intelligence and deep learning. Different software applications using these technologies were developed with the aim of image enhancement and resolution upscaling that show promising results. The aim of this study is to establish the strengths and weakness of the afore mentioned technology, by comparing images that have been upscaled using the traditional algorithms to those with AI upscaling. For the experiment, a human skull was photographed since it has a lot of fine detail that can be easily obscured by a photo with bad quality. The photo was then processed which resulted with 10 images, that were compared both visually and with a special software. The visual comparison shows that the Ai upscaled images appear sharper and with a finer detail than the traditionally upscaled ones. However, the data gained from the comparison software showed that the AI upscaled images contained more errors than the traditionally upscaled ones.
HTML PDFShare this article
Journal of Forensic Medicine received 165 citations as per Google Scholar report
