List your company

Temporal Action Segmentation on Breakfast (F1@10/25/50, Edit, Acc)

76.4AccuracyDiffAct
48.63255.84163.0570.259Mar 5, 2019May 15, 2020Jul 26, 2021Oct 7, 2022Dec 18, 2023Feb 27, 2025May 11, 2026

Evaluation results

MethodLinks
DiffActParams(M)=1.2, Backbone=I3D2024.0976.480.375.964.678.4—
C2F-TCNArchitecture Type=Complex, Base Backbone Model=C2F-TCN2026.0476.171.4167.857.2867.84—
FACT + OursArchitecture Type=Complex, Base Backbone Model=FACT2026.0476.0881.3676.4965.9879.44—
C2F-TCN + OursArchitecture Type=Complex, Base Backbone Model=C2F-TCN2026.0476.0372.468.5957.6968.54—
Temporal AggregatesBackbone=I3D, Evaluation protocol=GT segments (Upper bound)2020.0675.9—————
FACTArchitecture Type=Complex, Base Backbone Model=FACT2026.0475.8980.3675.7965.4779.23—
DiffActInference mode=offline2024.1175.180.375.975.176.4—
DiffActParams(M)=1.2, Backbone=ViT-B2024.0974.978.774.962.776.4—
OursSupervision Setting=Fully-supervised, Backbone=ASFormer2026.0574.678.873.15877.9—
KARI*Supervision Setting=Fully-supervised, Backbone=ASFormer2026.0574.577.371.65778.3—
Baseline [45]reproduced=false, refinement=false2023.1274.377.37259.475.6—
ASFormerre-produced=true, refinement=false, grammar induction=None2023.1274.377.37259.475.6—
ASFormerre-prod.=true, refinement=false, grammar induction=none2023.1274.377.37259.475.6—
Richard et al.reproduced=true, refinement=true2023.1274.177.272.259.477.3—
KARIreproduced=true, refinement=true2023.127478.873.760.877.8—
ASFormer + KARIre-produced=true, refinement=true, grammar induction=KARI2023.127478.873.760.877.8—
ASFormerre-prod.=true, refinement=true, grammar induction=KARI2023.127478.873.760.877.8—
ASFormerre-produced=false, refinement=false, grammar induction=None2023.1273.57670.657.475—
ASFormerre-prod.=false, refinement=false, grammar induction=none2023.1273.57670.657.475—
ASFormerInference mode=offline2024.1173.57670.657.475—
OriginalBackbone=ASFormer2025.0373.57670.657.475—
ASFormerParams(M)=1.1, Backbone=I3D2024.0973.57670.657.475—
Kuehne et al.reproduced=false, refinement=true2023.1272.97469.155.572.8—
ADIOS-ANDreproduced=false, refinement=true2023.1272.469.864.952.269.2—
ASFormer + ADIOS-ANDre-produced=true, refinement=true, grammar induction=ADIOS-AND2023.1272.469.864.952.269.2—
ASFormerre-prod.=true, refinement=true, grammar induction=ADIOS-AND2023.1272.469.864.952.269.2—
w/o constraintsSupervision Setting=Fully-supervised, Backbone=ASFormer2026.0572.474.168.755.572.8—
Temporal2Seq_batchParams(M)=6.9, Backbone=ViT-B2024.0972.277.571.659.775—
ADIOS-ORreproduced=true, refinement=true2023.1271.871.866.854.270.3—
ASFormer + ADIOS-ORre-produced=true, refinement=true, grammar induction=ADIOS-OR2023.1271.871.866.854.270.3—
ASFormerre-prod.=true, refinement=true, grammar induction=ADIOS-OR2023.1271.871.866.854.270.3—
Temporal2Seq_dataParams(M)=6.9, Backbone=ViT-B2024.0971.87771.560.374.7—
Baseline_TASParams(M)=6.9, Backbone=ViT-B2024.0970.775.770.857.873.4—
BCN2020.1270.468.765.55566.2—
BCNSupervision=Full2021.0370.468.765.55566.2—
SSTDALabeled training data percentage=100%2020.0370.27569.155.273.7—
UVASTParams(M)=1.3, Backbone=I3D2024.0969.776.971.55877.1—
MS-TCNre-produced=true, refinement=false, grammar induction=None2023.1269.470.765.152.669.7—
MS-TCNre-prod.=true, refinement=false, grammar induction=none2023.1269.470.765.152.669.7—
HASRParams(M)=19.2, Backbone=I3D2024.0969.474.769.55771.9—
MS-TCNInference mode=offline2024.1169.364.759.647.567.3—
MS-TCN + KARIre-produced=true, refinement=true, grammar induction=KARI2023.1268.874.668.755.174.9—
