Reducing the Sim-to-Real Gap for Event Cameras
This work addresses the challenge of training data quality for event camera applications, which is incremental but important for improving performance in high-speed, high dynamic range scenarios.
The paper tackles the problem of reducing the sim-to-real gap for event cameras by improving training data for event-based CNNs, resulting in a 20-40% performance boost for video reconstruction networks and up to 15% for optic flow networks. It introduces a new High Quality Frames (HQF) dataset to address the lack of quality ground truth in existing datasets.
Event cameras are paradigm-shifting novel sensors that report asynchronous, per-pixel brightness changes called 'events' with unparalleled low latency. This makes them ideal for high speed, high dynamic range scenes where conventional cameras would fail. Recent work has demonstrated impressive results using Convolutional Neural Networks (CNNs) for video reconstruction and optic flow with events. We present strategies for improving training data for event based CNNs that result in 20-40% boost in performance of existing state-of-the-art (SOTA) video reconstruction networks retrained with our method, and up to 15% for optic flow networks. A challenge in evaluating event based video reconstruction is lack of quality ground truth images in existing datasets. To address this, we present a new High Quality Frames (HQF) dataset, containing events and ground truth frames from a DAVIS240C that are well-exposed and minimally motion-blurred. We evaluate our method on HQF + several existing major event camera datasets.