Introduction and Motivation - 3D reconstruction determines the 3D profile of objects and the 3D coordinates of points on the profile. - It is a core technology in various fields such as Computer Aided Geometric Design (CAGD), computer graphics, computer vision, medical imaging, and virtual reality. - It offers a new and accurate approach in diagnosis, particularly in presenting lesion information in 3D for clinical value. - Digital elevation models can be reconstructed using methods like airborne laser altimetry and synthetic aperture radar.

Active and Passive Methods Active methods: - Use range data to reconstruct the 3D profile of objects by numerical approximation. - Interfere with the object using techniques like structured light, laser range finders, and time-of-flight lasers. - Acquire depth maps by actively measuring distance or emitted and reflected radiance from the object. - Applicable in various scenarios such as computer stereo vision and 3D ultrasound. - Enable accurate reconstruction but may require complex setups and computations.

Passive methods: - Do not interfere with the object and rely on sensors to measure the radiance reflected or emitted by the object's surface. - Use image sensors in cameras sensitive to visible light and process digital images or videos. - Can be applied in a wider range of situations compared to active methods. - Used for image-based reconstruction and output 3D models. - Commonly used in computer vision and computer graphics.

Monocular Cues and Stereo Vision Monocular cues methods: - Use one or more images from a single viewpoint to reconstruct 3D shapes. - Measure 3D shape using 2D characteristics such as silhouettes, shading, and texture. - Techniques include shape-from-shading, photometric stereo, and shape-from-texture.

Stereo vision: - Obtains 3D geometric information of an object from multiple images. - Uses two cameras simultaneously or a single camera at different times to restore the object's 3D profile and location. - Requires careful calibration and matching of image points to calculate depth information from disparity.

Techniques and Algorithms for 3D Reconstruction - Various techniques and algorithms are used for 3D reconstruction, including real-time non-rigid reconstruction, 3D building model reconstruction, investigating landslides, estimating pavement roughness, and shape-from-X. - Algorithms for 3D reconstruction include Delaunay triangulation, marching cubes, shape from shading, and photometric methods. - Challenges include loss of geometry precision and complexity of post-processing techniques.

Applications of 3D Reconstruction - 3D reconstruction has various applications, including 360 depth estimation for virtual reality, car shape reconstruction, translation with conditional vector-quantised code diffusion, learning cultural heritage, and digital archaeological exhibition. - It is also used in computer vision, robotics, medical imaging, and surface reconstruction. - Applications in medical imaging facilitate clinical routine, surgical planning, patient follow-up, and various clinical areas such as radiotherapy planning and neurointerventions.