Graph-Based Object Classification for Neuromorphic Vision Sensing
This work addresses the problem of object classification for NVS devices, which cannot use standard CNNs, by introducing a novel graph-based approach and providing a new real-world dataset, representing an incremental advancement in the field.
The authors tackled object classification with neuromorphic vision sensing (NVS) data by proposing a compact graph representation and residual graph CNN architectures, achieving improved spatial and temporal coherence with less computation and memory compared to conventional methods.
Neuromorphic vision sensing (NVS)\ devices represent visual information as sequences of asynchronous discrete events (a.k.a., ``spikes'') in response to changes in scene reflectance. Unlike conventional active pixel sensing (APS), NVS allows for significantly higher event sampling rates at substantially increased energy efficiency and robustness to illumination changes. However, object classification with NVS streams cannot leverage on state-of-the-art convolutional neural networks (CNNs), since NVS does not produce frame representations. To circumvent this mismatch between sensing and processing with CNNs, we propose a compact graph representation for NVS. We couple this with novel residual graph CNN architectures and show that, when trained on spatio-temporal NVS data for object classification, such residual graph CNNs preserve the spatial and temporal coherence of spike events, while requiring less computation and memory. Finally, to address the absence of large real-world NVS datasets for complex recognition tasks, we present and make available a 100k dataset of NVS recordings of the American sign language letters, acquired with an iniLabs DAVIS240c device under real-world conditions.