The application of neural networks in the work of forensic experts in child abuse cases

Abstract

Purpose: One of the tasks that court experts in the field of forensic sexology have to perform is the assessment of secured pornographic materials involving minors. Forensic experts use their specialist knowledge to answer questions posed by the procedural authorities, including whether the material may induce sexual stimulation and whether an offender may be identified as having a disorder of sexual preference in the form of pedophilia. The aim of the article is to present the possibility of using neural networks in forensic sexology. Views: Neural networks are mathematical structures whose basic elements are artificial neurons modelled on the work of biological neurons. They are used in a variety of commercial and scientific tasks. Models for classifying pornographic materials (both images and films) and for estimating the age of the minors presented in these images are introduced. Neural networks can be used to categorize pornographic materials in the context of the growing levels of sexualization of minors. Moreover, the training of neural networks to classify specific objects in the pornographic material shown in these images could allow for the differentiation between the various categories of pornographic materials involving minors. Conclusions: Neural networks can be widely used in forensic sexology as an element supporting the work of forensic experts. The presented research results seem to be very promising, but the area requires further research.