To reconstruct a digital 3D model of a stolen/damaged statue using photogrammetric methods. Besides using standard structure-from-motion and multi-view stereo methods, we match image pairs with low overlap using sliding windows and maximize the normalized cross-correlation based on patch-consistency to well align image pairs into a complete model to build the 3D mesh surface.
Limited number of available photos for a stolen statue; the object's appearance changes because of the dirt and the variation of lighting conditions in the photos taken at different time; the camera setup is unknown, without precise information about intrinsic or extrinsic parameters.
Centre for Perceptual and Interactive Intelligence was established by The Chinese University of Hong Kong in 2020. Our Centre Director Professor Helen Meng and Principal Investigators have rich backgrounds in computer vision, multilingual speech and language technologies, natural language processing, and AI-enabled design automation.