Learning Dense and Continuous Optical Flow From an Event Camera.

Area of Science:

• Computer Vision
• Robotics
• Machine Learning

Background:

• Event cameras offer high temporal resolution for motion capture.
• Existing optical flow methods struggle with dense, continuous estimation, especially in low-event regions.

Purpose of the Study:

• To develop a novel deep learning framework for dense and continuous optical flow estimation.
• To improve the perception of high-speed motion using event streams and intensity images.

Main Methods:

• An event-image fusion and correlation module to integrate data from both modalities.
• An iterative update network with bidirectional training for robust optical flow prediction.

Main Results:

• The proposed model achieves reliable dense flow estimation comparable to frame-based methods.
• It provides temporal continuous flow estimation, a key advantage of event-based methods.
• Experimental results on synthetic and real datasets show superior performance over state-of-the-art methods.

Conclusions:

• The framework effectively combines event and image data for accurate dense and continuous optical flow.
• This approach enhances the capability to perceive and analyze high-speed motion in dynamic scenes.