Action Segmentation On 50 Salads 1

评估指标

Acc
Edit
F1@10%
F1@25%
F1@50%

评测结果

各个模型在此基准测试上的表现结果

Paper TitleRepository
Br-Prompt+ASPnet (RGB, flow, accelerometer)91.487.592.791.688.5ASPnet: Action Segmentation With Shared-Private Representation of Multiple Data Sources-
SF-TMN(ASFormer)89.884.489.188.082.9SF-TMN: SlowFast Temporal Modeling Network for Surgical Phase Recognition-
BaFormer89.584.289.388.483.9Efficient Temporal Action Segmentation via Boundary-aware Query Voting
DiffAct88.985.090.189.283.7Diffusion Action Segmentation
Semantic2Graph88.689.191.590.287.3Semantic2Graph: Graph-based Multi-modal Feature Fusion for Action Segmentation in Videos-
Br-Prompt+ASFormer88.183.889.287.881.3Bridge-Prompt: Towards Ordinal Action Understanding in Instructional Videos
LTContext87.783.289.487.782.0How Much Temporal Long-Term Context is Needed for Action Segmentation?
EUT87.482.989.287.581Do we really need temporal convolutions in action segmentation?
UVAST87.483.989.187.681.7Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation
DPRN87.282.087.886.379.4Maximization and restoration: Action segmentation through dilation passing and temporal reconstruction-
CETNet86.981.787.686.580.1Cross-Enhancement Transformer for Action Segmentation
ASFormer+ASRF85.981.985.185.479.3ASFormer: Transformer for Action Segmentation
ASFormer85.679.685.183.476.0ASFormer: Transformer for Action Segmentation
C2F-TCN84.976.484.381.872.6Coarse to Fine Multi-Resolution Temporal Convolutional Network
ASRF84.579.384.983.577.3Alleviating Over-segmentation Errors by Detecting Action Boundaries
BCN84.474.382.381.374Boundary-Aware Cascade Networks for Temporal Action Segmentation-
ASRF + HASR83.981.086.685.778.5Refining Action Segmentation With Hierarchical Video Representations-
MS-TCN++83.774.380.778.570.1MS-TCN++: Multi-Stage Temporal Convolutional Network for Action Segmentation
SSTDA83.275.883.081.573.8Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation
DA83.275.282.080.172.5Action Segmentation with Mixed Temporal Domain Adaptation-
0 of 28 row(s) selected.
Action Segmentation On 50 Salads 1 | SOTA | HyperAI超神经