Max Robotics — Global Service Robot Directory + US Certification

About

Daniel DeTone is a leading researcher at the intersection of computer vision, robotics, and differentiable optimization. His work focuses on enabling machines to perceive and understand 3D environments from visual data, with major contributions in fiducial marker pose estimation, visual localization, and end-to-end learning. DeTone is best known for developing **Deep ChArUco**, a deep learning approach that robustly detects ChArUco calibration boards even in low-light conditions where classical methods fail—a critical advance for AR and robotics (80 citations). He also introduced **OrienterNet**, the first deep neural network capable of localizing a camera in 2D public maps without expensive 3D point clouds, achieving 78 citations for bridging a key gap between human and machine spatial reasoning. Additionally, DeTone created **Theseus**, an open-source PyTorch library for differentiable nonlinear least squares optimization (47 citations), providing a unified framework for end-to-end structured learning in robotics and vision. His system **ODAM** further advances 3D scene understanding by jointly detecting, associating, and mapping objects from posed RGB video. DeTone’s work consistently pushes the boundaries of practical, deployable perception systems.