List your company

Snitch Localization on CATER (test)

79.7Top-1 AccuracyTFC V3D Depthwise
-0.27620.48741.2562.013Mar 19, 2021Oct 2, 2021Apr 18, 2022Nov 2, 2022May 19, 2023Dec 3, 2023Jun 18, 2024

Evaluation results

MethodLinks
TFC V3D DepthwiseBackbone=ResNet50, total frames=1 × 32 × 1, parameters=24.64M, GFLOPs=1322022.0379.795.50.47
TFC V3DBackbone=ResNet50, total frames=1 × 32 × 1, parameters=32.68M, GFLOPs=1682022.0379.5940.5
V3DBackbone=ResNet50, total frames=1 × 32 × 1, parameters=31.63M, GFLOPs=167.72022.0377.9940.54
OPNetBackbone=OPNet, total frames=all frames, extra annotation/data=LA-CATER(Shamsian et al., 2020)2022.0374.8—0.54
Hopper-multihopFPS=1, # Frames=132021.0373.293.80.0085
HopperBackbone=Transformer, total frames=all frames, extra annotation/data=LA-CATER(Shamsian et al., 2020)2022.0373.293.80.85
OC-SlotSSMPre-train=Enabled2024.0669.2790.48—
Hopper-sinetFPS=1, # Frames=132021.0369.191.80.0102
TPN-101FPS=2.5, # Frames=322021.0365.3830.0109
Hopper-transformerFPS=1, # Frames=132021.0364.990.10.0111
TSM-50FPS=1, # Frames=132021.036485.70.0093
OC-SlotSSMPre-train=None2024.0661.5884—
Hopper-transformer (last frame)FPS=1, # Frames=132021.0361.186.60.0142
I3D-50 + LSTMFPS=5, # Frames=642021.0360.281.80.012
R3D+LSTMBackbone=ResNet50, total frames=64 × 10 × 12022.0360.281.81.2
I3D-50FPS=5, # Frames=642021.0357.478.40.014
R3DBackbone=ResNet50, total frames=64 × 10 × 32022.0357.478.41.4
SlotSSMPre-train=Enabled2024.0654.7374.42—
SlotTransformerPre-train=Enabled2024.0649.2170.24—
V4DBackbone=ResNet50, total frames=8 × 10 × 3, parameters=36.42M, GFLOPs=9342022.0347.272.1—
I3D-50 + NL + LSTMFPS=2.5, # Frames=322021.0346.269.90.015
R3D+NL+LSTMBackbone=ResNet50, total frames=32 × 10 × 12022.0346.269.91.5
Hungarian (Tracking - ours)2021.034652.70.019
Single State SSMPre-train=Enabled2024.0641.1565.7—
SlotTransformerPre-train=None2024.0641.0962.24—
DaSiamRPN (Tracking)2021.0333.940.80.024
TrackingBackbone=SiamRPN, total frames=all frames2022.0333.9—2.4
I3D-50 + NLFPS=2.5, # Frames=322021.0326.768.90.026
R3D+NLBackbone=ResNet50, total frames=32 × 10 × 32022.0326.768.92.6
SlotSSMPre-train=None2024.0625.6445.03—
TSN (RGB) + LSTMFPS=3, # Frames=32021.0325.667.20.026
SINetFPS=1, # Frames=132021.0321.147.10.0314
TSN (RGB)FPS=3, # Frames=32021.0314.138.50.032
TSN (Flow) + LSTMFPS=3, # Frames=32021.031443.50.032
TransformerFPS=1, # Frames=132021.0313.739.90.0353
Single State SSMPre-train=None2024.0610.2727.21—
TSN (Flow)FPS=3, # Frames=32021.039.632.30.037
Random2021.032.813.80.039