Optical Flow Estimation On Sintel Final

评估指标

Average End-Point Error

评测结果

各个模型在此基准测试上的表现结果

模型名称	Average End-Point Error	Paper Title	Repository
Spynet	8.36	Optical Flow Estimation using a Spatial Pyramid Network	-
DEQ-Flow-H	2.886	Deep Equilibrium Optical Flow Estimation	-
SelFlow	4.26	SelFlow: Self-Supervised Learning of Optical Flow	-
FastFlowNet-ft	6.08	FastFlowNet: A Lightweight Network for Fast Optical Flow Estimation	-
FDFlowNet-ft	5.11	FDFlowNet: Fast Optical Flow Estimation using a Deep Lightweight Network	-
RPKNet	2.657	Recurrent Partial Kernel Network for Efficient Optical Flow Estimation
MR-Flow	5.38	Optical Flow in Mostly Rigid Scenes	-
MaskFlownet	4.17	MaskFlownet: Asymmetric Feature Matching with Learnable Occlusion Mask	-
GMFlow	2.37	Unifying Flow, Stereo and Depth Estimation	-
CroCo-Flow	2.436	CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flow	-
VCN	4.40	Volumetric Correspondence Networks for Optical Flow
IRR-PWC	4.579	Iterative Residual Refinement for Joint Optical Flow and Occlusion Estimation	-
Perceiver IO	2.42	Perceiver IO: A General Architecture for Structured Inputs & Outputs	-
Ef-RAFT	2.60	Rethinking RAFT for Efficient Optical Flow	-
UnrolledCost	5.8	Cost Function Unrolling in Unsupervised Optical Flow	-
GMFlowNet	2.648	Global Matching with Overlapping Attention for Optical Flow Estimation	-
ScopeFlow	4.098	ScopeFlow: Dynamic Scene Scoping for Optical Flow	-
GMA	2.470	Learning to Estimate Hidden Motions with Global Motion Aggregation	-
ContinualFlow + ft	4.52	Continual Occlusions and Optical Flow Estimation	-
LiteFlowNet2-ft	4.69	A Lightweight Optical Flow CNN - Revisiting Data Fidelity and Regularization	-

0 of 27 row(s) selected.