Video Object Segmentation On Youtube Vos

评估指标

F-Measure (Seen)

F-Measure (Unseen)

Jaccard (Seen)

Jaccard (Unseen)

Overall

Params(M)

Speed (FPS)

评测结果

各个模型在此基准测试上的表现结果

								Paper Title
Cutie+ (base, MEGA)	91.0	90.1	86.6	82.2	87.5	-	-	Putting the Object Back into Video Object Segmentation
SwinB-AOTv2-L (all frames, MS)	90.7	88.9	85.6	80.7	86.5	65.6	-	Scalable Video Object Segmentation with Identification Mechanism
SwinB-DeAOT-L	90.6	88.4	85.6	80.0	86.2	70.3	-	Decoupling Features in Hierarchical Propagation for Video Object Segmentation
XMem (BL30K, MS)	90.3	90.2	85.6	81.7	86.9	-	-	XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model
R50-AOTv2-L (all frames)	90.2	87.3	85.1	78.9	85.4	15.1	-	Scalable Video Object Segmentation with Identification Mechanism
SwinB-AOTv2-L (all frames)	90.1	88.2	-	79.6	85.8	-	-	Scalable Video Object Segmentation with Identification Mechanism
SwinB-AOT-L (all frames)	90.1	86.9	85.1	78.4	85.1	65.4	5.2	Associating Objects with Transformers for Video Object Segmentation
XMem (MS)	89.9	89.9	85.3	81.7	86.7	-	-	XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model
R50-DeAOT-L	89.9	88.7	84.9	80.4	86.0	19.8	-	Decoupling Features in Hierarchical Propagation for Video Object Segmentation
XMem (BL30K)	89.8	89.2	85.1	80.3	86.1	-	22.6	XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model
R50-AOT-L (all frames)	89.5	88.2	84.5	79.6	85.5	14.9	6.4	Associating Objects with Transformers for Video Object Segmentation
DeAOT-L	89.4	-	84.2	78.6	84.8	-	-	Decoupling Features in Hierarchical Propagation for Video Object Segmentation
XMem	89.3	88.7	84.6	80.2	85.7	-	22.6	XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model
SwinB-AOT-L	89.3	86.4	84.3	77.9	84.5	65.4	9.3	Associating Objects with Transformers for Video Object Segmentation
DeAOT-B	88.9	87.0	83.9	78.5	84.6	13.2	-	Decoupling Features in Hierarchical Propagation for Video Object Segmentation
R50-AOST (L'=3)	88.8	87.9	83.8	79.3	85.0	15.4	-	Scalable Video Object Segmentation with Identification Mechanism
AOT-L (all frames)	88.8	87.1	83.7	78.4	84.5	8.3	6.5	Associating Objects with Transformers for Video Object Segmentation
R50-AOST (L'=2)	88.5	87.2	83.5	78.8	84.5	13.9	-	Scalable Video Object Segmentation with Identification Mechanism
AOT-B (all frames)	88.5	86.5	83.6	78.0	84.1	8.3	20.5	Associating Objects with Transformers for Video Object Segmentation
R50-AOT-L	88.5	86.1	83.7	78.1	84.1	14.9	14.9	Associating Objects with Transformers for Video Object Segmentation

0 of 53 row(s) selected.