Learning to Deblur and Generate High Frame Rate Video with an Event Camera

TL;DR

This paper introduces a novel method using event cameras and deep learning to deblur images and generate high frame rate videos, outperforming existing techniques in sharpness and temporal resolution.

Contribution

The paper proposes a residual learning framework with specialized neural networks for deblurring and high frame rate video synthesis guided by event camera data.

Findings

Restores sharper images than state-of-the-art methods.

Generates high frame rate videos with improved temporal detail.

Effective use of event data for deblurring and video synthesis.

Abstract

Event cameras are bio-inspired cameras which can measure the change of intensity asynchronously with high temporal resolution. One of the event cameras' advantages is that they do not suffer from motion blur when recording high-speed scenes. In this paper, we formulate the deblurring task on traditional cameras directed by events to be a residual learning one, and we propose corresponding network architectures for effective learning of deblurring and high frame rate video generation tasks. We first train a modified U-Net network to restore a sharp image from a blurry image using corresponding events. Then we train another similar network with different downsampling blocks to generate high frame rate video using the restored sharp image and events. Experiment results show that our method can restore sharper images and videos than state-of-the-art methods.