Digital Approaches in Forensic Dentistry Practice: Clustering or Fractal Differentiation?

Abstract

Purpose To evaluate the effectiveness of using nurb-to-nurb, line-to-line, and node-to-node superimposition principles during the person identification procedure based on dental status comparison considering clustering and fractal differentiation of targeted points. Methods.The study included a graphical analysis of 45 skulls that were originally scanned with the use of a CBCT device. Forty-five original *.dcm files were converted into 90 *.stl files (2 copies of each original *.dcm) after image processing. One set of 45 *.stl files was saved, while another 45 copies were deformed with alteration of graphical object integrity via MeshMixer Sofware. Forty-five deformed graphical objects of the skulls were superimposed with originally saved stl-objects using nurb-to-nurb, line-to-line, and node-to-node superimposition principles considering clustering and fractal differentiation of targeted points within original CBCT-files. Results.Nurb-to-nurb superimposition provides the highest validity with the level of “successful identification” up to 79.2±8.4% and the level of “possible identification” up to 94.4±3.2%. The level of graphical disintegration of study object ≥ 45% was considered critical from the point of possibility to reach “successful identification” result. Fractal differentiation provides a higher level of sensitivity regarding truthful identification as compared to the clustering approach both in integral and deform conditions of analyzed skulls (р < 0.05). Conclusion The development of modern digital approaches expands the validity of methods used for the forensic-dental identification process during post-mortem expertise. Digitalization is forming a specific pool of quantitative criteria that also helps to estimate the credibility of diagnosed changes during comparative or reconstructive identification.