Ma Minhua, Zheng Huiru, Lallie Harjinder
School of Computing, University of Derby, Derby, UK.
J Forensic Sci. 2010 Sep;55(5):1227-31. doi: 10.1111/j.1556-4029.2010.01453.x. Epub 2010 Jun 8.
Computer-generated three-dimensional (3D) animation is an ideal media to accurately visualize crime or accident scenes to the viewers and in the courtrooms. Based upon factual data, forensic animations can reproduce the scene and demonstrate the activity at various points in time. The use of computer animation techniques to reconstruct crime scenes is beginning to replace the traditional illustrations, photographs, and verbal descriptions, and is becoming popular in today's forensics. This article integrates work in the areas of 3D graphics, computer vision, motion tracking, natural language processing, and forensic computing, to investigate the state-of-the-art in forensic visualization. It identifies and reviews areas where new applications of 3D digital technologies and artificial intelligence could be used to enhance particular phases of forensic visualization to create 3D models and animations automatically and quickly. Having discussed the relationships between major crime types and level-of-detail in corresponding forensic animations, we recognized that high level-of-detail animation involving human characters, which is appropriate for many major crime types but has had limited use in courtrooms, could be useful for crime investigation.
计算机生成的三维（3D）动画是一种理想的媒介，能够向观众和法庭准确呈现犯罪或事故现场。基于事实数据，法医动画可以重现现场并展示不同时间点的活动。利用计算机动画技术重建犯罪现场正开始取代传统的插图、照片和文字描述，在当今法医学领域越来越受欢迎。本文整合了三维图形、计算机视觉、运动跟踪、自然语言处理和法医计算等领域的工作，以研究法医可视化的最新技术。它识别并回顾了三维数字技术和人工智能的新应用可用于增强法医可视化特定阶段，从而自动快速创建三维模型和动画的领域。在讨论了主要犯罪类型与相应法医动画中的细节程度之间的关系后，我们认识到涉及人物角色的高细节程度动画，虽然适用于许多主要犯罪类型，但在法庭上的使用有限，可能对犯罪调查有用。
