Evaluation of neural network algorithms for atmospheric turbulence mitigation
This work addresses image quality issues for applications like surveillance or astronomy, but it is incremental as it focuses on reusing existing methods.
The paper evaluated five neural network architectures for mitigating atmospheric turbulence-induced blur in images and videos, finding that an end-to-end trained network eliminated the need for a stabilization step.
A variety of neural networks architectures are being studied to tackle blur in images and videos caused by a non-steady camera and objects being captured. In this paper, we present an overview of these existing networks and perform experiments to remove the blur caused by atmospheric turbulence. Our experiments aim to examine the reusability of existing networks and identify desirable aspects of the architecture in a system that is geared specifically towards atmospheric turbulence mitigation. We compare five different architectures, including a network trained in an end-to-end fashion, thereby removing the need for a stabilization step.