Self-supervised Learning to Bring Dual Reversed Rolling Shutter Images Alive

Modern consumer cameras usually employ the rolling shutter (RS) mechanism, where images are captured by scanning scenes row-by-row, yielding RS distortions for dynamic scenes. To correct RS distortions, existing methods adopt a fully supervised learning manner, where high framerate global shutter (GS) images should be collected as ground-truth supervision. In this paper, we propose a Self-supervised learning framework for Dual reversed RS distortions Correction (SelfDRSC), where a DRSC network can be learned to generate a high framerate GS video only based on dual RS images with reversed distortions. In particular, a bidirectional distortion warping module is proposed for reconstructing dual reversed RS images, and then a self-supervised loss can be deployed to train DRSC network by enhancing the cycle consistency between input and reconstructed dual reversed RS images. Besides start and end RS scanning time, GS images at arbitrary intermediate scanning time can also be supervised in SelfDRSC, thus enabling the learned DRSC network to generate a high framerate GS video. Moreover, a simple yet effective self-distillation strategy is introduced in self-supervised loss for mitigating boundary artifacts in generated GS images. On synthetic dataset, SelfDRSC achieves better or comparable quantitative metrics in comparison to state-of-the-art methods trained in the full supervision manner. On real-world RS cases, our SelfDRSC can produce high framerate GS videos with finer correction textures and better temporary consistency. The source code and trained models are made publicly available at https://github.com/shangwei5/SelfDRSC.

1. Introduction

Recent years have witnessed an increasing demand for imaging sensors, due to the widespread applications of digital cameras and smartphones. Although the Charge-Coupled Device (CCD) has been the dominant technology for imaging sensors, it is recently popular that modern consumer cameras choose the Complementary Metal-Oxide Semiconductor (CMOS) as an alternative due to its many merits, e.g., easy integration with image processing pipeline and communication circuits, and low power consumption [14]. In CMOS sensors, rolling shutter (RS) scanning mechanism is generally deployed to capture images, i.e., each row of CMOS array is exposed in the sequential time, which is different from CCD with global shutter (GS) scanning at one instant. Therefore, RS images suffer from distortions when capturing dynamic scenes, which not only affect human visual perception but also yield performance degradation or even failure in computer vision tasks [2], [13].