Research map / Detection and segmentation / CNN features
"FARSE-CNN: Fully Asynchronous, Recurrent and Sparse Event-Based CNN"
Riccardo Santambrogio, Marco Cannici, Matteo Matteucci
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add state-initialization training strategy (ablation) | FARSE-CNN · DVS128 Gesture | accuracy: 80.7% -> 96.6% | Improved |
| increase Temporal Dropout window from 2 to 4 (ablation) | FARSE-CNN · N-Cars | accuracy / MFLOP per event: 0.949 -> 0.939 accuracy; 2.056 -> 0.495 MFLOP/ev | Mixed |
| increase Temporal Dropout window from 4 to 8 (ablation) | FARSE-CNN · N-Caltech101 | accuracy / MFLOP per event: 0.687 -> 0.660 accuracy; 0.339 -> 0.148 MFLOP/ev | Mixed |
| add Temporal Dropout (ablation) | FARSE-CNN · N-Cars | FLOPS/ev: complexity grows by 1 order of magnitude without it | Improved |
| use Submanifold FARSE-Conv instead of regular FARSE-Conv (ablation) | FARSE-CNN · N-Cars | FLOPS/ev: regular FARSE-Conv grows complexity by 2 orders of magnitude | Improved |
| add non-maximum suppression (ablation) | FARSE-CNN · Gen1 Automotive | mAP: 0.288 -> 0.300 | Improved |
| use deeper network (13 layers instead of 8) (ablation) | FARSE-CNN detection · Gen1 Automotive | mAP | No clear effect |
Later papers that compared with this method
1 later accepted paper reports FARSE-CNN as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Maximizing Asynchronicity in Event-based Neural Networks (ICLR 2026 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Event-based Asynchronous Sparse Convolutional Networks (ECCV 2020 · similarity 0.92)
- EventNet: Asynchronous Recursive Event Processing (CVPR 2019 · similarity 0.91)
- End-to-End Learning of Representations for Asynchronous Event-Based Data (ICCV 2019 · similarity 0.90)
- Event2Vec: Processing neuromorphic events directly by representations in vector space (ICML 2026 · similarity 0.89)
- Leveraging Asynchronous Spiking Neural Networks for Ultra Efficient Event-Based Visual Processing (AAAI 2025 · similarity 0.89)
- Efficient Event-Based Object Detection: A Hybrid Neural Network with Spatial and Temporal Attention (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
