NOCaL: Calibration-Free Semi-Supervised Learning of Odometry and Camera Intrinsics

There are a multitude of emerging imaging tech- nologies that could benefit robotics. However the need for bespoke models, calibration and low-level processing represents a key barrier to their adoption. In this work we present NOCaL, Neural Odometry and Calibration using Light fields, a semi-supervised learning architecture capable of interpret- ing previously unseen cameras without calibration. NOCaL learns to estimate camera parameters, relative pose, and scene appearance. It employs a scene-rendering hypernetwork pre- trained on a large number of existing cameras and scenes, and adapts to previously unseen cameras using a small supervised training set to enforce metric scale. We demonstrate NOCaL on rendered and captured imagery using conventional cam- eras, demonstrating calibration-free odometry and novel view synthesis. This work represents a key step toward automating the interpretation of general camera geometries and emerg- ing imaging technologies. Code and datasets are available at https://roboticimaging.org/Projects/NOCaL/.