Event-Based Motion Magnification

ECCV 2024

Yutian Chen1,2*, Shi Guo1*,†, Fangzheng Yu1,2, Feng Zhang1, Jinwei Gu3, Tianfan Xue3
1Shanghai AI Laboratory, 2Zhejiang University, 3The Chinese University of Hong Kong
*indicates equal contributions, indicates corresponding author

Our system enables a broad and cost-effective amplification of high-frequency motions.


Detecting and magnifying imperceptible high-frequency motions in real-world scenarios has substantial implications for industrial and medical applications. These motions are characterized by small amplitudes and high frequencies. Traditional motion magnification methods rely on costly high-speed cameras or active light sources, which limit the scope of their applications. In this work, we propose a dual-camera system consisting of an event camera and a conventional RGB camera for video motion magnification, containing temporally-dense information from the event stream and spatially-dense data from the RGB images. This innovative combination enables a broad and cost-effective amplification of high-frequency motions. By revisiting the physical camera model, we observe that estimating motion direction and magnitude necessitates the integration of event streams with additional image features. On this basis, we propose a novel deep network for event-based video motion magnification that addresses two primary challenges: firstly, the high frequency of motion induces a large number of interpolated frames (up to 80), which our network mitigates with a Second-order Recurrent Propagation module for better handling of long-term frame interpolations; and secondly, magnifying subtle motions is sensitive to noise, which we address by utilizing a temporal filter to amplify motion at specific frequencies and reduce noise impact. We demonstrate the effectiveness and accuracy of our dual-camera system and network through extensive experiments in magnifying small-amplitude, high-frequency motions, offering a cost-effective and flexible solution for motion detection and magnification.



      title={Event-Based Motion Magnification}, 
      author={Yutian Chen and Shi Guo and Fangzheng Yu and Feng Zhang and Jinwei Gu and Tianfan Xue},
      journal={arXiv preprint arXiv:2402.11957},