Akademik Veri Yönetim Sistemi
Eastern Journal of Medicine, cilt.31, sa.1, ss.53-63, 2026 (Scopus, TRDizin)
Estimating bruise age by color is unreliable due to low accuracy and confounding variables. The aims of the current study were 1) to compare the RGB values, obtained for age determination of post-traumatic bruises, as identified by four methods in sequence: traditional naked eye and photographic color identifi cation, ImageJ analysis, and artificial intelligence (AI)-supported color identification; 2) to statistically determine the accuracy prediction rates of the bruise aging phase using discriminant function analysis (DFA) of these RGB values; and 3) to assess the usability of these methods in forensic medicine and clinical practice. We examined 407 photographs from 43 patients with traumatic bruises at the University Hospital (2023 –2024). One researcher recorded RGB values during patient examination; two resea rchers used a scale to assess RGB values from photographs; two researchers used ImageJ; and AI analyzed bruise photos. Discriminant function analysis (DFA) assessed bruise-aging group classification using RGB means. The AI-assisted program demonstrated the highest overall accuracy in bruise age estimation (50.1%). In the yellow-dominant group, the 65-year-old researcher exhibited the lowest accuracy (18.3%), whereas the AI-assisted program achieved perfect accuracy (100.0%). Visual identification by the naked eye was more accurate compared to other non-AI digital methods. These findings indicate that AI-based color analysis, which uses computational techniques to assess bruising, outperforms traditional and digital met hods across specific bruise color groups. Determining bruise age remains unreliable with current methods, but AI-supported programs offer higher prediction rates in some color groups. These results suggest AI may improve accuracy as technology advances.