Anders Larsen

Papers

Total Citations

H-Index

About

Anders Larsen is a computer vision researcher whose work centers on local image descriptors and feature extraction methods. His most notable contribution, the 2012 paper "Jet-Based Local Image Descriptors," introduced a novel approach to representing image patches by leveraging differential invariants—or "jets"—to capture rich geometric and photometric information. This method improved robustness to common image transformations, offering a principled alternative to handcrafted descriptors like SIFT. Although the paper has accumulated 19 citations, its influence extends through subsequent work in texture recognition and 3D reconstruction, where jet-based representations have been adopted for their theoretical elegance and practical stability. Larsen’s research bridges classical differential geometry and modern machine learning, providing foundational tools for tasks requiring precise local feature matching. His work is particularly valuable for students and researchers exploring the intersection of mathematical modeling and computer vision, demonstrating how carefully designed descriptors can outperform purely data-driven approaches in constrained settings.