GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
This addresses the issue of unreliable optical flow for computer vision applications in adverse conditions, representing a novel integration of sensor data but incremental in its method.
The paper tackles the problem of inaccurate optical flow in challenging scenes like fog, rain, and night by introducing an unsupervised learning approach that fuses gyroscope data with image content, resulting in state-of-the-art performance in both regular and challenging scenes.
Existing optical flow methods are erroneous in challenging scenes, such as fog, rain, and night because the basic optical flow assumptions such as brightness and gradient constancy are broken. To address this problem, we present an unsupervised learning approach that fuses gyroscope into optical flow learning. Specifically, we first convert gyroscope readings into motion fields named gyro field. Second, we design a self-guided fusion module to fuse the background motion extracted from the gyro field with the optical flow and guide the network to focus on motion details. To the best of our knowledge, this is the first deep learning-based framework that fuses gyroscope data and image content for optical flow learning. To validate our method, we propose a new dataset that covers regular and challenging scenes. Experiments show that our method outperforms the state-of-art methods in both regular and challenging scenes. Code and dataset are available at https://github.com/megvii-research/GyroFlow.