Convex Optimization
Quasiconvex Optimization for Robust Geometric Reconstruction
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Multiple View Geometry and the L_"-norm
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Smooth Approximation of L_infinity-Norm for Multi-view Geometry
DICTA '09 Proceedings of the 2009 Digital Image Computing: Techniques and Applications
Hi-index | 0.00 |
Recent work on geometric vision problems has exploited convexity properties to obtain globally optimal solutions. The way based on L-infinity norm makes it possible to obtain a provably global optimal solution. But the computation time increases rapidly according to the size of measurement data, so the time cost is unbearable for large scale data. We validate that L-infinity geometry problems is a variational inequality problem essentially and present a hybrid steepest descent method instead of traditional interior point algorithm to compute L-infinity solutions for large scale geometry problem. We give both theoretic justification and experimental verification. Experimental results verify that our method is extremely fast than traditional ones while keeps the accuracy.