Code
|
Forward projection toolbox MATLAB toolbox for projecting 3D points in calibrated non-central catadioptric cameras. Supports all conic mirrors with on-axis or off-axis camera placement, enabling calibration and 3D reconstruction through reprojection-error minimization and bundle adjustment. |
|
Extrinsic Camera Calibration Without a Direct View Using a Spherical Mirror (ICCV 2013) Code for estimating a camera’s position and orientation relative to a reference 3D object that it cannot view directly. Uses a single reflection in a spherical mirror of known radius to recover both the camera pose and mirror location. Avoids the degenerate mirror configurations encountered with planar mirrors and does not require the sphere’s boundary to be visible. |
|
A Theory of Multi-Layer Flat Refractive Geometry (CVPR 2012) MATLAB code for calibrating cameras that view scenes through parallel, flat refractive layers, such as a water tank. Estimates layer orientation, thickness, refractive indices, and calibration-target pose from a single image. Includes analytical forward projection for one or two refractive layers, supporting reprojection-error minimization and 3D reconstruction with bundle adjustment. |
|
Analytical Projection Model for Non-Central Catadioptric Cameras with Quadric Mirrors (CVPR 2011) Code accompanying the CVPR 2011 paper on exact analytical projection for cameras viewing scenes through quadric mirrors. Supports unrestricted, including off-axis, camera placement and computes the image projection of a 3D point by solving an eighth-degree equation. Enables 3D reconstruction using bundle adjustment and plane sweep without approximating the camera as a central system. |
|
Surface Reconstruction from Gradient Fields (ECCV 2006) MATLAB code for reconstructing surfaces from noisy or inconsistent gradient fields. Implements a framework that includes Poisson integration and extends it with alpha-surfaces, robust estimation, and anisotropic diffusion to preserve edges and surface details while reducing the effects of noise and outliers. Useful for photometric stereo, shape from shading, and gradient-domain image processing. |
|
Edge Suppression by Gradient Field Transformation Using Cross-Projection Tensors (CVPR 2006) MATLAB code and example images for selectively suppressing edges using a reference image, even under different illumination. Computes cross-projection tensors to transform an image’s gradient field, then reconstructs the result by integration. Applications include foreground extraction, shadow removal with flash/no-flash image pairs, glass-reflection removal, and intrinsic-image recovery. |
|
Solver for the 2D Poisson Equation C++ code for solving the two-dimensional Poisson equation. |