Real-Time Intensity-Image Reconstruction for Event Cameras Using Manifold Regularisation
This work addresses the challenge of converting event camera data into usable intensity images for applications like robotics or computer vision, representing an incremental improvement by leveraging manifold regularization for real-time performance.
The authors tackled the problem of reconstructing intensity images from event cameras, which capture sparse per-pixel intensity changes at high frequency but lack traditional intensity information, by proposing a variational model that incorporates an event manifold based on relative timestamps, enabling real-time reconstruction with arbitrary frame rates and producing high-quality images without optical flow estimation.
Event cameras or neuromorphic cameras mimic the human perception system as they measure the per-pixel intensity change rather than the actual intensity level. In contrast to traditional cameras, such cameras capture new information about the scene at MHz frequency in the form of sparse events. The high temporal resolution comes at the cost of losing the familiar per-pixel intensity information. In this work we propose a variational model that accurately models the behaviour of event cameras, enabling reconstruction of intensity images with arbitrary frame rate in real-time. Our method is formulated on a per-event-basis, where we explicitly incorporate information about the asynchronous nature of events via an event manifold induced by the relative timestamps of events. In our experiments we verify that solving the variational model on the manifold produces high-quality images without explicitly estimating optical flow.