Object Detection on Gen1 (test)
50.4mAPERGO-12
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ERGO-12Time (ms)=77.2, Size (M)=59.6, Backbone=Swin Transformer V22023.11 | 50.4 | — | — | — | — | — | |
| ERGO-12Backbone=Transformer, Detection Head=YOLOv6, Runtime (ms)=69.9, Parameters (M)=59.62024.02 | 50.4 | — | — | — | — | — | |
| LEOD-RVT-STime (ms)=9.5, Size (M)=9.92023.11 | 48.7 | — | — | — | — | — | |
| GET-TBackbone=Transformer + RNN, Detection Head=YOLOX, Runtime (ms)=16.8, Parameters (M)=21.92024.02 | 47.9 | — | — | — | — | — | |
| S5-ViT-BBackbone=Transformer + SSM, Detection Head=YOLOX, Runtime (ms)=8.16, Parameters (M)=18.22024.02 | 47.7 | — | — | — | — | — | |
| RVT-BBackbone=Transformer + RNN, Detection Head=YOLOX [15], Time (ms)=10.2 (3.7), Params (M)=18.52022.12 | 47.2 | — | — | — | — | — | |
| RVT-BTime (ms)=10.2, Size (M)=18.52023.11 | 47.2 | — | — | — | — | — | |
| RVT-BBackbone=Transformer + RNN, Detection Head=YOLOX, Runtime (ms)=10.2, Parameters (M)=18.52024.02 | 47.2 | — | — | — | — | — | |
| HMNet-L3Time (ms)=7.9 (V100 GPU), Size (M)=33.22023.11 | 47.1 | — | — | — | — | — | |
| ASTMNetBackbone=(T)CNN + RNN, Detection Head=SSD, Time (ms)=35.6, Params (M)=> 100*2022.12 | 46.7 | — | — | — | — | — | |
| ASTMNetTime (ms)=35.6, Size (M)=> 1002023.11 | 46.7 | — | — | — | — | — | |
| S5-ViT-SBackbone=Transformer + SSM, Detection Head=YOLOX, Runtime (ms)=7.81, Parameters (M)=9.72024.02 | 46.6 | — | — | — | — | — | |
| RVT-SBackbone=Transformer + RNN, Detection Head=YOLOX, Time (ms)=9.5 (3.0), Params (M)=9.92022.12 | 46.5 | — | — | — | — | — | |
| RVT-STime (ms)=9.5, Size (M)=9.92023.11 | 46.5 | — | — | — | — | — | |
| Nested-TBackbone=Transformer + RNN, Detection Head=YOLOX, Runtime (ms)=25.9, Parameters (M)=22.22024.02 | 46.3 | — | — | — | — | — | |
| S4D-ViT-BBackbone=Transformer + SSM, Detection Head=YOLOX, Runtime (ms)=9.40, Parameters (M)=16.52024.02 | 46.2 | — | — | — | — | — | |
| Swin-T v2Backbone=Transformer + RNN, Detection Head=YOLOX, Runtime (ms)=26.6, Parameters (M)=21.12024.02 | 45.5 | — | — | — | — | — | |
| RVT-TBackbone=Transformer + RNN, Detection Head=YOLOX, Time (ms)=9.4 (2.3), Params (M)=4.42022.12 | 44.1 | — | — | — | — | — | |
| REDBackbone=CNN + RNN, Detection Head=SSD, Time (ms)=16.7, Params (M)=24.12022.12 | 40 | — | — | — | — | — | |
| REDTime (ms)=16.7, Size (M)=24.12023.11 | 40 | — | — | — | — | — | |
| REDBackbone=CNN + RNN, Detection Head=SSD, Runtime (ms)=16.7, Parameters (M)=24.12024.02 | 40 | — | — | — | — | — | |
| YOLOv3 EventsBackbone=CNN, Detection Head=YOLOV3, Time (ms)=22.3, Params (M)=> 60*2022.12 | 31.2 | — | — | — | — | — | |
| YOLOv3 EventsBackbone=CNN, Detection Head=YOLOv3, Runtime (ms)=22.3, Parameters (M)=> 60*2024.02 | 31.2 | — | — | — | — | — | |
| MatrixLSTMBackbone=RNN + CNN, Detection Head=YOLOV3, Params (M)=61.52022.12 | 31 | — | — | — | — | — | |
| MatrixLSTMBackbone=RNN + CNN, Detection Head=YOLOv3, Parameters (M)=61.52024.02 | 31 | — | — | — | — | — | |
| RRC-EventsBackbone=CNN, Detection Head=YOLOv3 [41], Time (ms)=21.5, Params (M)=> 100*2022.12 | 30.7 | — | — | — | — | — | |
| RRC-EventsBackbone=CNN, Detection Head=YOLOv3, Runtime (ms)=21.5, Parameters (M)=> 100*2024.02 | 30.7 | — | — | — | — | — | |
| Inception + SSDBackbone=CNN, Detection Head=SSD, Time (ms)=19.4, Params (M)=> 60*2022.12 | 30.1 | — | — | — | — | — | |
| Inception + SSDBackbone=CNN, Detection Head=SSD, Runtime (ms)=19.4, Parameters (M)=> 60*2024.02 | 30.1 | — | — | — | — | — | |
| Spiking DenseNetBackbone=SNN, Detection Head=SSD [30], Params (M)=8.22022.12 | 18.9 | — | — | — | — | — | |
| Spiking DenseNetBackbone=SNN, Detection Head=SSD, Parameters (M)=8.22024.02 | 18.9 | — | — | — | — | — | |
| AEGNNBackbone=GNN, Detection Head=YOLOv1, Params (M)=20.02022.12 | 16.3 | — | — | — | — | — | |
| AEGNNBackbone=GNN, Detection Head=YOLOv1, Parameters (M)=20.02024.02 | 16.3 | — | — | — | — | — | |
| AsynetBackbone=Sparse CNN, Detection Head=YOLOv1, Params (M)=11.42022.12 | 14.5 | — | — | — | — | — | |
| AsynetBackbone=Sparse CNN, Detection Head=YOLOv1, Parameters (M)=11.42024.02 | 14.5 | — | — | — | — | — | |
| NVS-SBackbone=GNN, Detection Head=YOLOv1 [40], Params (M)=0.92022.12 | 8.6 | — | — | — | — | — | |
| EvRT-DETR-B2026.06 | — | 52.7 | 50.6 | 45 | 42.2 | 32 | |
| FATE-E (PE + FAT + EvRT-DETR-B)Components=PE + FAT, Base Model=EvRT-DETR-B2026.06 | — | 53.14 | 51.6 | 48.4 | 46.8 | 42.38 | |
| FATE-S (PE + FAT + S5-ViT-B)Components=PE + FAT, Base Model=S5-ViT-B2026.06 | — | 48.2 | 46.9 | 47.02 | 43.18 | 40.8 | |
| GET2026.06 | — | 47.9 | 34.15 | 19.97 | 15.13 | 5.35 | |
| RVT-B2026.06 | — | 47.2 | 35.13 | 21.98 | 18.61 | 8.35 | |
| S5-ViT-B2026.06 | — | 47.4 | 46.44 | 45.08 | 42.49 | 39.84 |