MS-TCNre-prod.=true, refinement=true, grammar induction=KARI2023.1268.874.668.755.174.9—
Timestamp Supervision approachFraction=0.12021.0368.4—————
MS-TCN + ADIOS-ANDre-produced=true, refinement=true, grammar induction=ADIOS-AND2023.1268.466.7614868—
MS-TCNre-prod.=true, refinement=true, grammar induction=ADIOS-AND2023.1268.466.7614868—
DTGRM2020.1268.368.761.946.668.9—
MS-TCN + ADIOS-ORre-produced=true, refinement=true, grammar induction=ADIOS-OR2023.1268.269.263.350.369.6—
MS-TCNre-prod.=true, refinement=true, grammar induction=ADIOS-OR2023.1268.269.263.350.369.6—
UARLParams(M)=1.5, Backbone=I3D2024.0967.865.259.447.466.2—
MSTCN++2020.1267.664.158.645.965.6—
MS-TCN++Supervision=Full2021.0367.664.158.645.965.6—
ASRFSupervision=Full2021.0367.674.368.956.172.4—
MS-TCN++2026.0267.664.158.645.965.6—
MS-TCN++Params(M)=0.8, Backbone=I3D2024.0967.664.158.645.965.6—
ASRFParams(M)=1.2, Backbone=I3D2024.0967.674.368.956.172.4—
MS-TCN + OursArchitecture Type=Standard, Base Backbone Model=MS-TCN2026.0467.4655.9651.3940.7862.29—
Timestamp Supervision approachFraction=0.012021.0367.4—————
OriginalBackbone=MS-TCN2025.0367.250.546.336.860.6—
MS-TCNArchitecture Type=Standard, Base Backbone Model=MS-TCN2026.0466.9650.6146.4136.5861.75—
MS-TCNFeatures=I3D2019.0366.352.648.137.961.7—
MS-TCNBackbone=I3D2020.0666.352.648.137.961.7—
MSTCN2020.1266.352.648.137.961.7—
MS-TCNSupervision=Full2021.0366.352.648.137.961.7—
MS-TCNre-produced=false, refinement=false, grammar induction=None2023.1266.352.648.137.961.7—
MS-TCNre-prod.=false, refinement=false, grammar induction=none2023.1266.352.648.137.961.7—
MS-TCN2026.0266.352.648.137.961.7—
SSTDALabeled training data percentage=65%2020.0365.869.362.949.469—
Temporal AggregatesBackbone=I3D, Temporal scale=2s2020.0665.352.346.534.851.3—
MS-TCNFeatures=IDT2019.0365.158.252.940.861.4—
MS-TCNParams(M)=0.8, Backbone=I3D2024.0965.158.252.940.861.4—
MSTCN+GTRM2020.126557.55443.358.7—
MS-TCNsource=official code2020.0364.765.359.647.265.7—
Temporal AggregatesBackbone=I3D, Temporal scale=5s2020.0664.559.253.939.554.5—
w/o constraintsSupervision Setting=Semi-supervised (5% labeled), Backbone=C2F-TCN2026.0564.257.151.234.654.6—
Timestamp Supervision approachSupervision=Timestamps2021.0364.170.563.647.469.9—
OursSupervision Setting=Semi-supervised (5% labeled), Backbone=C2F-TCN2026.0563.958.552.43656.4—
Seg. Model M + plateauSupervision=Timestamps2021.0363.565.559.143.265.9—
Generative Network Inversion†Backbone=ASFormer2025.0361.162.45642.161.4—
HMM-RNNFraction=0.12021.0360.9—————
GRUFeatures=I3D2019.0360.6—————
GRU2020.0360.6—————
Generative Network InversionBackbone=ASFormer2025.0359.854.147.734.148.8—
HMM-RNNFraction=0.012021.0358.8—————
OnlineTASInference mode=semi-online, Post-processing=false2024.1157.417.814.810.119.6—
OnlineTASInference mode=online, Post-processing=false2024.1156.716.813.99.319.3—
HTK(64)Features=IDT2019.0356.3—————
Generative Network InversionBackbone=MS-TCN2025.0355.546.741.128.745.6—
Causal TCNInference mode=online2024.1155.315.111.78.318.7—
Generative Network Inversion†Backbone=MS-TCN2025.0354.149.844.333.153.3—
OnlineTASInference mode=semi-online, Post-processing=true2024.1153.856.447.331.457.5—
RoboSubtaskNet2026.0253.548.441.730.452.03—
OnlineTASInference mode=online, Post-processing=true2024.1152.954.845.830.555.7—
MeanBackbone=ASFormer2025.0352.243.538.326.743.2—
TCFPNFeatures=IDT2019.0352—————
TCFPN2020.0352—————
CoresetBackbone=ASFormer2025.035248.342.429.748.1—
HTKFeatures=IDT2019.0350.7—————
CDFLSupervision=Transcripts2021.0350.2—————
CoresetBackbone=MS-TCN2025.0349.732.327.519.336.8—