The segmentation accuracy and efficiency of lightweight MOS methods on CDnet 2014 dataset.
| No. | Year | Method | GPU | F-measure Mean (%) ↑ | Inference Speed (fps) ↑ | # Param (M) ↓ | Model Size (MB) ↓ |
|---|---|---|---|---|---|---|---|
| 1 | 2020 | Edge Aggregation Network [157] | Tesla V100 | 96.9 | 19.6 | 20 | 240 |
| 2 | 2021 | Lightweight U-Net-like [111] | Titan Xp | 97.7 | 250 | 0.1 | 0.4 |
| 3 | 2021 | 3DS_MM [75] | 95.2 | 151 | 0.4 | 1.5 | |
| 4 | 2020 | 2D_Separable CNN [74] | 91.5 | 149 | 1 | 3.8 | |
| 5 | 2021 | F3DsCNN [76] | 95.9 | 120 | 4.3 | 5 | |
| 6 | 2022 | ChangeDet [133] | 88.0 | 58.8 | 0.1 | 1.6 | |
| 7 | 2019 | 3DFR [127] | 86.0 | 33.3 | 0.1 | 2.8 | |
| 8 | 2019 | Frame-Level Weakly Supervised [139] | GTX 1080Ti | 84.8 | 134 | 3.7 | 14.8 |
| 9 | 2019 | MvRF-CNN [3] | 95.1 | 42 | 0.9 | 4.0 | |
| 10 | 2018 | Guided Multi-Scale CNN [107] | 75.9 | 28 | 0.3 | 1.3 | |
| 11 | 2019 | 3D CNN-LSTM [2] | 95.7 | 24 | 0.2 | 0.9 | |
| 12 | 2021 | BSUV-Net 2.0 [193] | Tesla P100 | 81.0 | 29 | 15.9 | 110 |
| 13 | 2018 | BScGAN [10] | GTX 1080 | 97.1 | 400 (BMC dataset) | - | - |
| 14 | 2019 | FCESNet [168] | Titan X | 86.0 | 112 | - | - |
| 15 | 2018 | MFCN [240] | GTX 1060 | 98.7 | 27 | 20.8 | 50 |
| 16 | 2018 | FgSegNet_M [96] | GTX 970 | 98.8 | 18 | 6.5 | 60.4 |
| 17 | 2016 | MSCNN+Cascade [216] | 95.0 | 13 | 0.5 | 3.8 | |
| 18 | 2019 | Trip-Net [147] | Titan Black | 84.2 | 282 | 0.3 | 2.5 |
| 19 | 2019 | BMN-BSN [142] | - | 80.0 | 48 | - | - |
| 20 | 2020 | RT-SBS [43] | - | 82.8 | 25 | 0.6 | 3.0 |
| No. | Year | Method | GPU | Inference Speed (fps) ↑ | # Param (M) ↓ | Model Size (MB) ↓ | |
|---|---|---|---|---|---|---|---|
| - | - | ||||||
| - | - | ||||||
| - | - | - | |||||
| - |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.