Deep Feedback Inverse Problem Solver

TL;DR

This paper introduces a generic neural feedback-based iterative method for solving inverse problems that is faster and more accurate than traditional and existing deep learning approaches.

Contribution

It proposes a novel feedback-driven neural network framework that does not require prior knowledge and can recover from early errors across various inverse problems.

Findings

Achieves comparable or better accuracy than traditional methods.

Operates two to three orders of magnitude faster.

Consistently outperforms existing deep learning approaches.

Abstract

We present an efficient, effective, and generic approach towards solving inverse problems. The key idea is to leverage the feedback signal provided by the forward process and learn an iterative update model. Specifically, at each iteration, the neural network takes the feedback as input and outputs an update on the current estimation. Our approach does not have any restrictions on the forward process; it does not require any prior knowledge either. Through the feedback information, our model not only can produce accurate estimations that are coherent to the input observation but also is capable of recovering from early incorrect predictions. We verify the performance of our approach over a wide range of inverse problems, including 6-DOF pose estimation, illumination estimation, as well as inverse kinematics. Comparing to traditional optimization-based methods, we can achieve comparable or better performance while being two to three orders of magnitude faster. Compared to deep learning-based approaches, our model consistently improves the performance on all metrics. Please refer to the project page for videos, animations, supplementary materials, etc.