As part of joint research on social digital twins, Fujitsu and Carnegie Mellon University in the U.S. are using AI to convert objects seen in images obtained from a fixed monocular camera into three-dimensional images and digitizing them to create images of people and objects. They have developed a technology that dynamically reconstructs three-dimensional shapes and positions with high precision. Both companies are conducting demonstration experiments to verify the effectiveness of the system using image data taken of intersections around Carnegie Mellon University. This technology uses AI that has learned the shapes of people and objects through deep learning to estimate the 3D shape of each 2D object that is captured by the camera, and includes 3D shape estimation technology that can estimate the 3D shape of each 2D object reflected in the camera, as well as buildings, terrain. It is composed of two core technologies: 3D projection technology that accurately estimates and reconstructs the position of people and objects on 3D models. For example, images taken in scenes where people and cars are crowded, such as at intersections, are anonymized and dynamically restored into 3D to identify potential issues such as the causes of traffic accidents that could not be captured by surveillance cameras. Visualize. Fujitsu aims to put it into practical use in fiscal 2025.