A General Approach to State Refinement

Publication Type:
Conference Proceeding
2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021, 00, pp. 8985-8991
Issue Date:
Filename Description Size
2104.10729.pdf5.42 MB
Adobe PDF
Full metadata record
Deep learning algorithms such as Convolutional Neural Networks (CNNs) are currently used to solve a range of robotics and computer vision problems. These networks typically estimate the desired representation in a single forward pass and must therefore learn to converge from a wide range of initial conditions to a precise result. This is challenging, and has led to increased interest in the development of separate refinement modules which learn to improve a given initial estimate, thus reducing the required search space. Such modules are usually developed ad-hoc for each given application, often requiring significant engineering investment. In this work we propose a generic innovation-based CNN. Our CNN is implemented along with a stochastic gradient descent (SGD) algorithm to iteratively refine a given initial estimate. The proposed approach provides a general framework for the development of refinement modules applicable to a wide range of robotics problems. We apply this framework to object pose estimation and depth estimation and demonstrate significant improvement over the initial estimates, in the range of 4.2 - 8.1%, for both applications.
Please use this identifier to cite or link to this item: