Action Segmentation On Breakfast 1

평가 지표

Acc

Average F1

Edit

F1@10%

F1@25%

F1@50%

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Acc	Average F1	Edit	F1@10%	F1@25%	F1@50%	Paper Title	Repository
EUT	75	69.3	74.6	76.2	71.8	59.8	Do we really need temporal convolutions in action segmentation?
ASQuery	77.9	74.6	78.4	80.7	76.5	66.5	ASQuery: A Query-based Model for Action Segmentation
TSA (Spectral)	63.2	-	-	-	-	-	Leveraging triplet loss for unsupervised action segmentation
SSTDA	70.2	66.4	73.7	75.0	69.1	55.2	Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation
G2L(SSTDA)	70.8	66.9	74.5	76.3	69.9	54.6	Global2Local: Efficient Structure Search for Video Action Segmentation
SF-TMN(ASFormer)	77.0	71.6	77.0	78.7	74.0	62.2	SF-TMN: SlowFast Temporal Modeling Network for Surgical Phase Recognition	-
TW-FINCH (K=avg/activity)	62.7	-	-	-	-	-	Temporally-Weighted Hierarchical Clustering for Unsupervised Action Segmentation
ASRF	67.6	66.4	72.4	74.3	68.9	56.1	Alleviating Over-segmentation Errors by Detecting Action Boundaries
TSA (FINCH)	65.1	-	-	-	-	-	Leveraging triplet loss for unsupervised action segmentation
TSA (Kmeans)	63.7	-	-	-	-	-	Leveraging triplet loss for unsupervised action segmentation
ETSN	67.8	66.4	70.3	74.0	69.0	56.2	Efficient Two-Step Networks for Temporal Action Segmentation
DPRN	71.7	67.9	75.1	75.6	70.5	57.6	Maximization and restoration: Action segmentation through dilation passing and temporal reconstruction	-
FACT (efficient hybrid of convolution and transformer model)	76.2	74.7	79.7	81.4	76.5	66.2	FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation
CETNet	74.9	71.8	77.8	79.3	74.3	61.9	Cross-Enhancement Transformer for Action Segmentation
DiffAct	76.4	73.6	78.4	80.3	75.9	64.6	Diffusion Action Segmentation
DS-TCN	70.75	59.6	69.02	67.70	62.05	49.18	Depthwise Separable Temporal Convolutional Network for Action Segmentation	-
ASRF + HASR	69.4	67.1	71.9	74.7	69.5	57.0	Refining Action Segmentation With Hierarchical Video Representations
ASPnet	75.9	70.6	76.3	78.1	72.9	60.8	ASPnet: Action Segmentation With Shared-Private Representation of Multiple Data Sources	-
MS-TCN (IDT)	65.1	50.6	61.4	58.2	52.9	40.8	MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation
UVAST	69.7	68.8	77.1	76.9	71.5	58	Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation

0 of 37 row(s) selected.