SCSim: A Realistic Spike Cameras Simulator

Liwen Hu,Lei Ma,Yijia Guo,Tiejun Huang
DOI: https://doi.org/10.1109/icme57554.2024.10688014
2024-01-01
Abstract:Spike cameras, with their exceptional temporal resolution, arerevolutionizing high-speed visual applications. Large-scale synthetic datasetshave significantly accelerated the development of these cameras, particularlyin reconstruction and optical flow. However, current synthetic datasets forspike cameras lack sophistication. Addressing this gap, we introduce SCSim, anovel and more realistic spike camera simulator with a comprehensive noisemodel. SCSim is adept at autonomously generating driving scenarios andsynthesizing corresponding spike streams. To enhance the fidelity of thesestreams, we've developed a comprehensive noise model tailored to the uniquecircuitry of spike cameras. Our evaluations demonstrate that SCSim outperformsexisting simulation methods in generating authentic spike streams. Crucially,SCSim simplifies the creation of datasets, thereby greatly advancingspike-based visual tasks like reconstruction. Our project refers tohttps://github.com/Acnext/SCSim.
What problem does this paper attempt to address?